技术领域technical field
本发明涉及视频数据的存储技术，尤其涉及一种视频数据的存储方法及设备。The invention relates to video data storage technology, in particular to a video data storage method and equipment.
背景技术Background technique
视频编码器所输出的视频数据，一般通过大容量的存储设备进行循环存储。视频编码器输出的视频数据通常都是连续的，而存储设备的存储空间不管是大还是小，容量总是确定的。因此，对于确定的一个存储设备而言，能够存储的视频数据相对而言都有一个相对确定的上限。The video data output by the video encoder is generally stored cyclically through a large-capacity storage device. The video data output by the video encoder is usually continuous, and the storage space of the storage device is always determined whether it is large or small. Therefore, for a certain storage device, the video data that can be stored has a relatively certain upper limit.
随着摄像头硬件以及周边技术的发展，视频清晰度等也跟着提升，同样时间段的视频数据越来越大，所需要的存储空间也越来越大。虽然随着硬件技术的不断进步，存储设备的容量也在快速攀升，但是社会生活的快速发展，视频数据的存储需求越来越广泛，存储设备的发展技术已经难以跟上视频数据存储需求以及高清晰度等的快速发展。一般是通过增加更多的存储设备来应对快速增长的视频数据的存储需求。With the development of camera hardware and peripheral technologies, the video resolution has also been improved. The video data in the same time period is getting larger and larger, and the required storage space is also getting larger. Although with the continuous advancement of hardware technology, the capacity of storage devices is also rising rapidly, but with the rapid development of social life, the storage requirements of video data are becoming more and more extensive, and the development technology of storage devices has been difficult to keep up with the storage requirements of video data and high Rapid development of clarity etc. Generally, more storage devices are added to cope with the rapidly growing video data storage requirements.
对于一块存储设备而言，通常在存储设备剩余存储空间较少时，是通过预设的循环覆盖策略删除较早存储的视频数据，腾出存储空间后存储最近接收的视频数据。如何以尽量小的存储空间存储尽量大的视频数据，或者说尽量延长视频数据的存储周期，是视频数据存储技术领域中亟待解决的重要技术问题之一。For a storage device, usually when the remaining storage space of the storage device is small, the video data stored earlier is deleted through the preset cycle overwriting strategy, and the recently received video data is stored after freeing up storage space. How to store video data as large as possible in as small a storage space as possible, or prolong the storage period of video data as much as possible, is one of the important technical problems to be solved urgently in the field of video data storage technology.
发明内容Contents of the invention
本发明所要解决的技术问题是为了提出一种新的视频存储技术，实现以尽量小的存储空间存储尽量大的视频数据。The technical problem to be solved by the present invention is to propose a new video storage technology to store as large video data as possible in as small a storage space as possible.
本发明提供了一种视频数据的存储方法，包括：将视频的连续数据段存储到存储设备的第一存储模块中；根据预设间隔从所述连续数据段中确定不相邻的两个I帧；根据所述不相邻的两个I帧的相似性确定抽帧步长；利用所述抽帧步长从所述连续数据段中抽取出一部分I帧作为剩余I帧；将所述剩余I帧存储到所述存储设备的第二存储模块中。The present invention provides a method for storing video data, comprising: storing continuous data segments of video in a first storage module of a storage device; determining two non-adjacent I frame; determine the frame extraction step size according to the similarity of the two non-adjacent I frames; utilize the frame extraction step size to extract a part of the I frame from the continuous data segment as the remaining I frame; The I frame is stored in the second storage module of the storage device.
优选地，根据所述不相邻的两个I帧的相似性确定抽帧步长，包括：获取所述不相邻的两个I帧的灰度图；根据所述不相邻的两个I帧的灰度图判断所述不相邻的两个I帧的相似性；根据所述相似性确定所述抽帧步长。Preferably, determining the frame drawing step size according to the similarity of the two non-adjacent I frames includes: acquiring the grayscale images of the two non-adjacent I frames; The grayscale image of the I frame judges the similarity of the two non-adjacent I frames; and determines the frame extraction step according to the similarity.
优选地，获取所述不相邻的两个I帧的灰度图，包括：降低所述不相邻的两个I帧的分辨率并保留灰度信息，获取所述不相邻的两个I帧的灰度缩略图；根据所述不相邻的两个I帧的灰度图判断所述不相邻的两个I帧的相似性，包括：根据所述不相邻的两个I帧的灰度缩略图判断所述不相邻的两个I帧的相似性。Preferably, obtaining the grayscale images of the two non-adjacent I frames includes: reducing the resolution of the two non-adjacent I frames and retaining the grayscale information, and obtaining the two non-adjacent I frames The gray scale thumbnail of the I frame; judging the similarity of the two non-adjacent I frames according to the gray scale images of the two non-adjacent I frames, including: according to the two non-adjacent I frames The grayscale thumbnail of the frame judges the similarity of the two non-adjacent I frames.
优选地，根据所述相似性确定所述抽帧步长，包括：所述相似性大于等于预设的相似度阈值时，将所述预设间隔作为所述抽帧步长；所述相似性小于所述相似度阈值时，缩小所述预设间隔，并根据缩小后的预设间隔重新从所述连续数据段中确定所述不相邻的两个I帧。Preferably, determining the frame drawing step according to the similarity includes: when the similarity is greater than or equal to a preset similarity threshold, using the preset interval as the frame drawing step; the similarity When it is smaller than the similarity threshold, reduce the preset interval, and re-determine the two non-adjacent I frames from the continuous data segment according to the reduced preset interval.
优选地，从所述连续数据段中确定不相邻的两个I帧，包括：接收服务器发送的命令，根据所述命令从所述连续数据段中确定所述不相邻的两个I帧；或者，在所述第一存储模块可用空间与所述存储器的容量的比值小于等于预设的比例阈值时，从所述连续数据段中确定所述不相邻的两个I帧。Preferably, determining two non-adjacent I frames from the continuous data segment includes: receiving a command sent by the server, and determining the two non-adjacent I frames from the continuous data segment according to the command or, when the ratio of the available space of the first storage module to the capacity of the memory is less than or equal to a preset ratio threshold, determine the two non-adjacent I frames from the continuous data segment.
优选地，该方法包括：从所述连续数据段中抽取出所述一部分I帧后，从所述第一存储模块中删除所述连续数据段。Preferably, the method includes: after extracting the part of I frames from the continuous data segment, deleting the continuous data segment from the first storage module.
优选地，该方法包括：接收循环覆盖时间点，在所述第二存储模块中删除结束时间在所述循环覆盖时间点之前的剩余I帧。Preferably, the method includes: receiving a cyclic coverage time point, and deleting the remaining I frames whose end time is before the cyclic coverage time point in the second storage module.
本发明还提供了一种视频数据的存储设备，包括存储器，所述存储器包括第一存储模块和第二存储模块，该设备还包括：输入单元，将视频的连续数据段存储到所述第一存储模块中，以及将剩余I帧存储到所述第二存储模块中；处理器，根据预设间隔从所述连续数据段中确定不相邻的两个I帧，根据所述不相邻的两个I帧的相似性确定抽帧步长，并利用所述抽帧步长从所述连续数据段中抽取出一部分I帧作为所述剩余I帧。The present invention also provides a video data storage device, which includes a memory, the memory includes a first storage module and a second storage module, and the device also includes: an input unit for storing continuous video data segments in the first In the storage module, and store the remaining I frames in the second storage module; the processor determines two non-adjacent I frames from the continuous data segment according to a preset interval, and according to the non-adjacent The similarity of two I frames determines the frame extraction step, and uses the frame extraction step to extract a part of I frames from the continuous data segment as the remaining I frame.
优选地，所述处理器获取所述不相邻的两个I帧的灰度图，根据所述不相邻的两个I帧的灰度图判断所述不相邻的两个I帧的相似性，根据所述相似性确定所述抽帧步长。Preferably, the processor acquires the grayscale images of the two non-adjacent I frames, and judges the grayscale images of the two non-adjacent I frames according to the grayscale images of the two non-adjacent I frames. Similarity, determining the frame drawing step size according to the similarity.
优选地，所述处理器降低所述不相邻的两个I帧的分辨率并保留灰度信息，获取所述不相邻的两个I帧的灰度缩略图，根据所述不相邻的两个I帧的灰度缩略图判断所述不相邻的两个I帧的相似性。Preferably, the processor reduces the resolution of the two non-adjacent I frames and retains the grayscale information, acquires the grayscale thumbnails of the two non-adjacent I frames, and according to the non-adjacent The gray-scale thumbnails of the two I-frames determine the similarity of the two non-adjacent I-frames.
优选地，所述处理器在所述相似性大于等于预设的相似度阈值时，将所述预设间隔作为所述抽帧步长；在所述相似性小于所述相似度阈值时，缩小所述预设间隔，并根据缩小后的预设间隔重新从所述连续数据段中确定所述不相邻的两个I帧。Preferably, when the similarity is greater than or equal to a preset similarity threshold, the processor uses the preset interval as the frame drawing step; when the similarity is smaller than the similarity threshold, zooms out the preset interval, and re-determine the two non-adjacent I frames from the continuous data segment according to the reduced preset interval.
优选地，所述处理器根据命令从所述连续数据段中确定所述不相邻的两个I帧；或者在所述第一存储模块可用空间与所述存储器的容量的比值小于等于预设的比例阈值时，从所述连续数据段中确定所述不相邻的两个I帧；其中，所述命令由所述设备的通讯单元从服务器接收。Preferably, the processor determines the two non-adjacent I frames from the continuous data segment according to the command; or the ratio of the available space of the first storage module to the capacity of the memory is less than or equal to a preset When the proportion threshold of , determine the two non-adjacent I frames from the continuous data segment; wherein, the command is received from the server by the communication unit of the device.
优选地，所述处理器从所述连续数据段中抽取出所述一部分I帧后，从所述第一存储模块中删除所述连续数据段。Preferably, the processor deletes the continuous data segment from the first storage module after extracting the part of I frames from the continuous data segment.
优选地，该设备包括：通讯单元，接收循环覆盖时间点；所述处理器从所述第二存储模块中删除结束时间在所述循环覆盖时间点之前的剩余I帧。Preferably, the device includes: a communication unit, configured to receive a cyclic coverage time point; the processor deletes the remaining I frames whose end time is before the cyclic coverage time point from the second storage module.
与现有技术相比，本发明通过去冗余技术缩小了视频数据的存储量，对于既定容量的存储设备而言，延长了视频数据的存储周期。本发明可以自适应地从视频数据中抽取合适的I帧以尽量减少视频数据的存储需求。本发明通过去冗余技术对连续的视频数据进行抽帧转存，缩小了数据的存储需求，可以高效地延长视频数据的存储周期。本发明在自适应地抽帧处理之后仅保留有用I帧，并非按固定时间间隔来抽取I帧并直接存储，体现了智能性和自适应性。本发明中，抽帧处理仅需将存储器区分为两个存储空间，比如设置两个存储器或者将同一个存储器划分为两个存储区域，通过连续存储周期到达来触发抽帧转存，操作便利。本发明中，连续数据段抽帧完成后立即作为冗余数据被删除，冗余数据生存期短，抽帧过程使得存储设备的存储空间使用率更高。Compared with the prior art, the present invention reduces the storage capacity of the video data through the de-redundancy technology, and prolongs the storage period of the video data for a storage device with a given capacity. The present invention can adaptively extract the appropriate I frame from the video data to reduce the storage requirement of the video data as much as possible. The present invention extracts frames and dumps continuous video data through the redundant technology, reduces the storage requirement of data, and can efficiently prolong the storage period of video data. The present invention only retains useful I frames after adaptively extracting frames, instead of extracting I frames at fixed time intervals and storing them directly, which embodies intelligence and adaptability. In the present invention, the frame extraction process only needs to divide the memory into two storage spaces, such as setting up two memories or dividing the same memory into two storage areas, and the frame extraction is triggered by the arrival of consecutive storage cycles, which is convenient to operate. In the present invention, the continuous data segment is immediately deleted as redundant data after frame extraction is completed, and the redundant data has a short lifetime, and the frame extraction process makes the storage space utilization rate of the storage device higher.
本发明的其它特征和优点将在随后的说明书中阐述，并且，部分地从说明书中变得显而易见，或者通过实施本发明的技术方案而了解。本发明的目的和其他优点可通过在说明书、权利要求书以及附图中所特别指出的结构和/或流程来实现和获得。Other features and advantages of the present invention will be set forth in the following description, and partly become obvious from the description, or can be understood by implementing the technical solutions of the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures and/or processes particularly pointed out in the specification, claims and accompanying drawings.
附图说明Description of drawings
附图用来提供对本申请的技术方案或现有技术的进一步理解，并且构成说明书的一部分。其中，表达本申请实施例的附图与本申请的实施例一起用于解释本申请的技术方案，但并不构成对本申请技术方案的限制。The accompanying drawings are used to provide a further understanding of the technical solutions of the present application or the prior art, and constitute a part of the specification. Wherein, the drawings expressing the embodiments of the present application are used together with the embodiments of the present application to explain the technical solutions of the present application, but do not constitute limitations on the technical solutions of the present application.
图1为本发明实施例的视频数据的存储方法的流程示意图。FIG. 1 is a schematic flowchart of a method for storing video data according to an embodiment of the present invention.
图2为本发明实施例的视频数据的存储方法中抽帧的流程示意图。FIG. 2 is a schematic flow chart of frame extraction in a method for storing video data according to an embodiment of the present invention.
图3为本发明实施例的视频数据的存储设备的构造示意图。FIG. 3 is a schematic structural diagram of a storage device for video data according to an embodiment of the present invention.
具体实施方式detailed description
以下将结合附图及实施例来详细说明本发明的实施方式，借此对本发明如何应用技术手段来解决技术问题，并达成相应技术效果的实现过程能充分理解并据以实施。本申请实施例以及实施例中的各个特征，在不相冲突前提下可以相互结合，所形成的技术方案均在本发明的保护范围之内。The implementation of the present invention will be described in detail below in conjunction with the accompanying drawings and examples, so as to fully understand and implement the implementation process of how to apply technical means to solve technical problems and achieve corresponding technical effects in the present invention. The embodiments of the present application and the various features in the embodiments can be combined with each other under the premise of no conflict, and the formed technical solutions are all within the protection scope of the present invention.
另外，附图的流程图示出的步骤可以在诸如一组计算机可执行指令的计算机系统中执行。并且，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。Additionally, the steps shown in the flowcharts of the figures may be performed in a computer system, such as a set of computer-executable instructions. Also, although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in an order different from that shown or described herein.
如图1所示，本发明实施例的视频数据的存储方法，以存储设备包含有两个存储模块为例进行说明，该方法主要包括如下步骤。As shown in FIG. 1 , the method for storing video data in the embodiment of the present invention is described by taking a storage device including two storage modules as an example. The method mainly includes the following steps.
步骤S110，将视频编码器输出的视频数据按段写入到存储设备的第一存储模块中，每一段称之为连续数据段，并为连续数据段生成索引信息存储在存储设备的数据索引区，还将连续数据段的索引信息发送给管理服务器。管理服务器存储该连续数据段的索引信息。Step S110, write the video data output by the video encoder into the first storage module of the storage device by segment, each segment is called a continuous data segment, and generate index information for the continuous data segment and store it in the data index area of the storage device , and also send the index information of the continuous data segment to the management server. The management server stores index information of the continuous data segment.
本发明的实施例中，存储设备包含有两个存储部分，分别称之为第一存储模块和第二存储模块。在不同的实施例中，这两个存储部分既可以独立设置，也可以是同一个物理存储器件的不同存储空间。第一存储模块主要用于存储视频编码器输出的视频数据，第二存储模块主要用于存储从第一存储模块存储的视频数据中抽取的I帧数据。In the embodiment of the present invention, the storage device includes two storage parts, which are respectively referred to as a first storage module and a second storage module. In different embodiments, the two storage parts can be set independently, or they can be different storage spaces of the same physical storage device. The first storage module is mainly used for storing video data output by the video encoder, and the second storage module is mainly used for storing I frame data extracted from the video data stored in the first storage module.
存储设备通过网络接收到视频编码器输出的视频数据后，先缓存在为每个编码器预分配的数据缓冲区中。然后通过解析码流，解析出视频数据的每一帧，包括I帧和非I帧，再将视频数据按帧写入第一存储模块。每写完一个时间段(比如2分钟或者5分钟等)的视频数据，存储设备就为此生成一个新的连续数据段。完成一个连续数据段的写入后，再切换继续写下一个连续数据段。After the storage device receives the video data output by the video encoder through the network, it first caches it in the data buffer pre-allocated for each encoder. Then, by analyzing the code stream, each frame of video data, including I frame and non-I frame, is analyzed, and then the video data is written into the first storage module frame by frame. Every time the video data of a time period (such as 2 minutes or 5 minutes, etc.) is written, the storage device generates a new continuous data segment for this. After completing the writing of a continuous data segment, switch to continue writing the next continuous data segment.
写入第一存储模块的时间段，可以根据需要进行选取。这个时间段的长度决定了后续抽帧时冗余数据生命期的长短和抽帧效率。这个时间段的长度越长，相应的连续数据段就越大，段内I帧数也越多，抽帧花的时间也越多。由于抽帧是按连续数据段进行的，连续数据段抽帧完成之后才会被删除，这样冗余数据存在的时间也越长，整个存储设备的存储效率就会受到一定影响。时间段越长，同样数据量所生成的连续数据段的数目就越小，这样进行并发抽帧的连续数据段也越少，整体的抽帧效率就较低。The time period for writing into the first storage module can be selected as required. The length of this time period determines the length of the redundant data lifetime and the frame extraction efficiency in the subsequent frame extraction. The longer the length of this time period, the larger the corresponding continuous data segment, the more I frames in the segment, and the more time it takes to draw frames. Since the frame extraction is performed according to continuous data segments, the continuous data segments will be deleted after frame extraction is completed, so the longer the redundant data exists, the storage efficiency of the entire storage device will be affected to a certain extent. The longer the time period, the smaller the number of continuous data segments generated by the same amount of data, so the fewer continuous data segments for concurrent frame extraction, the lower the overall frame extraction efficiency.
在将连续数据段写到第一存储模块中时，为每个连续数据段生成数据段索引信息，在存储设备的数据索引区记录每个连续数据段的段编号、存储位置、开始时间、结束时间、连续数据段内每个I帧的大小和存储位置。该数据段索引信息包括每个连续数据段的段编号、存储位置、开始时间、结束时间、连续数据段内每个I帧的大小和存储位置。I帧的存储位置主要用来后续抽帧处理时能够快速定位出I帧，提高抽帧效率。When writing continuous data segments into the first storage module, generate data segment index information for each continuous data segment, and record the segment number, storage location, start time, and end of each continuous data segment in the data index area of the storage device The time, size and storage location of each I-frame within the contiguous data segment. The data segment index information includes the segment number, storage location, start time, end time, size and storage location of each I frame in the continuous data segment for each continuous data segment. The storage location of the I frame is mainly used for quickly locating the I frame during subsequent frame extraction processing, and improving the efficiency of frame extraction.
步骤S120，管理服务器根据自己保存的存储设备上报的索引信息，定期检测连续存储周期是否到达。第一存储模块中的每个连续数据段都有自己的开始时间和结束时间，通过比较管理服务器当前时间和某一个连续数据段的结束时间，就可以得知该连续数据段是否达到了存储周期。如果管理服务器上的当前时间减去该连续数据段的结束时间，所得差值大于预设的连续存储周期，就认为该了连续数据段达到了连续存储周期。在连续存储周期到达时，向存储设备发送抽帧转存命令。In step S120, the management server periodically detects whether the continuous storage period is reached according to the index information reported by the storage device saved by the management server. Each continuous data segment in the first storage module has its own start time and end time. By comparing the current time of the management server with the end time of a certain continuous data segment, it can be known whether the continuous data segment has reached the storage cycle . If the end time of the continuous data segment is subtracted from the current time on the management server, and the resulting difference is greater than the preset continuous storage period, it is considered that the continuous data segment has reached the continuous storage period. When the continuous storage period arrives, a frame dump command is sent to the storage device.
步骤S130，存储设备收到抽帧转存命令时，根据该抽帧转存命令，对连续数据段进行抽帧处理，从连续数据段中抽取出I帧并进行I帧的去冗余处理，然后将去冗余处理之后获得的剩余I帧存储到第二存储模块中。存储设备还删除第一存储模块中对应的连续数据段，向管理服务器发送删除消息，通知管理服务器删除该连续数据段的索引信息。Step S130, when the storage device receives the frame extraction and dump command, according to the frame extraction and dump command, the continuous data segment is subjected to frame extraction processing, the I frame is extracted from the continuous data segment and the I frame is de-redundantly processed, Then store the remaining I frames obtained after de-redundancy processing into the second storage module. The storage device also deletes the corresponding continuous data segment in the first storage module, sends a deletion message to the management server, and notifies the management server to delete the index information of the continuous data segment.
管理服务器根据该删除消息，删除该连续数据段的索引信息。The management server deletes the index information of the continuous data segment according to the deletion message.
在其他的实施例中，存储设备根据第一存储模块的可用空间与自身的容量的比值与预设的比例阈值来确定何时进行抽帧处理。比如，存储设备在第一存储模块的可用空间与自身的容量的比值小于等于预设的比例阈值(比如10％)时，认为第一存储模块余下的可用空间已经较少，难以存储较大数据量的视频数据，此时开始执行从连续数据段中确定不相邻的两个I帧并进行后续的抽帧处理。In other embodiments, the storage device determines when to perform frame extraction processing according to the ratio of the available space of the first storage module to its own capacity and a preset ratio threshold. For example, when the ratio of the available space of the first storage module to its own capacity is less than or equal to a preset ratio threshold (such as 10%), the storage device considers that the remaining available space of the first storage module is less and it is difficult to store larger data At this time, start to determine two non-adjacent I frames from the continuous data segment and perform subsequent frame extraction processing.
本发明的实施例中，存储设备对连续数据段进行抽帧处理时，根据预设的去冗余策略，从连续数据段中抽取出I帧，或者说去除冗余I帧，将去除冗余I帧之后的剩余I帧也即抽取出来的I帧写入到第二存储模块，减小视频数据对存储空间的需求。In the embodiment of the present invention, when the storage device performs frame extraction processing on continuous data segments, it extracts I frames from the continuous data segments according to the preset de-redundancy strategy, or removes redundant I frames, and removes redundant I frames. The remaining I frames after the I frame, that is, the extracted I frames, are written into the second storage module, so as to reduce the storage space requirement of the video data.
本发明的实施例中，存储设备将剩余I帧写入第二存储模块时，是将剩余I帧组成一个I帧数据段，然后将该I帧数据段写入第二存储模块中。存储设备还为该I帧数据段生成索引信息，将该I帧数据段的索引信息发送给管理服务器。管理服务器收到该I帧索引信息后存储。In the embodiment of the present invention, when the storage device writes the remaining I frames into the second storage module, the remaining I frames are formed into an I frame data segment, and then the I frame data segment is written into the second storage module. The storage device also generates index information for the I-frame data segment, and sends the index information of the I-frame data segment to the management server. The management server stores the I frame index information after receiving it.
存储设备上的命令通讯单元接收管理服务器下发的各种命令。当接收到管理服务器发送的抽帧转存命令时，解析出该抽帧转存命令中的抽帧时间点参数。在存储设备的数据索引区中查找结束时间在该抽帧时间点之前的连续数据段，然后根据连续数据段的段编号，将满足条件的连续数据段缓存到待抽帧段队列中。在抽帧转存时，按照段编号等从待抽帧段队列中取出相应的连续数据段，对该连续数据段进行抽帧处理。The command communication unit on the storage device receives various commands issued by the management server. When receiving the frame extraction and dumping command sent by the management server, the parameter of the frame extraction time point in the frame extraction and dumping command is parsed out. Search the data index area of the storage device for continuous data segments whose end time is before the frame extraction time point, and then buffer the continuous data segments satisfying the conditions into the frame segment queue to be extracted according to the segment numbers of the continuous data segments. When frame dumping is performed, the corresponding continuous data segment is taken out from the segment queue to be extracted according to the segment number, etc., and the frame extraction process is performed on the continuous data segment.
本发明的实施例，可以并发进行多个连续数据段的抽帧处理。In the embodiment of the present invention, frame extraction processing of multiple continuous data segments can be performed concurrently.
上述实施例是以存储设备包含有两个相互独立的存储模块为例来说明本发明的视频数据的存储方法的，在其他的实施例中，该两个存储模块也可以共存于同一个存储器中，该存储器的不同存储区域被视为不同的存储模块。也就是说，本发明的技术方案中，只要是能够将连续数据段和剩余I帧进行分开存储的存储技术，均是可行的。The above-mentioned embodiment has taken the storage device including two mutually independent storage modules as an example to illustrate the video data storage method of the present invention. In other embodiments, the two storage modules can also coexist in the same storage , different storage areas of the memory are regarded as different storage modules. That is to say, in the technical solution of the present invention, as long as it is a storage technology capable of separately storing continuous data segments and remaining I frames, all are feasible.
如图2所示，本发明的实施例针对每个连续数据段的抽帧处理，主要包括如下步骤。As shown in FIG. 2 , the embodiment of the present invention mainly includes the following steps for frame extraction processing of each continuous data segment.
步骤S210，根据连续数据段的段编号，在数据索引区查询到段内I帧的索引信息，该索引信息包括了连续数据段中I帧的大小和存储位置；根据该索引信息获得连续数据段内的所有I帧。Step S210, according to the segment number of the continuous data segment, query the index information of the I frame in the data index area, the index information includes the size and storage location of the I frame in the continuous data segment; obtain the continuous data segment according to the index information All I-frames within.
步骤S220，根据预设间隔，从连续数据段的所有I帧中确定两个I帧，这两个I帧在连续数据段的所有I帧中并不相邻。比如，对连续数据段解码获得I帧序列，以n为间隔，就可以从该I帧序列中获得两个I帧。后续就可以根据本不相邻的这两个I帧之间的相似性，来去除连续数据段中的冗余I帧，最终获得用来存储的可以代表整个连续数据段的剩余I帧。Step S220, according to the preset interval, determine two I frames from all the I frames of the continuous data segment, and these two I frames are not adjacent among all the I frames of the continuous data segment. For example, to obtain an I frame sequence by decoding continuous data segments, and taking n as an interval, two I frames can be obtained from the I frame sequence. Subsequently, the redundant I frames in the continuous data segment can be removed according to the similarity between the two non-adjacent I frames, and finally the remaining I frames that can represent the entire continuous data segment for storage are obtained.
步骤S230，根据不相邻的这两个I帧之间的相似性确定从连续数据段的整个I帧序列中抽取I帧的抽帧步长，利用该抽帧步长去除I帧序列中的冗余I帧，获得连续数据段的整个I帧序列中的剩余I帧。Step S230, according to the similarity between the two non-adjacent I frames, determine the frame extraction step size for extracting the I frame from the entire I frame sequence of the continuous data segment, and use the frame extraction step size to remove the frame size in the I frame sequence Redundant I-frames, to obtain the remaining I-frames in the entire sequence of I-frames for consecutive data segments.
比如，以n为间隔，解码I帧序列中序号为k*n和(k+1)*n的I帧，通过分析这两个I帧之间的相似性，就可以从连续数据段的整个I帧序列中去除冗余I帧，留下或者抽取出剩余I帧。其中，n为预设间隔，k为I帧的序号。For example, take n as an interval to decode the I frames whose sequence numbers are k*n and (k+1)*n in the I frame sequence, and by analyzing the similarity between these two I frames, you can learn from the entire continuous data segment Redundant I frames are removed from the I frame sequence, leaving or extracting the remaining I frames. Wherein, n is a preset interval, and k is a serial number of an I frame.
对于一个既定的预设间隔n，如果k*n和(k+1)*n这一对I帧的相似性大于等于预设的相似度阈值，说明这两个I帧所表示的内容变化不大，可以将n作为抽帧步长进行抽帧。如果这一对I帧的相似性小于该相似度阈值，则说明这两个I帧所表示的内容变化较大，这种情况下将n作为抽帧步长进行抽帧的话，就容易漏掉比较多的信息。此时，缩小预设间隔n，并根据缩小后的预设间隔n重新从连续数据段中确定前述不相邻的两个I帧，再判断重新确定的相比之前较近的两个I帧之间的相似性，并根据新确定的两个I帧之间的相似性来判断是否能将缩小后的预设间隔n作为抽帧步长。For a given preset interval n, if the similarity of the pair of I frames k*n and (k+1)*n is greater than or equal to the preset similarity threshold, it means that the content changes represented by the two I frames are not different. If it is large, n can be used as the step size of frame extraction for frame extraction. If the similarity of this pair of I frames is less than the similarity threshold, it means that the content represented by the two I frames has changed greatly. In this case, if n is used as the frame extraction step to extract frames, it is easy to miss more information. At this time, reduce the preset interval n, and re-determine the aforementioned two non-adjacent I frames from the continuous data segment according to the reduced preset interval n, and then judge the re-determined two I frames that are closer than before According to the similarity between the newly determined two I-frames, it is judged whether the reduced preset interval n can be used as the frame extraction step size.
具体地，以k*n和(k+1)*n这一对I帧为例，获取这两个I帧的灰度图。然后根据这两个I帧的灰度图来判断这两个I帧的相似性如何，后续根据这两个I帧的相似性来确定从整个I帧序列中抽取出剩余I帧的抽帧步长。根据这两个I帧的灰度图来判断这两个I帧的相似性，主要是通过降低这两个I帧的分辨率并仅保留这两个I帧的像素的灰度信息，得到这两个I帧的灰度缩略图。后续根据灰度缩略图来进行相似性判断。Specifically, taking a pair of I frames k*n and (k+1)*n as an example, the grayscale images of these two I frames are acquired. Then determine the similarity of the two I frames according to the grayscale images of the two I frames, and then determine the frame extraction step for extracting the remaining I frames from the entire I frame sequence according to the similarity of the two I frames long. Judging the similarity of the two I frames according to the grayscale images of the two I frames is mainly by reducing the resolution of the two I frames and only retaining the grayscale information of the pixels of the two I frames to obtain the Grayscale thumbnails of the two I-frames. Follow-up similarity judgments are made based on the gray scale thumbnails.
具体地，解码I帧序列获得k*n和(k+1)*n这两个I帧后，降低这两个I帧的分辨率至M*N，且仅保留像素灰度信息，得到这两个I帧的灰度缩略图。该灰度缩略图为M*N的矩阵，矩阵中每个元素对应缩略图中像素灰度，取值范围是0～255。Specifically, after decoding the I-frame sequence to obtain the two I-frames k*n and (k+1)*n, reduce the resolution of these two I-frames to M*N, and only keep the grayscale information of the pixels, so that Grayscale thumbnails of the two I-frames. The grayscale thumbnail is an M*N matrix, and each element in the matrix corresponds to the grayscale of a pixel in the thumbnail, and the value range is 0 to 255.
本发明的实施例中，n＝32，k＝(0，1，2，…)，M*N＝352*288。In the embodiment of the present invention, n=32, k=(0, 1, 2, . . . ), M*N=352*288.
根据所得的灰度缩略图之间的相似性，判断这两个I帧之间的相似性。比如对于第k*n和(k+1)*n这两个I帧的相似性，可以分两步来进行。第一步是通过判断这两个I帧的灰度缩略图中对应像素的相似性来判断这两个I帧的灰度缩略图之间的相似性；第二步是根据这两个I帧的灰度缩略图之间的相似性来判断这两个I帧之间的相似性。According to the similarity between the obtained grayscale thumbnails, the similarity between the two I-frames is judged. For example, the similarity between the two I-frames k*n and (k+1)*n can be performed in two steps. The first step is to judge the similarity between the grayscale thumbnails of the two I frames by judging the similarity of the corresponding pixels in the grayscale thumbnails of the two I frames; the second step is to judge the similarity between the grayscale thumbnails of the two I frames; The similarity between the grayscale thumbnails is used to judge the similarity between the two I-frames.
在判断两幅灰度缩略图中同一位置的像素的相似性时，将两幅灰度缩略图分别记为f和f‘，相同位置上的像素的灰度差的绝对值为abs(x，y)＝|f(x，y)-f‘(x，y)|，预设阈值T。如果abs(x，y)<＝T，则认为这两个像素具有相似性；反之，则认为这两个像素不具有相似性。其中，(x，y)为像素在灰度缩略图矩阵中的位置，取值范围是(0，0)～(M，N)。阈值T在本实施例中取值为48，在其他实施例中，该阈值T也可根据需要取其它数值。When judging the similarity of pixels at the same position in two gray scale thumbnails, the two gray scale thumbnails are respectively recorded as f and f', and the absolute value of the gray scale difference of pixels at the same position is abs(x, y)=|f(x, y)-f'(x, y)|, preset threshold T. If abs(x, y)<=T, it is considered that the two pixels have similarity; otherwise, it is considered that the two pixels have no similarity. Wherein, (x, y) is the position of the pixel in the gray scale thumbnail matrix, and the value range is (0, 0) ~ (M, N). The threshold T takes a value of 48 in this embodiment, and in other embodiments, the threshold T can also take other values as required.
在获得两幅灰度缩略图中同一位置上的像素的相似性结果之后，就可以判断两幅灰度缩略图之间的相似性了。具体地，通过像素相似性的判断，可以得到两个缩略图之间的相似性度量矩阵，该矩阵大小同为M*N，其元素记为cor(x，y)。两幅灰度缩略图中同一位置上的像素具有相似性时cor(x，y)取1，反之取0。将所相似性度量矩阵的有元素求和，记为D，再除以矩阵大小M*N，得到缩略图相似性系数r，即r＝D/(M*N)。显然，r<＝1。r越大，则表示两幅灰度缩略图之间的相似性越大，反之则相似性越小。其中，x，y，M，N取值范围请参考前文。当然，如果像素具有相似性时cor(x，y)取0，反之取1，则缩略图相似性系数r越小，则表示缩略图相似性越大，反之则相似性越小。After obtaining the similarity results of the pixels at the same position in the two grayscale thumbnails, the similarity between the two grayscale thumbnails can be judged. Specifically, by judging the similarity of pixels, a similarity measurement matrix between two thumbnails can be obtained, and the size of the matrix is M*N, and its elements are recorded as cor(x, y). Cor(x, y) takes 1 when the pixels at the same position in the two grayscale thumbnails have similarity, and takes 0 otherwise. The sum of all elements of the similarity measurement matrix is recorded as D, and then divided by the matrix size M*N to obtain the thumbnail similarity coefficient r, that is, r=D/(M*N). Obviously, r<=1. The larger r is, the greater the similarity between the two gray scale thumbnails is, otherwise the smaller the similarity is. Among them, the value range of x, y, M, N, please refer to the previous article. Of course, cor(x, y) takes 0 if the pixels have similarity, and takes 1 otherwise, the smaller the thumbnail similarity coefficient r is, the greater the similarity of the thumbnails is, otherwise the smaller the similarity is.
灰度缩略图的相似性即可以作为两个I帧的相似性。The similarity of the gray scale thumbnails can be regarded as the similarity of two I frames.
当然，在本发明的其他实施例中，不降低I帧的分辨率，直接利用I帧的像素的灰度信息来进行相邻两个I帧的相似性判断，也是可行的。通过相邻两个I帧的灰度图中同一位置上像素之间的相似性，得到两幅灰度图之间的相似性。然后根据两幅灰度图之间的相似性，就可以得到对应的两个I帧之间的相似性。Of course, in other embodiments of the present invention, it is also feasible to directly use the grayscale information of the pixels of the I frame to determine the similarity between two adjacent I frames without reducing the resolution of the I frame. The similarity between two grayscale images is obtained by the similarity between pixels at the same position in the grayscale images of two adjacent I frames. Then, according to the similarity between the two grayscale images, the similarity between the corresponding two I frames can be obtained.
根据这两个I帧之间的相似性，比如根据第k*n和(k+1)*n这两个I帧的相似性，去除整个I帧序列中的冗余I帧，抽取获得的是连续数据段中的剩余I帧，这些剩余I帧就可以代表该连续数据段。然后，就可以将该些剩余I帧写入存储设备的第二存储模块。According to the similarity between the two I frames, for example, according to the similarity between the k*n and (k+1)*n I frames, remove redundant I frames in the entire I frame sequence, and extract the obtained are the remaining I frames in the continuous data segment, and these remaining I frames can represent the continuous data segment. Then, these remaining I frames can be written into the second storage module of the storage device.
本发明实施例的视频数据的存储方法中，从第一存储模块301中删除已进行了抽帧处理的连续数据段的时机非常灵活，可以在将剩余I帧写入到第二存储模块302中后执行，也可以是在去除冗余I帧之后、将剩余I帧写入到第二存储模块302中之前进行。In the storage method of the video data of the embodiment of the present invention, from the first storage module 301, delete the timing of the continuous data segment that has carried out the frame drawing process very flexible, can write in the second storage module 302 in remaining I frame It may also be performed after removing redundant I frames and before writing the remaining I frames into the second storage module 302 .
本发明的实施例中，最后存储到第二存储模块中的剩余I帧，并不是通过预设间隔从连续数据段中抽取出的那一部分I帧，而是通过灵活的抽帧步长来确定从I帧序列中抽取那些I帧。抽取获得的这些剩余的I帧的总量，相比连续数据段而言已经大幅缩小，显著降低了对存储空间的要求。对于既定容量的存储设备而言，大幅延长了视频的存储周期。In the embodiment of the present invention, the remaining I frames stored in the second storage module are not the part of the I frames extracted from the continuous data segments at preset intervals, but are determined by flexible frame extraction steps Those I-frames are extracted from the sequence of I-frames. The total amount of these remaining I frames obtained by extraction has been greatly reduced compared with the continuous data segment, which significantly reduces the requirement for storage space. For a storage device with a given capacity, the storage period of the video is greatly extended.
通过相似性系数r来确定有用I帧的抽帧步长s，r越小，s越小，即其中冗余I帧越少。本发明的一个实施例中，r和s具有如下对应关系：The frame extraction step s of useful I frames is determined by the similarity coefficient r, the smaller r is, the smaller s is, that is, the fewer redundant I frames. In one embodiment of the present invention, r and s have the following correspondence:
r<＝0.2，s＝2；r<=0.2, s=2;
0.2<r<＝0.4，s＝4；0.2<r<=0.4, s=4;
0.4<r<＝0.6，s＝8；0.4<r<=0.6, s=8;
0.6<r<＝0.8，s＝16；0.6<r<=0.8, s=16;
0.8<r，s＝24。0.8<r, s=24.
将k*n和(k+1)*n之间的I帧，从第k*n个开始，以s为步长，根据该段I帧索引信息，将满足条件的I帧从连续数据段中取出并写入第二存储模块，写入完成后生成一个相应的I帧数据段。The I frame between k*n and (k+1)*n, starting from the k*nth, with s as the step size, according to the I frame index information of the segment, the I frame that satisfies the condition is separated from the continuous data segment Take out and write into the second storage module, and generate a corresponding I frame data segment after the writing is completed.
本发明的实施例中，存储设备还接触管理服务器下发的循环覆盖时间点。在接收到该循环覆盖时间点的信息后，从第二存储模块中删除结束时间在该循环覆盖时间点之前的剩余I帧，实现第二存储模块的存储空间的循环存储。In the embodiment of the present invention, the storage device also contacts the cycle coverage time point issued by the management server. After receiving the information of the cyclic coverage time point, delete the remaining I frame whose end time is before the cyclic coverage time point from the second storage module, so as to realize the cyclic storage of the storage space of the second storage module.
如图3所示，本发明实施例的视频数据的存储设备包括有存储器300，该存储器300包括第一存储模块301和第二存储模块302，这两个存储模块共存于存储器300中，是存储器300不同的存储区域。在其他的实施例中，该两个存储模块也可以是相互独立的两个物理存储器件。也就是说，本发明的技术方案中，只要是能够将连续数据段和剩余I帧进行分开存储的存储技术，均是可行的。该存储设备还包括输入单元310以及处理器320。As shown in FIG. 3 , the video data storage device of the embodiment of the present invention includes a memory 300, and the memory 300 includes a first storage module 301 and a second storage module 302. These two storage modules coexist in the memory 300 and are memory 300 different storage areas. In other embodiments, the two storage modules may also be two physical storage devices independent of each other. That is to say, in the technical solution of the present invention, as long as it is a storage technology capable of separately storing continuous data segments and remaining I frames, all are feasible. The storage device also includes an input unit 310 and a processor 320 .
输入单元310，与存储器300中的第一存储模块301及第二存储模块302分别相连，将视频数据按段存储到第一存储模块301中，将后续抽帧之后获得的剩余I帧存储到第二存储模块302中。The input unit 310 is connected to the first storage module 301 and the second storage module 302 in the memory 300 respectively, stores the video data in the first storage module 301 by segment, and stores the remaining I frames obtained after the subsequent frame extraction into the second storage module. In the second storage module 302 .
处理器320，与输入单元310及存储器300中的第一存储模块301相连，解码连续数据段获得I帧序列，根据预设间隔从连续数据段的该I帧序列中获取原本不相邻的两个I帧之后，根据这两个I帧之间的相似性确定从该I帧序列中抽取剩余I帧的抽帧步长，并利用所确定的抽帧步长，从该连续数据段的I帧序列中抽取出一部分I帧作为前述的剩余I帧。The processor 320 is connected to the input unit 310 and the first storage module 301 in the memory 300, decodes the continuous data segment to obtain an I frame sequence, and obtains two non-adjacent frames from the I frame sequence of the continuous data segment according to a preset interval. After the first I frame, determine the frame step size of extracting the remaining I frames from the I frame sequence according to the similarity between the two I frames, and utilize the determined frame step size to extract from the I frame of the continuous data segment A part of I frames are extracted from the frame sequence as the aforementioned remaining I frames.
本发明实施例的视频数据的存储设备，还包括接收管理服务器通过抽帧转存命令发送的抽帧时间点的通讯单元(图中未示出)。管理服务器将抽帧时间点作为参数封装在抽帧转存命令发送给存储设备，存储设备的通讯单元在收到该抽帧转存命令后，从中解析出该抽帧时间点。处理器320根据该抽帧时间点，从第一存储模块301中存储的该抽帧时间点以前的连续数据段中抽取出所述的剩余I帧。The storage device for video data in the embodiment of the present invention further includes a communication unit (not shown in the figure) that receives the frame extraction time point sent by the management server through the frame extraction and transfer command. The management server encapsulates the frame extraction time point as a parameter in the frame extraction dump command and sends it to the storage device, and the communication unit of the storage device parses the frame extraction time point after receiving the frame capture dump command. The processor 320 extracts the remaining I frames from the continuous data segments before the frame extraction time point stored in the first storage module 301 according to the frame extraction time point.
在其他的实施例中，处理器320根据第一存储模块301的可用空间与自身的容量的比值与预设的比例阈值来确定何时进行抽帧处理。比如，处理器320在第一存储模块301的可用空间与自身的容量的比值小于等于预设的比例阈值(比如10％)时，认为第一存储模块301余下的可用空间已经较少，难以存储较大数据量的视频数据，此时开始执行从连续数据段中确定不相邻的两个I帧并进行后续的抽帧处理。In other embodiments, the processor 320 determines when to perform frame capture processing according to the ratio of the available space of the first storage module 301 to its own capacity and a preset ratio threshold. For example, when the ratio of the available space of the first storage module 301 to its own capacity is less than or equal to a preset ratio threshold (such as 10%), the processor 320 considers that the remaining available space of the first storage module 301 is too small to store For video data with a large amount of data, at this time, start to determine two non-adjacent I frames from continuous data segments and perform subsequent frame extraction processing.
本发明实施例的视频数据的存储设备中，处理器320从连续数据段中抽取出剩余I帧后，从第一存储模块301中删除已进行了抽帧处理的连续数据段。处理器320从第一存储模块301中删除已进行了抽帧处理的连续数据段的时机非常灵活，可以是抽取出剩余I帧之后、输入单元310将剩余I帧写到存储器300的第二存储模块302中之前进行，也可以是在输入单元310将剩余I帧写入到第二存储模块302中之后再进行。连续数据段抽帧完成后及时被删除，生存期短，可以提高第一存储模块301的存储空间使用率。In the storage device for video data in the embodiment of the present invention, after the processor 320 extracts the remaining I frames from the continuous data segments, it deletes the continuous data segments that have been subjected to frame extraction processing from the first storage module 301 . The processor 320 deletes from the first storage module 301 the opportunity to delete the continuous data segments that have been processed by frame extraction. It can be that after the remaining I frames are extracted, the input unit 310 writes the remaining I frames to the second storage of the memory 300. before the module 302, or after the input unit 310 writes the remaining I frames into the second storage module 302. The continuous data segment is deleted in time after the frame extraction is completed, and the lifetime is short, which can improve the storage space usage rate of the first storage module 301 .
通讯单元接收管理服务器通过循环覆盖删除命令发送循环覆盖时间点这一参数。管理服务器将循环覆盖时间点作为参数封装在循环覆盖删除命令中发送给存储设备，存储设备的通讯单元在收到该循环覆盖删除命令后，从中解析出循环覆盖时间点。处理器根据该循环覆盖时间点从第二存储模块302中删除结束时间在该循环覆盖时间点之前的剩余I帧。删除这些剩余I帧之后，就可以在第二存储模块302中腾出存储空间，以利后续继续存储新的剩余I帧。The communication unit receives the parameter of cyclic coverage time point sent by the management server through the cyclic coverage deletion command. The management server encapsulates the cycle coverage time point as a parameter in the cycle coverage deletion command and sends it to the storage device. After receiving the cycle coverage deletion command, the communication unit of the storage device parses the cycle coverage time point from it. The processor deletes, from the second storage module 302, the remaining I frames whose end time is before the loop coverage time point according to the loop coverage time point. After these remaining I frames are deleted, storage space can be freed in the second storage module 302 to facilitate subsequent storage of new remaining I frames.
本发明的实施例中，去冗余模块330根据所抽取出的一部分I帧中相邻两个I帧之间的相似性，从这一部分I帧中去除冗余I帧。In the embodiment of the present invention, the de-redundancy module 330 removes redundant I frames from the extracted part of I frames according to the similarity between two adjacent I frames in the extracted part of I frames.
本发明的实施例中，处理器320从连续数据段的I帧序列中获取不相邻的两个I帧的灰度图之后，根据这两个不相邻的I帧的灰度图判断这两个I帧的相似性，再根据这两个I帧的相似性确定抽帧步长。在这个过程中，处理器320降低这两个I帧的分辨率并保留各自的灰度信息，获取这两个I帧的灰度缩略图，再根据这两个I帧的灰度缩略图，来判断这两个I帧之间的相似性。这两个I帧的相似性大于等于预设的相似度阈值，则说明这两个I帧所表示的内容变化不大，处理器320将这两个I帧之间的间距，也即前述的预设间隔作为抽帧步长进行抽帧处理，通常而言不会漏掉重要的信息。如果这两个I帧的相似性小于该相似度阈值，则说明这两个I帧所表示的内容变化较大，这种情况下将前述的预设间隔作为抽帧步长进行抽帧的话，就容易漏掉比较多的信息。此时，处理器320缩小前述的预设间隔，使得这两个I帧相距的更近，这样避免抽帧遗漏较多的信息。处理器320再根据缩小后的预设间隔重新从连续数据段中确定两个I帧，再判断重新确定的相比之前较近的两个I帧之间的相似性，并根据新确定的两个I帧之间的相似性来判断是否能将缩小后的预设间隔作为抽帧步长。In the embodiment of the present invention, after the processor 320 obtains the grayscale images of two non-adjacent I frames from the I frame sequence of the continuous data segment, it judges the grayscale image according to the two nonadjacent I frame grayscale images. The similarity of the two I frames, and then determine the frame extraction step according to the similarity of the two I frames. In this process, the processor 320 reduces the resolution of the two I frames and retains their respective grayscale information, obtains the grayscale thumbnails of the two I frames, and then according to the grayscale thumbnails of the two I frames, to judge the similarity between the two I frames. If the similarity of these two I frames is greater than or equal to the preset similarity threshold, it means that the content represented by these two I frames does not change much, and the processor 320 calculates the distance between these two I frames, that is, the aforementioned The preset interval is used as the frame extraction step to perform frame extraction processing, and generally speaking, important information will not be missed. If the similarity of the two I frames is less than the similarity threshold, it means that the content represented by the two I frames has changed greatly. In this case, if the aforementioned preset interval is used as the frame extraction step to extract frames, It is easy to miss more information. At this time, the processor 320 reduces the aforementioned preset interval, so that the two I frames are closer to each other, so as to avoid omitting more information when drawing frames. The processor 320 then re-determines two I-frames from the continuous data segment according to the reduced preset interval, then judges the similarity between the re-determined two I-frames that are closer than before, and based on the newly determined two I-frames The similarity between two I frames is used to judge whether the reduced preset interval can be used as the frame extraction step.
本发明的实施例中，一个连续数据段在抽帧完成后，就可以从第一存储模块中删除，以腾出存储空间，提高整个存储设备的使用效率。并从数据索引区中删除该连续数据段的索引信息，写入新生成的I帧数据段的索引信息。删除该连续数据段的索引信息以及写入I帧数据段的索引信息，也将上报给管理服务器。In the embodiment of the present invention, a continuous data segment can be deleted from the first storage module after frame extraction is completed, so as to free up storage space and improve the use efficiency of the entire storage device. And delete the index information of the continuous data segment from the data index area, and write the index information of the newly generated I frame data segment. Deleting the index information of the continuous data segment and writing the index information of the I frame data segment will also be reported to the management server.
本发明的实施例，在I帧存储周期到达时，管理服务器向存储设备发送循环覆盖删除命令。存储设备接收到管理服务器发送的循环覆盖删除命令后，解析出该循环覆盖删除命令中的循环覆盖时间点参数。在存储设备的数据索引区查找结束时间在该循环覆盖时间点之前的I帧数据段，然后将结束时间在该循环覆盖时间点之前的I帧数据段从第二存储模块中删除，以腾出空间存储后续生成的I帧数据段。In the embodiment of the present invention, when the I-frame storage period arrives, the management server sends a cycle overwriting delete command to the storage device. After receiving the cyclic overwriting deletion command sent by the management server, the storage device parses out the cyclic overwriting time point parameter in the cyclic overwriting deletion command. In the data index area of storage device, the I frame data segment whose end time is searched before the cycle coverage time point is deleted from the second storage module with the end time before the cycle coverage time point, so as to free up Space for storing subsequent I-frame data segments.
本发明的实施例中，管理服务器将存储设备上报的段信息保存下来，根据连续数据段的开始时间和结束时间维护第一存储模块两个检测时间：第一存储模块记录连续数据段的开始时间(TStartC)和结束时间(TEndC)，其对应的是第一存储模块中所存储的所有连续数据段的开始时间的最小值和结束时间的最大值。管理服务器还根据I帧数据段的开始时间和结束时间维护第二存储模块两个检测时间：第二存储模块记录I帧数据段的开始时间(TStartI)和结束时间(TEndI)，其对应的是第二存储模块所存储的所有I帧数据段的开始时间的最小值和结束时间的最大值。In the embodiment of the present invention, the management server saves the segment information reported by the storage device, and maintains two detection times of the first storage module according to the start time and end time of the continuous data segment: the first storage module records the start time of the continuous data segment (TStartC) and end time (TEndC), which correspond to the minimum value of the start time and the maximum value of the end time of all continuous data segments stored in the first storage module. The management server also maintains two detection times of the second storage module according to the start time and the end time of the I frame data segment: the second storage module records the start time (TStartI) and the end time (TEndI) of the I frame data segment, and its corresponding The minimum value of the start time and the maximum value of the end time of all I-frame data segments stored in the second storage module.
管理服务器定期检测周期是否到达。如果管理服务器的抽帧时间点TCheckC(系统当前时间减去连续存储周期)在第一存储模块所存储的连续数据段所对应的时间范围之内，即TStartC<TCheckC<TEndC，则管理服务器给存储设备下发抽帧转存命令，命令中含有抽帧时间点TCheckC。如果管理服务器的循环覆盖时间点TCheckI(系统当前时间减去循环覆盖周期)在第二存储模块所存储的I帧数据段所对应的时间范围之内，即TStartI<TCheckI<TEndI，则管理服务器给存储设备下发循环覆盖删除命令，命令中含有循环覆盖时间点TCheckI。The management server periodically checks whether the cycle is reached. If the time point TCheckC (the current system time minus the continuous storage cycle) of the management server is within the time range corresponding to the continuous data segment stored in the first storage module, that is, TStartC<TCheckC<TEndC, then the management server will give the storage The device issues a frame dump command, which contains the frame capture time point TCheckC. If the cycle coverage time point TCheckI (system current time minus the cycle coverage period) of the management server is within the corresponding time range of the I frame data segment stored by the second storage module, that is, TStartI<TCheckI<TEndI, then the management server gives The storage device issues a cyclic overwriting delete command, and the command includes a cyclic overwriting time point TCheckI.
本发明的实施例中，存储设备除了在本地记录连续数据段的索引信息和I帧数据段的索引信息外，也可以将写入连续数据生成的连续数据段、抽帧转存时生成的I帧数据段、删除的连续数据段、循环覆盖删除的I帧数据段的索引信息上报给管理服务器。In the embodiment of the present invention, in addition to locally recording the index information of the continuous data segment and the index information of the I frame data segment, the storage device may also write the continuous data segment generated by writing continuous data, and the I frame generated when the frame is dumped. The index information of the frame data segment, the deleted continuous data segment, and the I-frame data segment deleted by cyclic coverage is reported to the management server.
本发明的实施例适用于主要由管理服务器和存储设备组成的视频存储系统。管理服务器保存存储设备上存储数据的相关索引信息，还可以为存储设备设置存储空间和存储周期。并且，管理服务器还可以根据存储周期，给存储设备下发抽帧转存命令和循环覆盖删除命令。存储设备将视频编码器输出的视频数据以段为单位存储在第一存储模块内，收到抽帧转存命令时，进行抽帧转存；收到循环覆盖删除命令时，进行循环覆盖删除。The embodiment of the present invention is applicable to a video storage system mainly composed of a management server and a storage device. The management server saves the relevant index information of the data stored on the storage device, and can also set the storage space and storage period for the storage device. In addition, the management server can also issue a frame dump command and a cycle overwrite delete command to the storage device according to the storage period. The storage device stores the video data output by the video encoder in the first storage module in units of segments, and performs frame dumping when receiving a frame dumping command; and performs looping overwriting deletion when receiving a loop overwriting delete command.
本发明的实施例可以尽可能地延长视频数据的存储周期。The embodiments of the present invention can prolong the storage period of video data as much as possible.
本发明的实施例中，存储设备包含有第一存储模块和第二存储模块，更新视频数据的存储周期包含有连续存储周期和I帧存储周期。连续存储周期和I帧存储周期之和为循环覆盖周期。视频编码器输出的视频数据，分段写入存储设备的第一存储模块，记录段内所有I帧信息，同时将段编号、开始和结束时间等索引信息上报给管理服务器记录下来。In the embodiment of the present invention, the storage device includes a first storage module and a second storage module, and the storage period for updating video data includes a continuous storage period and an I-frame storage period. The sum of the continuous storage period and the I frame storage period is the cycle coverage period. The video data output by the video encoder is segmented into the first storage module of the storage device, and all I frame information in the record segment is reported to the management server to record index information such as segment number, start and end time.
当连续存储周期到达时，管理服务器给存储设备下发抽帧转存命令，存储设备收到后执行抽帧转存，将第一存储模块内在抽帧时间点之前的连续数据段内部分I帧数据解码出灰度缩略图，根据缩略图之间的相似性确定抽帧步长，利用该抽帧步长从连续数据段的I帧序列中去除冗余I帧，写入剩余的有用I帧到该设备第二存储模块内生成I帧数据段，同时删除对应的连续数据段，并上报管理服务器新段索引信息和原段删除索引信息。循环覆盖周期达到时，管理服务器给存储设备下发循环覆盖删除命令，存储设备收到后执行循环覆盖删除，将循环覆盖时间点之前的第二存储模块内I帧数据段删除。When the continuous storage period arrives, the management server sends a frame dump command to the storage device, and the storage device executes the frame dump command after receiving it, and transfers part of the I frames in the continuous data segment before the frame capture time point in the first storage module The data is decoded into a gray-scale thumbnail, and the frame extraction step is determined according to the similarity between the thumbnails, and the redundant I frame is removed from the I frame sequence of the continuous data segment by using the frame extraction step, and the remaining useful I frame is written Generate the I frame data segment in the second storage module of the device, delete the corresponding continuous data segment simultaneously, and report the new segment index information and the original segment deletion index information to the management server. When the cyclic coverage period is reached, the management server sends a cyclic coverage deletion command to the storage device, and the storage device executes the cyclic coverage deletion after receiving it, and deletes the I frame data segment in the second storage module before the cyclic coverage time point.
本发明的实施例，可以根据视频存储系统中存储设备的容量、接入视频编码器的路数、码率以及预期的存储周期来设置第一、第二存储模块以及连续存储周期和I帧存储周期。其中，预设的存储周期可以根据需要进行调整，以适用于不同的使用场合。In the embodiment of the present invention, the first and second storage modules, the continuous storage period and the I frame storage can be set according to the capacity of the storage device in the video storage system, the number of channels connected to the video encoder, the code rate and the expected storage period. cycle. Wherein, the preset storage period can be adjusted according to needs, so as to be suitable for different usage occasions.
具体地，存储设备的存储空间以及存储周期等与存储设备的存储能力之间，具有如下对应关系。Specifically, the storage space and the storage cycle of the storage device have the following corresponding relationship with the storage capability of the storage device.
存储空间S＝[存储周期T1+存储周期T2*I帧比列系数]*(码率/(8*1024*1024)*24*3600)*视频接入数。Storage space S=[storage period T1+storage period T2*I frame ratio coefficient]*(code rate/(8*1024*1024)*24*3600)*video access number.
其中，存储空间S，为系统存储设备总容量(GB)。存储周期T1为连续存储周期(以天为单位)。存储周期T2为I帧存储周期(以天为单位)。码率为视频编码器的输出码率(Kbps)。视频接入数为视频编码器的数量。I帧比例系数为I帧数据容量与连续数据容量之比，其中连续数据包含有I帧数据和非I帧数据。I帧比例系数可调，具体与视频编码I帧间隔相关，抽取全部I帧时，一般可取33％，仅抽取有用I帧，这个值可取得更小。Wherein, the storage space S is the total capacity (GB) of the system storage device. The storage period T1 is a continuous storage period (in days). The storage period T2 is an I frame storage period (in days). The bit rate is the output bit rate (Kbps) of the video encoder. The number of video accesses is the number of video encoders. The I-frame proportional coefficient is the ratio of the I-frame data capacity to the continuous data capacity, wherein the continuous data includes I-frame data and non-I-frame data. The I-frame ratio coefficient is adjustable, and is specifically related to the video coding I-frame interval. When extracting all I-frames, it is generally 33%, and only useful I-frames are extracted, and this value can be smaller.
存储空间S、码率、视频接入路数都是不变量，对用户来说，循环覆盖周期T尽可能大点好，这样数据能保存更久。The storage space S, bit rate, and number of video access channels are all invariant. For users, it is better for the cycle coverage period T to be as large as possible, so that the data can be stored for a longer period of time.
根据上述分析，假设T1＝30天，如果不进行抽帧转存，即T2＝0，则T＝T1＝30天。如果抽全部I帧并转存，若T1＝25天，则T2＝15天，T＝T1+T2＝40天；仅抽有用I帧并转存，则T2>15天，T>40天。由此对比可见，本发明的实施例相比传统的技术方案而言，可以将存储周期延长33％甚至更多。According to the above analysis, it is assumed that T1=30 days, if no frame dumping is performed, that is, T2=0, then T=T1=30 days. If all I frames are drawn and dumped, if T1=25 days, then T2=15 days, T=T1+T2=40 days; only useful I frames are drawn and dumped, then T2>15 days, T>40 days. From this comparison, it can be seen that the embodiment of the present invention can extend the storage period by 33% or more than the traditional technical solution.
本领域的技术人员应该明白，上述的本申请实施例所提供的设备的各组成部分，以及方法中的各步骤，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上。可选地，它们可以用计算装置可执行的程序代码来实现。从而，可以将它们存储在存储装置中由计算装置来执行，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本发明不限制于任何特定的硬件和软件结合。Those skilled in the art should understand that the various components of the equipment provided by the above-mentioned embodiments of the present application, as well as the various steps in the method, they can be concentrated on a single computing device, or distributed in multiple computing devices. online. Alternatively, they may be implemented in program code executable by a computing device. Therefore, they can be stored in a storage device to be executed by a computing device, or they can be fabricated into individual integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. As such, the present invention is not limited to any specific combination of hardware and software.
虽然本发明所揭露的实施方式如上，但所述的内容仅为便于理解本发明技术方案而采用的实施方式，并非用以限定本发明。任何本发明所属领域内的技术人员，在不脱离本发明所揭露的精神和范围的前提下，可以在实施的形式及细节上进行任何的修改与变化，但本发明的专利保护范围，仍须以所附的权利要求书所界定的范围为准。Although the embodiments disclosed in the present invention are as above, the described content is only an embodiment adopted to facilitate understanding of the technical solution of the present invention, and is not intended to limit the present invention. Anyone skilled in the field of the present invention can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed by the present invention, but the patent protection scope of the present invention must still be The scope defined by the appended claims shall prevail.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510501686.7ACN106470323B (en) | 2015-08-14 | 2015-08-14 | The storage method and equipment of video data |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510501686.7ACN106470323B (en) | 2015-08-14 | 2015-08-14 | The storage method and equipment of video data |
| Publication Number | Publication Date |
|---|---|
| CN106470323Atrue CN106470323A (en) | 2017-03-01 |
| CN106470323B CN106470323B (en) | 2019-08-16 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510501686.7AActiveCN106470323B (en) | 2015-08-14 | 2015-08-14 | The storage method and equipment of video data |
| Country | Link |
|---|---|
| CN (1) | CN106470323B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109816110A (en)* | 2019-01-24 | 2019-05-28 | 杭州嘉楠耘智信息科技有限公司 | Scrypt algorithm workload proving method and device |
| CN110300134A (en)* | 2018-03-22 | 2019-10-01 | 杭州海康威视系统技术有限公司 | Memory space method of adjustment, device and the cloud storage system of cloud storage resource pool |
| CN110728699A (en)* | 2019-10-11 | 2020-01-24 | 智慧视通(杭州)科技发展有限公司 | Track post-processing method based on characteristic distance |
| CN111698555A (en)* | 2020-06-19 | 2020-09-22 | 浙江大华技术股份有限公司 | Video frame extraction processing method and device |
| CN111866457A (en)* | 2020-07-14 | 2020-10-30 | 广州市宏视电子技术有限公司 | Monitoring image processing method, electronic device, storage medium and system |
| CN112434185A (en)* | 2020-10-26 | 2021-03-02 | 国家广播电视总局广播电视规划院 | Method, system, server and storage medium for searching similar video clips |
| CN112541390A (en)* | 2020-10-30 | 2021-03-23 | 四川天翼网络服务有限公司 | Frame-extracting dynamic scheduling method and system for violation analysis of examination video |
| CN112995676A (en)* | 2021-04-25 | 2021-06-18 | 北京金和网络股份有限公司 | Audio and video compression method, device, equipment, system and storage medium |
| CN114245167A (en)* | 2021-11-08 | 2022-03-25 | 浙江大华技术股份有限公司 | Video storage method and device and computer readable storage medium |
| CN114430469A (en)* | 2022-04-01 | 2022-05-03 | 浙江大华技术股份有限公司 | Video data storage method, video data reading method, electronic device and readable storage medium |
| CN115618050A (en)* | 2022-12-06 | 2023-01-17 | 苏州浪潮智能科技有限公司 | Video data storage, analysis method, device, system, communication equipment and storage medium |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070061727A1 (en)* | 2005-09-15 | 2007-03-15 | Honeywell International Inc. | Adaptive key frame extraction from video data |
| CN101674444A (en)* | 2009-10-15 | 2010-03-17 | 杭州华三通信技术有限公司 | Storage method of video data, system and encoder thereof |
| CN102761733A (en)* | 2011-04-29 | 2012-10-31 | 富泰华工业（深圳）有限公司 | Monitoring system and video data storage method |
| CN103810711A (en)* | 2014-03-03 | 2014-05-21 | 郑州日兴电子科技有限公司 | Keyframe extracting method and system for monitoring system videos |
| CN104618679A (en)* | 2015-03-13 | 2015-05-13 | 南京知乎信息科技有限公司 | Method for extracting key information frame from monitoring video |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070061727A1 (en)* | 2005-09-15 | 2007-03-15 | Honeywell International Inc. | Adaptive key frame extraction from video data |
| CN101674444A (en)* | 2009-10-15 | 2010-03-17 | 杭州华三通信技术有限公司 | Storage method of video data, system and encoder thereof |
| CN102761733A (en)* | 2011-04-29 | 2012-10-31 | 富泰华工业（深圳）有限公司 | Monitoring system and video data storage method |
| CN103810711A (en)* | 2014-03-03 | 2014-05-21 | 郑州日兴电子科技有限公司 | Keyframe extracting method and system for monitoring system videos |
| CN104618679A (en)* | 2015-03-13 | 2015-05-13 | 南京知乎信息科技有限公司 | Method for extracting key information frame from monitoring video |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110300134B (en)* | 2018-03-22 | 2022-10-04 | 杭州海康威视系统技术有限公司 | Storage space adjusting method and device of cloud storage resource pool and cloud storage system |
| CN110300134A (en)* | 2018-03-22 | 2019-10-01 | 杭州海康威视系统技术有限公司 | Memory space method of adjustment, device and the cloud storage system of cloud storage resource pool |
| CN109816110A (en)* | 2019-01-24 | 2019-05-28 | 杭州嘉楠耘智信息科技有限公司 | Scrypt algorithm workload proving method and device |
| CN110728699A (en)* | 2019-10-11 | 2020-01-24 | 智慧视通(杭州)科技发展有限公司 | Track post-processing method based on characteristic distance |
| CN111698555A (en)* | 2020-06-19 | 2020-09-22 | 浙江大华技术股份有限公司 | Video frame extraction processing method and device |
| CN111866457A (en)* | 2020-07-14 | 2020-10-30 | 广州市宏视电子技术有限公司 | Monitoring image processing method, electronic device, storage medium and system |
| CN112434185A (en)* | 2020-10-26 | 2021-03-02 | 国家广播电视总局广播电视规划院 | Method, system, server and storage medium for searching similar video clips |
| CN112434185B (en)* | 2020-10-26 | 2023-07-14 | 国家广播电视总局广播电视规划院 | Method, system, server and storage medium for searching similar video clips |
| CN112541390A (en)* | 2020-10-30 | 2021-03-23 | 四川天翼网络服务有限公司 | Frame-extracting dynamic scheduling method and system for violation analysis of examination video |
| CN112541390B (en)* | 2020-10-30 | 2023-04-25 | 四川天翼网络股份有限公司 | Frame extraction dynamic scheduling method and system for examination video violation analysis |
| CN112995676A (en)* | 2021-04-25 | 2021-06-18 | 北京金和网络股份有限公司 | Audio and video compression method, device, equipment, system and storage medium |
| CN114245167A (en)* | 2021-11-08 | 2022-03-25 | 浙江大华技术股份有限公司 | Video storage method and device and computer readable storage medium |
| CN114430469A (en)* | 2022-04-01 | 2022-05-03 | 浙江大华技术股份有限公司 | Video data storage method, video data reading method, electronic device and readable storage medium |
| CN115618050A (en)* | 2022-12-06 | 2023-01-17 | 苏州浪潮智能科技有限公司 | Video data storage, analysis method, device, system, communication equipment and storage medium |
| CN115618050B (en)* | 2022-12-06 | 2023-03-21 | 苏州浪潮智能科技有限公司 | Video data storage, analysis method, device, system, communication equipment and storage medium |
| Publication number | Publication date |
|---|---|
| CN106470323B (en) | 2019-08-16 |
| Publication | Publication Date | Title |
|---|---|---|
| CN106470323B (en) | The storage method and equipment of video data | |
| CN110610510B (en) | Target tracking method, device, electronic device and storage medium | |
| US11531482B2 (en) | Data deduplication method and apparatus | |
| CN105430480A (en) | Method and system for storing video data | |
| CN104363403A (en) | Method and device for storing video data in video monitoring system | |
| CN111263211B (en) | Method for caching video data and terminal equipment | |
| KR102147633B1 (en) | Method and apparatus for decoding variable length coded files | |
| US10070134B2 (en) | Analytics assisted encoding | |
| CN104811646B (en) | The storage method of the modulation of multiple video strems Data Concurrent and buffering based on Coutinuous store model | |
| CN102629235A (en) | Method for increasing read-write speed of double data rate (DDR) memory | |
| CN107229516B (en) | A data processing method and device | |
| CN116132719B (en) | Video processing method, device, electronic device and readable storage medium | |
| WO2021051912A1 (en) | Media data transmission method and related device | |
| CN111405358A (en) | Cache-based video frame extraction method, apparatus, medium, and system | |
| CN104580804A (en) | Image compression uploading method and system | |
| CN103891272B (en) | Multiple stream processing for video analysis and encoding | |
| CN104394453A (en) | Video prerecording method and device | |
| US11743474B2 (en) | Shot-change detection using container level information | |
| CN104301648A (en) | Method for avoiding data loss when SD card stores video | |
| CN114650391B (en) | Video processing method, terminal device and computer-readable storage medium | |
| CN114626994A (en) | An image processing method, a video processing method, a computer device, and a storage medium | |
| CN108876703B (en) | Data storage method | |
| CN105138330B (en) | The processing method and processing unit of dygraphs Drawing Objects | |
| CN117176979B (en) | Method, device, equipment and storage medium for extracting content frames of multi-source heterogeneous video | |
| CN103918002B (en) | Memory controller for video analysis and encoding |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | Effective date of registration:20250925 Address after:Rooms 602 and 605, No. 85 Xiangxue Avenue Middle, Huangpu District, Guangzhou City, Guangdong Province 510000 Patentee after:Guangzhou Gaohang Technology Transfer Co.,Ltd. Country or region after:China Address before:Hangzhou City, Zhejiang province 310000 Binjiang District Qianmo Road No. 555 Patentee before:HANGZHOU HIKVISION SYSTEM TECHNOLOGY Co.,Ltd. Country or region before:China |