CN112950400A - Data processing platform - Google Patents

Abstract

The application provides a data processing platform, and belongs to the technical field of computer application. The data processing platform comprises: the data acquisition system is used for acquiring monitoring data in the production process and sending the monitoring data to the edge computing system; the edge computing system is used for acquiring monitoring data, forwarding the monitoring data to the cloud computing system through the network transmission system, analyzing the monitoring data according to a preset monitoring algorithm, and generating a control strategy to control the production process; the network transmission system is used for transmitting the monitoring data to the cloud computing system and transmitting a preset monitoring algorithm generated by the cloud computing system to the edge computing system; and the cloud computing system is used for acquiring the monitoring data forwarded by the edge computing system, and performing big data analysis and model training on the monitoring data to generate a preset monitoring algorithm and send the preset monitoring algorithm to the edge computing system. Therefore, through the data processing platform, the automation and the intelligence degree of coal mine production are improved, and the coal mine production safety is ensured.

Description

Data processing platform

Technical Field

The application relates to the technical field of computer application, in particular to a data processing platform.

Background

With the development of new-generation information technology, industrial internet becomes an important enabling tool for economic and digital transformation of domestic and foreign entities, and the core idea is to realize comprehensive data perception, dynamic transmission, real-time analysis, scientific decision and intelligent control in the production and operation process of enterprises by combining emerging information technology with traditional manufacturing industry and taking an information physical system as a core technology system, and promote the upgrading and transformation of the traditional industry.

In the coal mining industry, the underground production process is monitored in real time, and the underground production process is an important guarantee for safe production. Therefore, how to realize real-time monitoring and automatic control of coal mining through an industrial internet technology to improve the automation and the intelligent degree of coal mining and provide service for safety production of coal mines is a problem to be solved urgently.

Disclosure of Invention

The data processing platform that this application provided for in solving the correlation technique, how to realize coal mining's real-time supervision and automatic control through industry internet technology to promote coal mining's automation and intelligent degree, provide the problem of service for the safety in production in colliery.

The data processing platform that this application embodiment provided includes: the system comprises a data acquisition system, a video data compression system, an edge computing system, a network transmission system and a cloud computing system; the data acquisition system comprises image acquisition equipment, voice acquisition equipment, a sensor, positioning equipment and equipment operation parameter detection equipment, and is used for acquiring monitoring data in the production process and sending the monitoring data to the edge computing system; the edge computing system comprises a plurality of edge computing nodes and is used for acquiring monitoring data acquired by the data acquisition system, forwarding the monitoring data to the cloud computing system through the network transmission system, and carrying out data analysis on the monitoring data according to a preset monitoring algorithm so as to generate a control strategy to control the production process; the network transmission system comprises at least one of the following network types: the wireless fidelity network, the wired Ethernet, the virtual private network, the 4G, the 5G and the 6G are used for sending the monitoring data forwarded by the edge computing system to the cloud computing system and sending the preset monitoring algorithm generated by the cloud computing system to the edge computing system; the cloud computing system is configured to obtain the monitoring data forwarded by the edge computing system, perform big data analysis and model training on the monitoring data, generate the preset monitoring algorithm, and send the preset monitoring algorithm to the edge computing system through the network transmission system.

Optionally, in a possible implementation manner of the embodiment of the present application, the monitoring data includes video data acquired by the image acquisition device, and the data processing platform further includes:

the video compression system is used for compressing the video data;

the edge computing system is further configured to obtain the compressed video data and forward the compressed video data to the cloud computing system.

Optionally, in another possible implementation manner of the embodiment of the present application, the cloud computing system is further configured to:

determining the type of each piece of monitoring data according to the data acquisition equipment corresponding to each piece of monitoring data;

determining a data annotation model corresponding to each piece of monitoring data according to the type of each piece of monitoring data;

inputting each piece of monitoring data into a corresponding data labeling model respectively to generate labeling data corresponding to each piece of monitoring data;

sequentially inputting each piece of monitoring data and the corresponding labeled data into an initial deep learning model so as to train the initial deep learning model and generate a deep learning model after training;

and issuing the trained deep learning model serving as the preset monitoring algorithm to the edge computing system.

Optionally, in another possible implementation manner of the embodiment of the present application, the monitoring data includes video data acquired by the image acquisition device, the data annotation model includes an image annotation model, and the cloud computing system is further configured to:

performing image frame extraction on the video data to determine each image frame included in the video data;

sequentially inputting each image frame into the image annotation model to determine image annotation data corresponding to each image frame, wherein the image annotation data comprises at least one of the following data: image type, abnormal behavior framing data and pixel labeling data.

Optionally, in another possible implementation manner of the embodiment of the present application, the cloud computing system is further configured to:

acquiring a development task establishing instruction, wherein the development task establishing instruction comprises at least one of a target development task name, a target development task function description, a target developer permission and a type of a target initial deep learning model;

according to the development task establishing instruction, determining target monitoring data and a target initial deep learning model corresponding to the development task;

inputting the target monitoring data into a target initial deep learning model to train the target initial deep learning model and generate a trained target deep learning model;

and issuing the target deep learning model to the edge computing system as a preset monitoring algorithm.

Optionally, in another possible implementation manner of the embodiment of the present application, the cloud computing system is further configured to:

acquiring an issued task establishing instruction, wherein the issued task establishing instruction comprises a target algorithm identifier and a target edge node identifier;

acquiring a target algorithm according to the target algorithm identification;

and issuing the target algorithm to the target edge node according to the target edge node identifier.

Optionally, in another possible implementation manner of the embodiment of the present application, the edge computing system is further configured to:

acquiring network parameters among the edge nodes, wherein the network parameters comprise static network parameters and dynamic network parameters;

acquiring a calculation effectiveness index of each edge node, wherein the calculation effectiveness index comprises a hardware performance index and a real-time calculation rate;

and determining the distribution rule of each preset monitoring algorithm in each edge node according to the network parameters among the edge nodes and the calculation effectiveness indexes of the edge nodes.

The data processing platform provided by the embodiment of the application acquires monitoring data in a production process through the data acquisition system and sends the monitoring data to the edge computing system, the edge computing system performs data analysis on the monitoring data acquired by the data acquisition system and generates a control strategy to control the production process in real time, the cloud computing system performs big data analysis and model training on the monitoring data forwarded by the edge computing system, generates a monitoring algorithm and sends the monitoring algorithm to the edge computing system, and therefore cooperative control of the cloud computing system and the edge computing system is achieved. Therefore, various factors such as production equipment, environment and personnel in the production environment are comprehensively considered, the real-time control is carried out on the production process through the edge computing system, the big data analysis is carried out on long-term monitoring data through the cloud computing system, an accurate monitoring algorithm is formulated, the real-time monitoring and control on the coal mine production process are effectively realized through the cooperative control of the edge computing system and the cloud computing system, the automation and the intelligent degree of coal mine production are improved, and the coal mine production safety is ensured.

Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.

Drawings

The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a schematic structural diagram of a data processing platform according to an embodiment of the present disclosure;

fig. 2 is a schematic structural diagram of another data processing platform according to an embodiment of the present disclosure;

fig. 3 is a schematic structural diagram of another data processing platform according to an embodiment of the present disclosure.

Detailed Description

Reference will now be made in detail to the embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the like or similar elements throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.

The embodiment of the application provides a data processing platform aiming at the problems that in the related technology, the real-time monitoring and automatic control of coal mining are realized through an industrial internet technology, so that the automation and the intelligent degree of coal mining are improved, and the service is provided for the safety production of coal mines.

The data processing platform provided by the embodiment of the application acquires monitoring data in a production process through the data acquisition system and sends the monitoring data to the edge computing system, the edge computing system performs data analysis on the monitoring data acquired by the data acquisition system and generates a control strategy to control the production process in real time, the cloud computing system performs big data analysis and model training on the monitoring data forwarded by the edge computing system, generates a monitoring algorithm and sends the monitoring algorithm to the edge computing system, and therefore cooperative control of the cloud computing system and the edge computing system is achieved. Therefore, various factors such as production equipment, environment and personnel in the production environment are comprehensively considered, the real-time control is carried out on the production process through the edge computing system, the big data analysis is carried out on long-term monitoring data through the cloud computing system, an accurate monitoring algorithm is formulated, the real-time monitoring and control on the coal mine production process are effectively realized through the cooperative control of the edge computing system and the cloud computing system, the automation and the intelligent degree of coal mine production are improved, and the coal mine production safety is ensured.

The data processing platform provided by the present application is described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic flowchart of a data processing platform according to an embodiment of the present disclosure.

As shown in fig. 1, the data processing platform 10 includes: data acquisition system 110, edge computing system 120, network transmission system 130, and cloud computing system 140.

The data acquisition system 110 may include an image acquisition device, a voice acquisition device, a sensor, a positioning device, and a device operation parameter detection device, and is configured to acquire monitoring data in a production process and send the monitoring data to the edge computing system 120. As shown in fig. 2, based on the embodiment shown in fig. 1, the data acquisition system 110 may include an image acquisition device 111, a voice acquisition device 112, a sensor 113, a positioning device 114, and a device operation parameter detection device 115;

the edge computing system 120, as shown in fig. 2, may include a plurality ofedge computing nodes 121, configured to obtain monitoring data acquired by the data acquisition system 110, forward the monitoring data to the cloud computing system 140 through the network transmission system 130, and perform data analysis on the monitoring data according to a preset monitoring algorithm to generate a control policy to control the production process;

the network transmission system 130 may include at least one of the following network types: a Wireless Fidelity (Wi-Fi) Network, a wired ethernet, a Virtual Private Network (VPN), 4G, 5G, and 6G, configured to send monitoring data forwarded by the edge computing system 120 to the cloud computing system 140, and send a preset monitoring algorithm generated by the cloud computing system 140 to the edge computing system 120;

the cloud computing system 140 is configured to obtain the monitoring data forwarded by the edge computing system 120, perform big data analysis and model training on the monitoring data to generate a preset monitoring algorithm, and send the preset monitoring algorithm to the edge computing system 120 through the network transmission system 130.

It should be noted that the data processing platform 10 according to the embodiment of the present application may be applied to a coal mine production environment to monitor real-time data of production equipment, production personnel, and the production environment in the coal mine production environment, so as to ensure safety and efficiency of coal mine production.

As a possible implementation manner, the image capturing device 111 may be any monitoring device, such as a monocular camera, a binocular camera, an RGB-D camera, a laser radar, an infrared camera, and the like, which is not limited in this embodiment of the present application. The image acquisition device 111 can be arranged at any position in the production environment as required, and is used for acquiring real-time video data in the production environment; accordingly, the monitoring data collected by the data collection system 110 may include video data.

The voice collecting device 112 may be a voice collecting device disposed in or near the production device to collect noise generated during operation of the production device, so that the operation condition of the production device can be monitored by sound generated during operation of the production device; alternatively, the voice collecting device 112 may also be a communication device such as an intercom used by the production staff to collect the voice data for communication between the production staff, and further monitor the production condition through the voice data for communication between the production staff. Accordingly, the monitoring data collected by the data collection system 110 may include voice data.

The sensor 113 may include various sensors for monitoring production environment data, and may include at least one of a temperature sensor, a humidity sensor, a gas concentration sensor, and a pressure sensor, so as to monitor environmental data such as temperature, humidity, gas concentration (such as gas concentration), and mine pressure in the production environment. Accordingly, when the sensor 113 includes at least one of a temperature sensor, a humidity sensor, a gas concentration sensor, and a pressure sensor, the monitoring data collected by the data collection system 110 may include at least one of an ambient temperature, an ambient humidity, an ambient gas concentration, and a mine pressure.

A locating device 114, which may comprise at least one of a person locating device, a device locating device; accordingly, the monitoring data collected by the data collection system 110 may include at least one of personnel location information and equipment location information. For example, in a coal mine production scene, positioning devices such as position sensors can be arranged in movable production devices such as underground production personnel and underground coal transportation vehicles, so as to obtain real-time position information of the underground production personnel and the movable production devices. It should be noted that, in a coal mine production scene, due to the special coal mine production environment, a position sensor or a positioning service dedicated for underground use can be used for acquiring position information, so as to ensure the accuracy of the position information acquisition.

The device operation parameter detection device 115 may be disposed in each production device to collect real-time operation parameters of the production device during the production process. The device operation parameter detection device 115 may be a component that is provided in the production device itself and can record its own operation parameter, or may be a component that is separately provided in each production device according to actual monitoring needs. For example, for a coal mining machine with an automation function, the coal mining machine itself may include sensors for measuring parameters such as its working temperature, working humidity, working pressure, and inclination angle during working, and these sensors may be used as the device operation parameter detection device 115.

Further, the data transmission pressure of the data processing platform 10 is too large because the video data collected by the image collecting device 111 usually has a huge data volume. Therefore, before transmitting the video data captured by the image capturing device 111, the video data may also be compressed first to reduce the data transmission pressure. That is, in a possible implementation manner of the embodiment of the present application, as shown in fig. 3, the data processing platform 10 may further include:

a video compression system 150 for compressing video data;

correspondingly, the edge computing system 120 is further configured to obtain the compressed video data and forward the compressed video data to the cloud computing system 140.

In the embodiment of the present application, a Software Development Kit (SDK) corresponding to a video compression algorithm may be included in the video compression system, and is used to compress video data. Therefore, after the image capturing device 111 acquires the video data, the video data may be sent to the video compression system 150, and the video compression system 150 efficiently compresses the video data according to the encoding characteristics of the video data by using a video compression algorithm, so as to reduce the data amount of the video data. Further, the compressed video data may be sent to the edge computing system 120, and then the edge computing system 120 may forward the obtained compressed video data to the cloud computing system 140 through the network transmission system 130.

In the embodiment of the present application, the network transmission system 130 may be implemented by using at least one of Wi-Fi, wired ethernet, VPN, 4G, 5G, and 6G, so as to form a hybrid network transmission system including multiple network types, thereby improving reliability of data transmission between systems in the data processing platform 10.

It should be noted that the implementation manner of the network transmission system 130 may include, but is not limited to, the above-listed situations. In actual use, a proper implementation mode can be selected according to an actual production scene so as to ensure the reliability of data transmission among the systems.

In this embodiment, the edge computing system 120 may send the monitoring data acquired from the data acquisition system 110 to the cloud computing system 140 through the network transmission system 130, and the computing system 140 may integrate all the monitoring data acquired within a preset time period and perform big data analysis, or perform model training by using the integrated data to generate a preset monitoring algorithm and send the preset monitoring algorithm to the edge computing system 130.

It should be noted that the preset time period may be a longer time period, such as a week, a month, a year, and the like, which is not limited in the embodiment of the present application.

As a possible implementation manner, the preset monitoring algorithm may include an image classification algorithm, an object detection algorithm, a semantic segmentation algorithm, a three-dimensional reconstruction algorithm, a positioning and mapping algorithm, an equipment operation parameter detection algorithm, and the like, so that the edge computing system 120 may process the monitoring data acquired by the data acquisition system in real time by using a corresponding algorithm, so as to implement real-time monitoring and control of the production process. And when it is determined by the monitoring algorithm that the warning is required or the production process needs to be controlled, a control strategy can be sent to an execution device (not shown in the figure) so that the execution device can execute the control strategy, thereby realizing the real-time control of the production process.

It should be noted that the execution device may be a production device, or may be a device dedicated to execute a control policy to control the production device; or, the alarm device may also be an alarm device, such as a buzzer, a voice prompt, etc., for performing alarm prompt.

For example, when the execution device acquires the control instruction issued by the edge computing system 120 and the control instruction is "close the production device a", the execution device may be the production device a itself, and the production device a may automatically stop operating when the control instruction is acquired. For another example, if the execution device acquires the alarm instruction issued by the edge computing system 120, the execution device may be a voice prompter, and the voice prompter may issue a voice prompt according to the alarm instruction when acquiring the alarm instruction.

Further, the edge computing system 110 may include a plurality ofedge nodes 121, and theedge nodes 121 may be configured to execute different monitoring algorithms to balance data processing pressures of theedge nodes 121, so as to improve the processing efficiency of the edge computing system 110 and improve the reliability and real-time performance of real-time control. That is, in a possible implementation manner of the embodiment of the present application, the edge computing system 110 may be further configured to:

acquiring network parameters among all edge nodes, wherein the network parameters comprise static network parameters and dynamic network parameters;

acquiring a calculation effectiveness index of each edge node, wherein the calculation effectiveness index comprises a hardware performance index and an actual calculation rate;

and determining the distribution rule of each preset monitoring algorithm in each edge node according to the network parameters among the edge nodes and the calculation effectiveness indexes of the edge nodes.

As a possible implementation manner, any oneedge node 121 in the edge computing system 120 may be configured with a non-edge cloud, and then the edge cloud adjusts an allocation rule of each preset monitoring algorithm in each edge node.

The static network parameters may include communication reliability corresponding to each network type. The dynamic network parameters may include parameters such as real-time transmission rate and real-time transmission quality corresponding to each network type.

The hardware performance index may include a processor type corresponding to each edge node; the real-time computation rate may include the current processor occupancy.

Optionally, the edge cloud may determine a target edge node to which a new preset monitoring algorithm may be installed when acquiring the new preset monitoring algorithm issued by the cloud computing system 140.

Specifically, after the edge computing system 120 is established, parameters such as network types between theedge nodes 121 in the edge computing system 120 and communication reliability corresponding to each network type may be recorded into an edge cloud as static network parameters, and hardware performance indexes capable of identifying the computing performance of theedge nodes 121, such as processor types of theedge nodes 121, may be recorded into the edge cloud; therefore, when a new preset monitoring algorithm needs to be allocated to theedge nodes 121, the edge cloud may obtain locally stored static network parameters between theedge nodes 121 and hardware performance indexes of theedge nodes 121; in addition, the edge cloud may obtain the dynamic network parameters between theedge nodes 121 and the real-time computing rate of eachedge node 121 through a heartbeat program. Then, the edge cloud can construct a target function and a constraint condition according to the obtained network parameters among the edge nodes and the calculation effectiveness indexes of the edge nodes, determine a target edge node which can enable the current processing efficiency and reliability of the system to be optimal through a branch-and-bound method, and issue a new preset monitoring algorithm to the target edge node.

For example, in determining the target edge node, the edge cloud may consider both computational efficiency and reliability factors. For the calculation efficiency, the edge cloud may determine, according to the processor model (CPU, GPU, etc.), processor occupancy rate, real-time network transmission rate, and other indexes of eachedge node 121, and according to the objective function and the constraint condition, theedge node 121 having a relatively high processor performance, a relatively low processor occupancy rate, and a relatively high real-time playing transmission rate as the target edge node. For reliability, the edge cloud may determine, according to the network reliability between eachedge node 121, anedge node 121 with higher network reliability as a target edge node according to an objective function and a constraint condition. For example, in theedge node 121 with the network type of wired ethernet, Wi-Fi, 4G, theedge node 121 with the network type of wired ethernet may be determined as the target edge node.

Optionally, the edge cloud may also uninstall the monitoring algorithm that cannot be run by theedge node 121 toother edge nodes 121 to run when the monitoring algorithm that is installed by theedge node 121 cannot be run normally, so that normal running of the edge computing system 120 is affected. In this case, the manner of determining the target edge node by the edge cloud is the same as the foregoing process, and details are not described here. After the edge cloud determines the target edge node, the monitoring algorithm that cannot be run by theedge node 121 may be copied to the target edge node, and the target edge node runs the monitoring algorithm.

It should be noted that, in actual use, the number of target edge nodes may be 1, or may be multiple, which is not limited in this embodiment of the present application.

As a possible implementation, the cloud computing system 140 may perform model training on the monitoring data collected by the data collection system 110 to generate a monitoring algorithm and send the monitoring algorithm to the edge computing system. That is, in a possible implementation manner of the embodiment of the present application, the cloud computing system 140 may be further configured to:

determining the type of each piece of monitoring data according to the data acquisition equipment corresponding to each piece of monitoring data;

determining a data annotation model corresponding to each piece of monitoring data according to the type of each piece of monitoring data;

inputting each piece of monitoring data into a corresponding data labeling model respectively to generate labeling data corresponding to each piece of monitoring data;

sequentially inputting each piece of monitoring data and the corresponding labeled data into the initial deep learning model to train the initial deep learning model and generate a deep learning model after training;

and issuing the trained deep learning model serving as a preset monitoring algorithm to the edge computing system.

In this embodiment, since the monitoring data collected by the data collection system 110 may be collected by a plurality of data collection devices, the monitoring data collected by different data collection devices may be used to train different monitoring algorithms respectively. For example, the video data may be used to train an image classification algorithm, an image segmentation algorithm, an abnormal behavior recognition algorithm, and the like; the audio data may be used to train a semantic segmentation algorithm, etc.; the equipment operation parameter data can be used for training an equipment operation state recognition algorithm and the like.

Therefore, in this embodiment of the application, the cloud platform computing system 140 may determine a type of each piece of monitoring data for the data acquisition device corresponding to each piece of monitoring data (for example, the data acquisition device corresponding to the monitoring data is an image acquisition device, the type of the monitoring data may be an image or a video, the data acquisition device corresponding to the monitoring data is a voice acquisition device, the type of the monitoring data may be an audio), then may determine a data annotation model matching each type of monitoring data according to the type of each piece of monitoring data, and further annotate each type of monitoring data respectively by using the corresponding data annotation model to generate annotation data corresponding to each piece of monitoring data.

Then, each piece of monitoring data and the corresponding labeled data thereof may be sequentially input into the corresponding initial deep learning model to train the initial deep learning model, and the trained deep learning model is further issued to the edge computing system 120 as a preset monitoring algorithm, so as to realize cooperative control of the cloud computing system 140 and the edge computing system 120.

As an example, video data is used as a type of important monitoring data, and various labels can be performed on the video data to train a monitoring algorithm with various functions. That is, in a possible implementation manner of the embodiment of the present application, the monitoring data may include video data acquired by the image acquisition device 111, and the data annotation model may include an image annotation model; accordingly, the cloud computing system 140 may be further configured to:

performing image frame extraction on the video data to determine each image frame included in the video data;

sequentially inputting each image frame into an image annotation model to determine image annotation data corresponding to each image frame, wherein the image annotation data comprises at least one of the following data: image type, abnormal behavior framing data and pixel labeling data.

In this embodiment of the application, when the monitoring data includes video data, the cloud computing system 140 may perform image frame extraction on the video data to obtain each image frame included in the video data, and then perform online image labeling through an image labeling model, for example, identify a category of the image frame to label the category of the image; identifying abnormal behaviors contained in the image frame, and performing rectangular frame selection on an image area corresponding to the abnormal behaviors to generate abnormal behavior frame selection data; pixel-by-pixel labeling of image frames to generate pixel labeling data, and so on.

Further, the cloud computing system 140 may further include a development management program, so that a user may establish a development task according to actual needs and generate a corresponding monitoring algorithm. That is, in a possible implementation manner of the embodiment of the present application, the cloud computing system 140 may be further configured to:

acquiring a development task establishing instruction, wherein the development task establishing instruction comprises at least one of a target development task name, a target development task function description, a target developer permission and a type of a target initial deep learning model;

according to a development task establishing instruction, determining target monitoring data and a target initial deep learning model corresponding to a development task;

inputting target monitoring data into a target initial deep learning model to train the target initial deep learning model and generate a trained target deep learning model;

the target deep learning model is issued to the edge computing system 120 as a preset monitoring algorithm.

The target development task name, the target development task function description, the target developer permission and the type of the target initial deep learning model are respectively the task name, the task function description, the developer permission and the type of the initial deep learning model which are set by a user with the development management permission through a development management program.

If necessary, the type of the initial deep learning model may be a development framework corresponding to the initial deep learning model, such as PyTorch, TensorFlow, and the like.

In this embodiment, the cloud computing platform 140 may obtain a development task establishment instruction set by a user having a development management authority through a development management program, determine target monitoring data corresponding to a development task according to a target development task name, a target development task function description, and the like included in the development task establishment instruction, input the target monitoring data into the target initial deep learning model to train the target initial deep learning model, and send the trained target deep learning model to the edge computing system 120 as a preset monitoring algorithm.

For example, if the name of the target development task is "image classification", it may be determined that the target monitoring data is video data or image data acquired by the image acquisition device 111, and then the target monitoring data may be input into the target initial deep learning model, so as to generate a trained target image classification model, and send the trained target image classification model to the edge computing system 120.

Further, the cloud computing system 140 may also establish a distribution task through a development management program, so as to distribute the generated monitoring algorithm to thecorresponding edge node 121 according to actual requirements. That is, in a possible implementation manner of the embodiment of the present application, the cloud computing system 140 may be further configured to:

acquiring an issued task establishing instruction, wherein the issued task establishing instruction comprises a target algorithm identifier and a target edge node identifier;

acquiring a target algorithm according to the target algorithm identification;

and issuing the target algorithm to the target edge node according to the target edge node identifier.

The target algorithm identifier refers to an identifier corresponding to a target algorithm that needs to be currently issued to a target edge node, and may be an algorithm name, a code, and the like of the target algorithm, which is not limited in this application.

The target edge node identifier refers to an identifier corresponding to an edge node to which a target algorithm needs to be installed at present, and may be information that can uniquely determine the edge node, such as an IP address and a hardware address of the target edge node.

As a possible implementation manner, the cloud computing system 140 may obtain, through the development management program, an issued task establishment instruction set by a user having a development right, determine a target algorithm identifier and a target edge node identifier included in the issued task establishment instruction, further obtain a target algorithm from the local according to the target algorithm identifier, and issue the target algorithm to the target edge node according to the target edge node identifier through the network transmission system 130.

Optionally, when the edge computing system 120 includes an edge cloud, the cloud computing system 140 may further issue the target algorithm to the edge cloud, and then the edge cloud issues the target algorithm to the target edge node according to the target edge node identifier.

It should be noted that the target edge node may be specified by a user with a development right, or may be determined by the cloud computing system 140 by integrating the computation rate and reliability of eachedge node 121 in the edge computing system 120. In actual use, the manner in which the cloud computing system 140 determines the target edge node may be the same as the manner in which the edge computing system 120 determines the target edge node, and is not described herein again.

The data processing platform provided by the embodiment of the application acquires monitoring data in a production process through the data acquisition system and sends the monitoring data to the edge computing system, the edge computing system performs data analysis on the monitoring data acquired by the data acquisition system and generates a control strategy to control the production process in real time, the cloud computing system performs big data analysis and model training on the monitoring data forwarded by the edge computing system, generates a monitoring algorithm and sends the monitoring algorithm to the edge computing system, and therefore cooperative control of the cloud computing system and the edge computing system is achieved. Therefore, various factors such as production equipment, environment and personnel in the production environment are comprehensively considered, the real-time control is carried out on the production process through the edge computing system, the big data analysis is carried out on long-term monitoring data through the cloud computing system, an accurate monitoring algorithm is formulated, the real-time monitoring and control on the coal mine production process are effectively realized through the cooperative control of the edge computing system and the cloud computing system, the automation and the intelligent degree of coal mine production are improved, and the coal mine production safety is ensured.

Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.

It will be understood that the present application is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.

Claims (7)

1. A data processing platform, comprising: the system comprises a data acquisition system, a video data compression system, an edge computing system, a network transmission system and a cloud computing system;

the data acquisition system comprises image acquisition equipment, voice acquisition equipment, a sensor, positioning equipment and equipment operation parameter detection equipment, and is used for acquiring monitoring data in the production process and sending the monitoring data to the edge computing system;

the edge computing system comprises a plurality of edge computing nodes and is used for acquiring monitoring data acquired by the data acquisition system, forwarding the monitoring data to the cloud computing system through the network transmission system, and carrying out data analysis on the monitoring data according to a preset monitoring algorithm so as to generate a control strategy to control the production process;

the network transmission system comprises at least one of the following network types: the wireless fidelity network, the wired Ethernet, the virtual private network, the 4G, the 5G and the 6G are used for sending the monitoring data forwarded by the edge computing system to the cloud computing system and sending the preset monitoring algorithm generated by the cloud computing system to the edge computing system;

the cloud computing system is configured to obtain the monitoring data forwarded by the edge computing system, perform big data analysis and model training on the monitoring data, generate the preset monitoring algorithm, and send the preset monitoring algorithm to the edge computing system through the network transmission system.

2. The data processing platform of claim 1, wherein the monitoring data comprises video data captured by the image capture device, the data processing platform further comprising:

the video compression system is used for compressing the video data;

the edge computing system is further configured to obtain the compressed video data and forward the compressed video data to the cloud computing system.

3. The data processing platform of claim 1, wherein the cloud computing system is further to:

determining the type of each piece of monitoring data according to the data acquisition equipment corresponding to each piece of monitoring data;

determining a data annotation model corresponding to each piece of monitoring data according to the type of each piece of monitoring data;

inputting each piece of monitoring data into a corresponding data labeling model respectively to generate labeling data corresponding to each piece of monitoring data;

sequentially inputting each piece of monitoring data and the corresponding labeled data into an initial deep learning model so as to train the initial deep learning model and generate a deep learning model after training;

and issuing the trained deep learning model serving as the preset monitoring algorithm to the edge computing system.

4. The data processing platform of claim 3, wherein the monitoring data comprises video data collected by the image capture device, the data annotation model comprises an image annotation model, and the cloud computing system is further configured to:

performing image frame extraction on the video data to determine each image frame included in the video data;

sequentially inputting each image frame into the image annotation model to determine image annotation data corresponding to each image frame, wherein the image annotation data comprises at least one of the following data: image type, abnormal behavior framing data and pixel labeling data.

5. The data processing platform of claim 1, wherein the cloud computing system is further to:

acquiring a development task establishing instruction, wherein the development task establishing instruction comprises at least one of a target development task name, a target development task function description, a target developer permission and a type of a target initial deep learning model;

according to the development task establishing instruction, determining target monitoring data and a target initial deep learning model corresponding to the development task;

inputting the target monitoring data into a target initial deep learning model to train the target initial deep learning model and generate a trained target deep learning model;

and issuing the target deep learning model to the edge computing system as a preset monitoring algorithm.

6. The data processing platform of claim 5, wherein the cloud computing system is further to:

acquiring an issued task establishing instruction, wherein the issued task establishing instruction comprises a target algorithm identifier and a target edge node identifier;

acquiring a target algorithm according to the target algorithm identification;

and issuing the target algorithm to the target edge node according to the target edge node identifier.

7. The data processing platform of any of claims 1-6, wherein the edge computing system is further to:

acquiring network parameters among the edge nodes, wherein the network parameters comprise static network parameters and dynamic network parameters;

acquiring a calculation effectiveness index of each edge node, wherein the calculation effectiveness index comprises a hardware performance index and a real-time calculation rate;

and determining the distribution rule of each preset monitoring algorithm in each edge node according to the network parameters among the edge nodes and the calculation effectiveness indexes of the edge nodes.

Images