技术领域technical field
本发明属于海洋装备技术领域，更具体地说，涉及海洋浮游生物光学成像 装置及成像处理方法。The invention belongs to the technical field of marine equipment, and more particularly, relates to an optical imaging device for marine plankton and an imaging processing method.
背景技术Background technique
浮游生物是海洋生态系统中非常重要的类群,由于浮游生物的复杂性,当前 海洋浮游生物观测研究面临的一个瓶颈问题是难以在较大的时空尺度上快速测 量其数量与种类组成的变化并完成其细结构的观测。Plankton is a very important group in the marine ecosystem. Due to the complexity of plankton, a bottleneck problem faced by the current marine plankton observation and research is that it is difficult to quickly measure and complete the changes in the number and species composition on a large temporal and spatial scale. observation of its fine structure.
传统的利用浮游生物网具采样的方式仍然是当前浮游生物采样技术的核心, 也是许多长时间序列研究和海洋研究计划的基础。随着科技发展，现有技术中 的取样设备上也做了一些更新，比如多联网通过压力感应在同一剖面中采集不 同水层的样品。虽然新的技术在一定程度上取代了一些传统采样技术，但其所 面临的一个挑战是只能在相对较低的时空精度采集样品,且样品分析的周期很 长，需要人为进行室内的镜检，由于采样及储存条件的影响因素，比如样品由 于水流冲刷导致个体易碎，或者储存时福尔马林长期保存的个体易分解等，往 往会导致采样结果与实际海区的分布情况认知有一定的差异。现在科研人员已 经普遍认识到许多与浮游生物有关的重要生态学过程发生在更为精细的时空尺度上,而使用传统的采样方法难以达到要求。Traditional methods of sampling with plankton nets remain at the heart of current plankton sampling techniques and the basis for many long-term studies and marine research programs. With the development of science and technology, some updates have also been made on the sampling equipment in the existing technology, for example, multiple networks can collect samples of different water layers in the same section through pressure sensing. Although the new technology has replaced some traditional sampling technologies to a certain extent, one of the challenges it faces is that samples can only be collected with relatively low spatial and temporal precision, and the sample analysis cycle is very long, requiring manual indoor microscopy. , Due to the influencing factors of sampling and storage conditions, such as the samples are brittle due to water scouring, or the individuals stored in formalin for a long time are easily decomposed during storage, etc., the sampling results and the actual distribution of the sea area are often cognition to a certain extent. difference. Now researchers have generally recognized that many important ecological processes related to plankton occur at finer spatiotemporal scales, which are difficult to achieve using traditional sampling methods.
另外利用网具采样经常会扰乱浮游生物的分布特征(如斑块分布、薄层分 布等),这些特征对于更好地了解浮游生物群落结构和分布非常重要,只有通过 现场原位观测系统才能保持，而对于生态学上重要但易受损伤的类群来说,如胶 质类浮游动物,使用传统网具采集也容易低估它们的数量。In addition, sampling with nets often disturbs the distribution characteristics of plankton (such as patch distribution, thin layer distribution, etc.), which are very important for a better understanding of plankton community structure and distribution and can only be maintained by in situ observation systems. , while for ecologically important but vulnerable taxa, such as glial zooplankton, collection using traditional nets tends to underestimate their numbers.
发明内容SUMMARY OF THE INVENTION
本发明的目的在于提供一种海洋浮游生物光学成像装置及成像处理方法， 利用机器视觉技术实现对海底浮游生物种群和颗粒物质的快速识别及量化，并 且避免了对海底浮游生物种群和颗粒物质自然维度及水文形态形式的改变，从 而可以实现对浮游生物种群行为较为准确的推测。The purpose of the present invention is to provide a marine plankton optical imaging device and an imaging processing method, which utilizes machine vision technology to realize the rapid identification and quantification of seabed plankton populations and particulate matter, and avoids the natural occurrence of seafloor plankton populations and particulate matter. Changes in dimensions and hydrological forms, so that more accurate predictions of plankton population behavior can be achieved.
实现本发明所采用的技术方案是：一种海洋浮游生物光学成像装置，包括 主架体、闪光灯组件、相机组件和主控制单元，闪光灯组件和相机组件均设于 主架体上，所述闪光灯组件包括旋转驱动装置、旋转叶轮和闪光灯，所述旋转 叶轮通过所述旋转驱动装置驱动旋转，在所述旋转叶轮上沿着圆周方向均布有 多个闪光灯，所述相机组件包括直线驱动装置、光学成像仪和相机镜头基座， 所述相机镜头基座朝向所述闪光灯组件的一侧设有透镜，在所述相机镜头基座 内设有光学成像仪，且所述光学成像仪通过所述直线驱动装置驱动移动以使合 适的镜头与所述透镜对位，所述旋转驱动装置、闪光灯、直线驱动装置和光学 成像仪均通过所述主控制单元控制。The technical scheme adopted for realizing the present invention is as follows: an optical imaging device for marine plankton, comprising a main frame body, a flash unit, a camera unit and a main control unit, the flash unit and the camera unit are both arranged on the main frame, and the flash unit The assembly includes a rotary driving device, a rotary impeller and a flash, the rotary impeller is driven to rotate by the rotary driving device, and a plurality of flashes are evenly distributed along the circumferential direction on the rotary impeller, and the camera assembly includes a linear driving device, An optical imager and a camera lens base, the camera lens base is provided with a lens on the side facing the flash assembly, an optical imager is arranged in the camera lens base, and the optical imager passes through the A linear drive drives movement to align a suitable lens with the lens, and the rotary drive, flash, linear drive and optical imager are all controlled by the main control unit.
所述主架体内设有闪光灯单元和相机驱动单元，所述闪光灯单元和相机驱 动单元平行设置且后端均与主控制单元前端相连，所述闪光灯组件设于闪光灯 单元前端，所述相机镜头基座设于相机驱动单元前端，且所述相机驱动单元内 设有所述直线驱动装置。The main frame is provided with a flash unit and a camera drive unit, the flash unit and the camera drive unit are arranged in parallel and the rear ends are connected to the front end of the main control unit, the flash unit is arranged at the front end of the flash unit, and the camera lens base is The seat is arranged at the front end of the camera driving unit, and the linear driving device is arranged in the camera driving unit.
所述闪光灯单元内设有为旋转驱动装置和闪光灯供电的驱动电源，且所述 驱动电源通过主控制单元控制通断电。The flash unit is provided with a driving power supply for supplying power to the rotary drive device and the flash, and the driving power is controlled by the main control unit to be turned on and off.
所述主控制单元包括监控控制器和嵌入式PC模块，所述直线驱动装置和驱 动电源通过所述嵌入式PC模块控制，所述光学成像仪通过所述监控控制器监控； 所述主控制单元通过连接缆与一个数据综合处理及分析单元相连；所述主控制 单元上设有电池组件和数据存储单元。The main control unit includes a monitoring controller and an embedded PC module, the linear drive device and the drive power are controlled by the embedded PC module, and the optical imager is monitored by the monitoring controller; the main control unit A data integrated processing and analysis unit is connected through a connecting cable; the main control unit is provided with a battery assembly and a data storage unit.
所述主架体内设有叶绿素浊度传感器和温盐深传感器；所述主架体上侧设 有上导流翼板、下侧设有下导流翼板；所述主架体呈前小后大的梭形；所述主 架体上设有预留挂载孔。The main frame body is provided with a chlorophyll turbidity sensor and a temperature and salinity depth sensor; the upper side of the main frame body is provided with an upper guide vane, and the lower side is provided with a lower guide vane; the main frame body is small in the front. The rear is large in the shape of a shuttle; the main frame body is provided with a reserved mounting hole.
一种海洋浮游生物光学成像处理方法，包括以下步骤：A marine plankton optical imaging processing method, comprising the following steps:
主控制单元接收数据综合处理及分析单元的指令，控制闪光灯组件照射海 洋生物，控制相机组件采集海洋浮游生物的原始图像，控制叶绿素浊度传感器 和温盐深传感器采集海洋数据并存储，通过图像处理的程序步骤获取海洋生物 的种群图像类别并分类，用于分析得到该海洋生物种群的分布。The main control unit receives the instructions from the data comprehensive processing and analysis unit, controls the flash unit to illuminate marine organisms, controls the camera unit to collect the original images of marine plankton, controls the chlorophyll turbidity sensor and the temperature, salinity and depth sensor to collect and store marine data, and processes the image through image processing. The program steps of obtaining the population image categories of marine organisms and classifying them are used to analyze and obtain the distribution of the marine organism populations.
主控制单元输出信号控制旋转驱动装置使旋转叶轮转动，从而旋转叶轮上 沿着圆周方向均布的多个闪光灯照射海洋生物；所述主控制单元输出信号控制 直线驱动装置驱动光学成像仪在各个透镜之间切换位置，对海洋生物进行对焦 拍照。The output signal of the main control unit controls the rotary drive device to rotate the rotary impeller, so that a plurality of flash lamps uniformly distributed along the circumferential direction on the rotary impeller illuminate the marine organisms; the output signal of the main control unit controls the linear drive device to drive the optical imager in each lens. Switch between positions to focus and take pictures of marine life.
所述图像处理的程序步骤包括：The program steps of the image processing include:
a.加载采集的海洋浮游生物的原始图像；a. Load the original image of the collected marine plankton;
b.对原始图像进行预处理与对焦物体检测、分段阈值计算、梯度分析，初 步获取海洋浮游生物的轮廓；b. Perform preprocessing and focus object detection, segmentation threshold calculation, and gradient analysis on the original image to initially obtain the outline of marine plankton;
c.特征向量提取的步骤：在样本空间构建最优超平面；计算不同样本集与 超平面之间的分离距离；计算每个分离距离的均值矩阵和距离矩阵并归一化处 理；计算上述归一化后矩阵的对比度、校正度、方差作为特征向量；c. Steps of feature vector extraction: construct the optimal hyperplane in the sample space; calculate the separation distance between different sample sets and the hyperplane; calculate the mean matrix and distance matrix of each separation distance and normalize them; calculate the above normalization The contrast, correction, and variance of the normalized matrix are used as eigenvectors;
d.多环境因子参数分析：根据各个传感器获取的温度值、电导率值、压力 值、光学叶绿素浓度、浑浊度值、经纬度值分析所在观测位置的剖面生物丰度 分布、水文环境因子分布；d. Multi-environmental factor parameter analysis: According to the temperature value, conductivity value, pressure value, optical chlorophyll concentration, turbidity value, longitude and latitude value obtained by each sensor, analyze the profile biological abundance distribution and hydrological environmental factor distribution at the observation location;
e.机器学习及深度学习的步骤：在特征向量提取后的图像中插入预设数量 的特征点并筛选比较，获取浮游生物的种类特征图像并存储至图像专家库用于 深度学习；对图像专家库中存储的图像，采用非线性变换提取多层次多角度特 征，根据不同种类的目标形态特征选择具有不同大小、形状和方向特性的结构 元素用于学习浮游生物的种群类别，获取自适应特征选取的快速线形分类器；e. Steps of machine learning and deep learning: insert a preset number of feature points into the image after feature vector extraction, filter and compare, obtain the type feature image of plankton and store it in the image expert database for deep learning; From the images stored in the library, nonlinear transformation is used to extract multi-level and multi-angle features, and structural elements with different size, shape and direction characteristics are selected according to different types of target morphological features to learn the population types of plankton and obtain adaptive feature selection. The fast linear classifier of ;
f.利用自适应特征选取的快速线形分类器处理现场采集的图像，按照判别 的种群类别对图像进行分类，再结合人工鉴定剔除分类错误的图像。f. Use the fast linear classifier of adaptive feature selection to process the images collected on the spot, classify the images according to the identified species category, and then remove the wrongly classified images combined with manual identification.
还包括根据该种群的丰度数据和环境因子参数，分析得到该种群的分布。It also includes analyzing the distribution of the population based on the abundance data and environmental factor parameters of the population.
所述预处理与对焦物体检测包括采用灰度校正、图像分割并标记；所述分 段阈值计算包括根据设定的阈值进行二值化处理，对感兴趣的区域ROI计算相 邻像素的灰度差并设置sobel参数。The preprocessing and focused object detection include grayscale correction, image segmentation and marking; the segmentation threshold calculation includes binarization processing according to the set threshold, and the grayscale of adjacent pixels is calculated for the region of interest ROI. difference and set the sobel parameter.
本发明的有益效果是：The beneficial effects of the present invention are:
1、本发明利用机器视觉技术实现对海底浮游生物种群和颗粒物质的快速识 别及量化，其中直线驱动装置驱动光学成像仪移动，以完成不同CCD镜头的切 换，扩展了拍摄视野，并且能够实现针对不同尺寸浮游生物和颗粒物质的自动 对焦，另外拍摄时，闪光灯组件中的各个闪光灯随旋转叶轮旋转，将各个闪光 灯发出的光有效的聚焦，从而保证相机拍照的清晰度，并且也避免了拍摄过程 中的阴影部分，经过后续单元数据处理后可以实现对浮游生物种群行为较为准 确的推测。1. The present invention utilizes machine vision technology to realize the rapid identification and quantification of plankton populations and particulate matter on the seafloor, wherein the linear drive device drives the optical imager to move to complete the switching of different CCD lenses, expand the shooting field of view, and can achieve Automatic focusing of plankton and particulate matter of different sizes. In addition, when shooting, each flash in the flash unit rotates with the rotating impeller to effectively focus the light emitted by each flash, so as to ensure the clarity of the camera and avoid the shooting process. The shaded part in , after subsequent unit data processing, a more accurate prediction of plankton population behavior can be achieved.
2、本发明中的闪光灯组件和相机镜头基座相对设置，拍摄时不会扰动浮游 生物种群和颗粒物质，避免了对海底浮游生物种群和颗粒物质自然维度及水文 形态形式的改变。2. The flash unit and the camera lens base in the present invention are arranged opposite to each other, so that plankton populations and particulate matter are not disturbed during shooting, and changes to the natural dimensions and hydrological forms of plankton populations and particulate matter on the seafloor are avoided.
3、本发明中经过图像处理步骤后，能快速将采集到的浮游生物进行识别并 分类，从而掌握浮游生物类群的分布情况，对于研究浮游生物的昼夜垂直移动、 斑块分布等海洋生态现象具有原位精细化观测与快速鉴定优势。3. After the image processing step in the present invention, the collected plankton can be quickly identified and classified, so as to grasp the distribution of plankton groups, which is useful for studying marine ecological phenomena such as the vertical movement of plankton in day and night and the distribution of patches. In situ refined observation and rapid identification advantages.
附图说明Description of drawings
图1为本发明的结构示意图，Fig. 1 is the structural representation of the present invention,
图2为图1中的A处放大图，Fig. 2 is an enlarged view of A in Fig. 1,
图3为图1中的主控制单元、闪光灯单元及相机驱动单元示意图，FIG. 3 is a schematic diagram of the main control unit, the flash unit and the camera driving unit in FIG. 1 ,
图4为本发明的光学数据后处理流程示意图，FIG. 4 is a schematic diagram of the optical data post-processing flow diagram of the present invention,
图5为多环境因子参数分析-剖面生物丰度分布图，Figure 5 is a multi-environmental factor parameter analysis-section biological abundance distribution map,
图6为为多环境因子参数分析-水文环境因子分布图，Figure 6 is a multi-environmental factor parameter analysis-hydrological environment factor distribution diagram,
图7为本发明的光学成像效果示意图。FIG. 7 is a schematic diagram of the optical imaging effect of the present invention.
其中，1为上导流翼板，2为叶绿素浊度传感器，3为主架体，4为主控制 单元，5为电池组件，6为温盐深传感器，7为下导流翼板，8为闪光灯单元，9 为闪光灯组件，10为相机镜头基座，11为相机驱动单元，12为连接缆，13为 连板，14为数据综合处理及分析单元，15为数据存储单元，16为监控控制器， 17为嵌入式PC模块，18为驱动电源，19为旋转驱动装置，20为旋转叶轮，21 为直线驱动装置，22为光学成像仪，23为透镜。Among them, 1 is the upper guide vane, 2 is the chlorophyll turbidity sensor, 3 is the main frame, 4 is the main control unit, 5 is the battery assembly, 6 is the temperature and salt depth sensor, 7 is the lower guide vane, 8 is the flash unit, 9 is the flash unit, 10 is the camera lens base, 11 is the camera drive unit, 12 is the connecting cable, 13 is the connecting board, 14 is the data integrated processing and analysis unit, 15 is the data storage unit, 16 is the monitoring unit The controller, 17 is an embedded PC module, 18 is a driving power supply, 19 is a rotary driving device, 20 is a rotary impeller, 21 is a linear driving device, 22 is an optical imager, and 23 is a lens.
具体实施方式Detailed ways
下面结合附图和实施例，对本发明的具体实施方式作进一步详细描述。以 下实施例用于说明本发明的目的，但不用来限定本发明的范围。The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments. The following examples are provided for the purpose of illustrating the present invention, but are not intended to limit the scope of the present invention.
如图1～3所示，本发明包括主架体3、闪光灯组件9、相机组件和主控制 单元4，所述闪光灯组件9设于主架体3前端，所述闪光灯组件9包括旋转驱动 装置19、旋转叶轮20和闪光灯，旋转叶轮20通过所述旋转驱动装置19驱动旋 转，在旋转叶轮20上沿着圆周方向均布有多个闪光灯，所述相机组件包括直线 驱动装置21、光学成像仪22和相机镜头基座10，其中相机镜头基座10设于主 架体3前端，且所述相机镜头基座10朝向所述闪光灯组件9的一侧设有透镜23 作为镜头，在所述相机镜头基座10内设有光学成像仪22，且所述光学成像仪 22通过所述直线驱动装置21驱动移动以使合适的镜头与所述相机镜头基座10 上的透镜23对位，从而实现拍照焦距的变更，扩大拍照范围，另外拍照时闪光 灯组件9中的各个闪光灯随旋转叶轮20旋转，以补偿拍照的阴影区域，所述旋 转驱动装置19、闪光灯、直线驱动装置21和光学成像仪22均通过所述主控制 单元4控制。所述旋转驱动装置19、闪光灯、直线驱动装置21和光学成像仪 22均为市购产品，本实施例中，所述旋转驱动装置19为旋转马达，所述直线驱 动装置21为电动缸，所述光学成像仪22为CCD传感器，且闪光灯频率为CCD传感器拍照频率的2倍以上，以避免拍照的无效动作。As shown in FIGS. 1 to 3 , the present invention includes a main frame body 3 , a flash unit 9 , a camera assembly and a main control unit 4 , the flash unit 9 is arranged at the front end of the main frame 3 , and the flash unit 9 includes a rotation driving device 19. Rotary impeller 20 and flashing light, the rotating impeller 20 is driven to rotate by the rotary driving device 19, and a plurality of flashing lights are evenly distributed along the circumferential direction on the rotary impeller 20, and the camera assembly includes a linear driving device 21, an optical imager 22 and the camera lens base 10, wherein the camera lens base 10 is arranged at the front end of the main frame body 3, and the side of the camera lens base 10 facing the flash unit 9 is provided with a lens 23 as a lens, and the camera The lens base 10 is provided with an optical imager 22, and the optical imager 22 is driven and moved by the linear drive device 21 to make the appropriate lens align with the lens 23 on the camera lens base 10, so as to realize The change of the focal length of the photo will expand the range of the photo. In addition, each flash in the flash unit 9 rotates with the rotating impeller 20 when taking pictures to compensate for the shadow area of the photo. The rotary drive device 19, flash, linear drive device 21 and optical imager 22 All are controlled by the main control unit 4 . The rotary driving device 19, the flash, the linear driving device 21 and the optical imager 22 are all commercially available products. In this embodiment, the rotary driving device 19 is a rotary motor, and the linear driving device 21 is an electric cylinder. The optical imager 22 is a CCD sensor, and the flash frequency is more than twice the photographing frequency of the CCD sensor, so as to avoid the invalid action of photographing.
如图1和图3所示，所述主架体3内设有闪光灯单元8和相机驱动单元11， 所述闪光灯单元8和相机驱动单元11平行设置且后端分别通过连板13与主控 制单元4前端相连，所述连板13内部中空以供各种线路穿过，所述闪光灯组件 9设于闪光灯单元8前端，所述相机镜头基座10设于相机驱动单元11前端，且 所述相机驱动单元11内设有所述直线驱动装置21。As shown in FIG. 1 and FIG. 3 , the main frame body 3 is provided with a flash unit 8 and a camera driving unit 11 . The flash unit 8 and the camera driving unit 11 are arranged in parallel, and the rear ends are connected to the main control unit through the connecting plate 13 respectively. The front end of the unit 4 is connected to the front end, the connecting plate 13 is hollow inside for various lines to pass through, the flash unit 9 is arranged at the front end of the flash unit 8, the camera lens base 10 is arranged at the front end of the camera driving unit 11, and the The camera driving unit 11 is provided with the linear driving device 21 .
如图3所示，所述闪光灯单元8内设有为旋转驱动装置19和闪光灯供电的 驱动电源18，且所述驱动电源18通过主控制单元4控制通断电，进而使闪光灯 组件9中的旋转驱动装置19和闪光灯通电或断电。所述驱动电源18为本领域 公知技术，另外本实施例中，在所述旋转叶轮20中部轮轴后端设有一个旋转接 头，所述驱动电源18通过线路与所述旋转接头相连，再由所述旋转接头分出多 个线路分别与各个闪光灯连接，为闪光灯供电同时不会影响旋转，所述旋转接 头为市购产品。As shown in FIG. 3 , the flash unit 8 is provided with a driving power supply 18 for supplying power to the rotary driving device 19 and the flash, and the driving power 18 is controlled by the main control unit 4 to be turned on and off, thereby enabling the flash unit 9 to be powered on and off. The rotary drive 19 and the flash are energized or de-energized. The driving power source 18 is a well-known technology in the art. In addition, in this embodiment, a rotary joint is provided at the rear end of the axle in the middle of the rotating impeller 20. The driving power source 18 is connected to the rotary joint through a line, and is The rotary joint is divided into a plurality of lines to be connected to each flashlight respectively, so as to supply power to the flashlight without affecting the rotation, and the rotary joint is a commercially available product.
如图1和图3所示，所述主控制单元4包括监控控制器16和嵌入式PC模 块17，其中嵌入式PC模块17可以根据需要控制直线驱动装置21伸长与缩短， 以推动光学成像仪22移动，从而使得光学成像仪22上的不同镜头与相机镜头 基座上的透镜23对位，实现拍照焦距的变更，扩大拍照视野，另外嵌入式PC 模块17还控制闪光灯单元8内的驱动电源18为闪光灯和旋转叶轮20提供电力， 监控控制器16能不间断的对光学成像仪22的成像情况进行监控。所述监控控 制器16和嵌入式PC模块17均为本领域公知技术。As shown in FIG. 1 and FIG. 3 , the main control unit 4 includes a monitoring controller 16 and an embedded PC module 17 , wherein the embedded PC module 17 can control the linear drive device 21 to extend and shorten as required to promote optical imaging The instrument 22 is moved, so that the different lenses on the optical imager 22 are aligned with the lens 23 on the camera lens base, so as to realize the change of the focal length of the photographing and expand the photographing field of view. In addition, the embedded PC module 17 also controls the drive in the flash unit 8. The power supply 18 provides power for the flash lamp and the rotating impeller 20 , and the monitoring controller 16 can continuously monitor the imaging condition of the optical imager 22 . The monitoring controller 16 and the embedded PC module 17 are known in the art.
如图1所示，所述主控制单元4通过两条连接缆12与船体上的数据综合处 理及分析单元14相连，其中嵌入式PC模块17通过一条连接缆12与数据综合 处理及分析单元14相连，嵌入式PC模块(17)内部设有目标生物图像采集程 序，当作为光学成像仪22的CCD传感器拍照后会通过关键参数设置作为边缘条 件对图片进行识别，判断图片中包含的浮游生物或颗粒物质的种类，并将信息 进行反馈交由数据综合处理及分析单元14处理，监控控制器16通过另一连接 缆12与数据综合处理及分析单元14相连以进行数据通讯，保证实时监控。所 述数据综合处理及分析单元14为本领域公知技术。As shown in FIG. 1 , the main control unit 4 is connected to the data integrated processing and analysis unit 14 on the hull through two connecting cables 12 , wherein the embedded PC module 17 is connected to the data integrated processing and analysis unit 14 through one connecting cable 12 Connected, the embedded PC module (17) is internally provided with a target biological image acquisition program. When the CCD sensor as the optical imager 22 takes a picture, it will identify the picture by setting key parameters as the edge condition, and determine the plankton or plankton contained in the picture. The type of particulate matter, and the information is fed back to the data integrated processing and analysis unit 14 for processing. The monitoring controller 16 is connected to the data integrated processing and analysis unit 14 through another connecting cable 12 for data communication to ensure real-time monitoring. The data integrated processing and analysis unit 14 is a well-known technology in the art.
如图1～2所示，所述主控制单元4上设有电池组件5和数据存储单元15， 所述电池组件5为整个主控制单元4供电，所述数据存储单元15用于存储采集 的数据信息。所述电池组件5和数据存储单元15均为本领域公知技术。As shown in Figures 1-2, the main control unit 4 is provided with a battery assembly 5 and a data storage unit 15, the battery assembly 5 supplies power to the entire main control unit 4, and the data storage unit 15 is used to store the collected data. Data information. The battery assembly 5 and the data storage unit 15 are known in the art.
如图1～2所示，所述主架体3内设有叶绿素浊度传感器2和温盐深传感器 6，其中叶绿素浊度传感器2用于测量叶绿素浓度和水体浑浊度，温盐深传感器 6用于测量剖面温盐深数据，所述叶绿素浊度传感器2和温盐深传感器6的测量 数据传回至数据综合处理及分析单元14处理，以作为浮游生物分布的环境背景 因子，用于验证水文特征对浮游生物的影响，所述叶绿素浊度传感器2和温盐 深传感器6均为市购产品。As shown in Figures 1-2, the main frame body 3 is provided with a chlorophyll turbidity sensor 2 and a temperature and salinity sensor 6, wherein the chlorophyll turbidity sensor 2 is used to measure the chlorophyll concentration and water turbidity, and the temperature and salinity sensor 6 For measuring profile temperature and salinity data, the measured data of the chlorophyll turbidity sensor 2 and the temperature and salinity sensor 6 are sent back to the data comprehensive processing and analysis unit 14 for processing, as an environmental background factor for plankton distribution, for verification For the influence of hydrological characteristics on plankton, the chlorophyll turbidity sensor 2 and the temperature-saline-depth sensor 6 are both commercially available products.
如图1～2所示，所述主架体3上侧设有上导流翼板1、下侧设有下导流翼 板7，所述上导流翼1和下导流翼7呈燕尾型设计，能有效减少航行过程中海水 阻力，节省牵引力，提高效率，所述主架体3则设计成前小后大的梭形，适合 在海水中航行，整体框架使用螺栓连接，减少因海水对焊接部位腐蚀而降低寿 命的影响，同时增强设备了的可维护性，另外主架体3中部设计有多个预留挂 载孔，可以根据需要搭载不同的水文传感器，以便满足环境要素观测的需求。As shown in Figures 1-2, the main frame body 3 is provided with an upper guide vane 1 on the upper side and a lower guide vane 7 on the lower side. The upper guide vane 1 and the lower guide vane 7 are arranged in a The dovetail design can effectively reduce the seawater resistance during sailing, save traction and improve efficiency. The main frame body 3 is designed into a shuttle shape with a small front and a large rear, suitable for sailing in seawater. The impact of seawater on the corrosion of the welded part reduces the service life, and at the same time enhances the maintainability of the equipment. In addition, there are multiple reserved mounting holes in the middle of the main frame body 3, which can be equipped with different hydrological sensors according to needs, so as to meet the observation of environmental elements. demand.
本发明装置的工作原理为：The working principle of the device of the present invention is:
本发明利用机器视觉技术实现对海底浮游生物种群和颗粒物质的快速识别 及量化，并且避免了对海底浮游生物种群和颗粒物质自然维度及水文形态形式 的改变，从而可以实现对浮游生物种群行为较为准确的推测，其中本发明利用 直线驱动装置21驱动作为光学成像仪22的CCD传感器移动，以使CCD传感器 上不同镜头分别与所述相机镜头基座10上的透镜23对位，从而完成不同CCD 镜头的切换，扩展了拍摄视野，并且能够实现针对不同尺寸浮游生物和颗粒物 质的自动对焦，所述闪光灯组件9与相机镜头基座10相对设置，当光学成像仪 22拍摄时，闪光灯组件9中的各个闪光灯随旋转叶轮20旋转，并通过旋转叶轮20的转动将闪光灯发出的光有效的聚焦，从而保证在一定范围内相机拍照的清 晰度，并且闪光灯随旋转叶轮20一起旋转，也避免了拍摄过程中的阴影部分。The invention utilizes the machine vision technology to realize the rapid identification and quantification of the plankton populations and particulate matter on the seafloor, and avoids changing the natural dimensions and hydrological forms of the plankton populations and particulate matter on the seafloor, so that the behavior of the plankton populations can be compared. Accurate speculation, wherein the present invention uses the linear drive device 21 to drive the CCD sensor as the optical imager 22 to move, so that the different lenses on the CCD sensor are aligned with the lens 23 on the camera lens base 10 respectively, so as to complete the different CCD sensors. The switching of the lens expands the shooting field of view, and can realize automatic focusing for plankton and particulate matter of different sizes. Each flashing light rotates with the rotating impeller 20, and the light emitted by the flashing light is effectively focused by the rotation of the rotating impeller 20, thereby ensuring the clarity of the camera within a certain range, and the flashing light rotates with the rotating impeller 20, which also avoids shooting. The shaded part of the process.
在所述主控制单元4中设有嵌入式PC模块17和监控控制器16，可以实现 对闪光灯组件9和作为光学成像仪22的CCD传感器的实时监控与控制，其中嵌 入式PC模块17内部设有目标生物图像采集程序，在CCD拍照后会通过关键参 数设置作为边缘条件对图片进行识别，判断图片中包含的浮游生物或颗粒物质 的种类，并将信息进行反馈，并通过数据综合处理及分析单元14进行处理， 其具体处理过程如图4所示，采集处理后的光学影像效果如图7所示。The main control unit 4 is provided with an embedded PC module 17 and a monitoring controller 16, which can realize real-time monitoring and control of the flash unit 9 and the CCD sensor as the optical imager 22, wherein the embedded PC module 17 is internally provided with There is a target biological image acquisition program. After the CCD takes a picture, it will identify the image through key parameter settings as edge conditions, determine the type of plankton or particulate matter contained in the image, and feed back the information, and comprehensively process and analyze the data. The unit 14 performs processing, and its specific processing process is shown in FIG. 4 , and the optical image effect after collection and processing is shown in FIG. 7 .
如图3所示，所述的一种海洋浮游生物光学成像数据综合处理及分析，具 体的实现如下：原始的图像数据输入后设定相应的参数(如“threshold”和 “sigma”等)，目标生物依托黑白二值化的设定将感兴趣的区域(ROI)用白色 线框图圈定，从而获得连续而封闭的目标轮廓。同步用线框图将潜在的目标生 物个体圈定后单独显示并储存在相关的文件夹中，每个照片均有对应的标号和 环境因子的相关信息。As shown in Fig. 3, the above-mentioned comprehensive processing and analysis of marine plankton optical imaging data is implemented as follows: after inputting the original image data, set corresponding parameters (such as "threshold" and "sigma", etc.), The target organism relies on the setting of black and white binarization to delineate the region of interest (ROI) with a white wireframe, so as to obtain a continuous and closed target contour. Simultaneously, wireframes are used to delineate potential target organisms and then display them separately and store them in relevant folders. Each photo has a corresponding label and information about environmental factors.
主要处理步骤如下：1、浮游生物光学成像装置原始的图像数据导入，2、 关键参数设计，3、特征向量提取，4、多环境因子参数分析，5、机器学习 及深度学习，6、种类鉴定，7、数据质控及分析。The main processing steps are as follows: 1. Import the original image data of the plankton optical imaging device, 2. Key parameter design, 3. Feature vector extraction, 4. Multi-environmental factor parameter analysis, 5. Machine learning and deep learning, 6. Type identification , 7, data quality control and analysis.
其中的2关键参数设计。包含预处理与对焦物体检测、分段阈值计算、梯 度分析三个主要内容。2 key parameters of which are designed. It includes three main contents: preprocessing and focused object detection, segmentation threshold calculation, and gradient analysis.
其中预处理及对焦物体检测为：对原始图像依次进行灰度校正、边缘检测 并设置边缘阈值获取感兴趣的区域ROI。具体包括灰度校正(通过灰度变换对像 素进行灰度拉伸使得像素分布在可视的灰度阶内)、分割(达到分割背景区域与 前景轮廓边缘模糊区域中的不可修复区域分割)、标记(将分割区域进行标注以 便下一步分析，水深及环境要素信息同步导入)。Among them, preprocessing and focused object detection are: performing grayscale correction, edge detection on the original image in turn, and setting the edge threshold to obtain the ROI of the region of interest. Specifically, it includes grayscale correction (grayscale stretching of pixels through grayscale transformation so that the pixels are distributed in the visible grayscale), segmentation (to achieve segmentation of the background area and the irreparable area in the blurred area of the foreground contour edge), Marking (mark the segmented area for further analysis, and import water depth and environmental element information synchronously).
分段阈值计算即：根据设定的阈值进行二值化处理；具体为使用聚焦检测 算法对这些检测点进行选定和分析；引入某些参数如“threshold”、“sigma” 和“sobel”等进行参数设定高低阈值，低值通常设为0，高值有一个参考值， 通常高于该高值时，图像显示为高亮的白值，低于时为黑色的暗值；The segmentation threshold calculation is: perform binarization processing according to the set threshold; specifically, use the focus detection algorithm to select and analyze these detection points; introduce certain parameters such as "threshold", "sigma" and "sobel", etc. Set the high and low thresholds by parameter. The low value is usually set to 0, and the high value has a reference value. Usually, when the high value is higher than the high value, the image is displayed as a bright white value, and when it is lower than the black dark value;
对感兴趣的区域ROI进行梯度分析获取轮廓线；具体为在选定区域内进行 计算相邻像素的灰度差并进行边缘检测，初步检测出图像的大致轮廓；进而设置 sobel参数，有利于对受噪声影响且灰度渐变图像的分割。Perform gradient analysis on the ROI of the region of interest to obtain contour lines; specifically, calculate the grayscale difference of adjacent pixels in the selected area and perform edge detection to initially detect the rough outline of the image; and then set the sobel parameter, which is conducive to Segmentation of grayscale gradient images affected by noise.
3.特征向量提取，结合图像的纹理、构图的形状及各部分之间的空间关系， 对图像噪声抑制处理，提取有意义的目标信息，使图像数据变成容易进行处理 的有组织的数据。3. Feature vector extraction, combined with the texture of the image, the shape of the composition and the spatial relationship between each part, the image noise is suppressed, the meaningful target information is extracted, and the image data becomes organized data that is easy to process.
在样本空间(样本空间为上一步骤初步检测图像轮廓的图像集合)，通过构 建最优超平面从而获取感兴趣区域的目标个体。主要是利用不同样本集与超平 面之间的分离距离最大，从而达到最大的泛化能力。利用不变矩(平移、旋转 和尺度)为主要特征的数学形态学参数，每个分离距离的均值矩阵和距离矩阵 实现相对旋转不变性。对所得矩阵进行归一化处理，实现尺度不变性。计算了 这些矩阵的对比度、校正度、方差等，并将其作为特征向量，当同一种浮游生 物目标有位移或尺度变化时，仍可以将其分为同一类目标，避免了误分离，该 方法对遮挡和投影敏感性较低。对于浮游生物，采用包含浮游生物细胞组织表 面结构排列的重要信息的纹理特征提取，在识别中起到重要作用。与其他类别 特征相比,纹理特征能更好地反映浮游生物图像的宏观与微观结构性质。In the sample space (the sample space is the image set for the initial detection of the image contour in the previous step), the target individual in the region of interest is obtained by constructing the optimal hyperplane. The main purpose is to use the maximum separation distance between different sample sets and hyperplanes, so as to achieve the maximum generalization ability. Using the mathematical morphological parameters with invariant moments (translation, rotation and scale) as the main features, the mean matrix and distance matrix of each separation distance achieve relative rotation invariance. The resulting matrix is normalized to achieve scale invariance. The contrast, correction degree, variance, etc. of these matrices are calculated and used as eigenvectors. When the same plankton target has displacement or scale change, it can still be classified into the same type of target to avoid false separation. This method Less sensitive to occlusion and projection. For plankton, texture feature extraction, which contains important information about the arrangement of plankton cell tissue surface structure, plays an important role in identification. Compared with other categories of features, texture features can better reflect the macroscopic and microstructural properties of plankton images.
4、多环境因子参数分析：获取温盐深传感器、叶绿素浊度传感器检测的数 据。主要包含温度值、电导率值、压力值、光学叶绿素浓度、浑浊度值、经纬 度值。其中的数据格式如下所示：4. Analysis of multi-environmental factor parameters: Obtain the data detected by the temperature and salinity sensor and the chlorophyll turbidity sensor. It mainly includes temperature value, conductivity value, pressure value, optical chlorophyll concentration, turbidity value, latitude and longitude value. The format of the data is as follows:
Output.Format("ctd％08u:％f,％f,％f,％f,％f,％f,％f,％f,％f\r\n",ms,Conduc tivity,Temperature,Pressure,Salinity,ChloroRef,ChloroSig,TurbRef,TurbSig,Altitude)；Output.Format("ctd％08u:％f,％f,％f,％f,％f,％f,％f,％f,％f\r\n",ms,Conduc tivity,Temperature,Pressure, Salinity, ChloroRef, ChloroSig, TurbRef, TurbSig, Altitude);
ctd6********(##):ROI time(ctd********)记录获取关注个体采样 时间(ms)，若同一时间提取出多个个体，则##自01开始依次记录ctd6********(##): ROI time(ctd********) records the sampling time (ms) for obtaining the concerned individual. If multiple individuals are extracted at the same time, ##self 01 Start recording sequentially
Conductivity:3.573590,温盐深测量仪记录采集图像时电导率值Conductivity: 3.573590, the temperature and salinity measuring instrument records the conductivity value when the image is collected
Temperature:9.619100,温盐深测量仪记录采集图像时温度值Temperature: 9.619100, the temperature and salinity measuring instrument records the temperature value when the image is collected
Pressure:67.970000,温盐深测量仪记录采集图像时的压力值Pressure: 67.970000, the temperature and salinity measuring instrument records the pressure value when the image is collected
Salinity:32.917400,温盐深测量仪记录采集图像时盐度值Salinity: 32.917400, the salinity value when the temperature and salinity measuring instrument records the image
ChloroRef:0.000000,搭载的光学叶绿素传感器采集时波段值(nm)ChloroRef: 0.000000, the band value (nm) when the optical chlorophyll sensor on board is collected
ChloroSig:0.000000,该单波段采集叶绿素时记录的原始电压值ChloroSig: 0.000000, the original voltage value recorded when collecting chlorophyll with this single band
TurbRef:0.000000,搭载的光学浊度传感器采集时波段值(nm)TurbRef: 0.000000, the band value (nm) when the optical turbidity sensor is installed
TurbSig:0.000000,该单波段采集浊度时记录的原始电压值TurbSig: 0.000000, the original voltage value recorded when the single band collected turbidity
Altitude:0.000000接入GPS时记录的经纬度Altitude: 0.000000 Longitude and latitude recorded when GPS is connected
通过数据综合处理及分析技术在黄海3600-05站位得到的同一站位的剖面 生物丰度分布、数据如下所示：The profile biological abundance distribution and data obtained at the same station in the Yellow Sea at station 3600-05 through data comprehensive processing and analysis technology are as follows:
如图5所示，为多环境因子参数分析-剖面生物丰度分布图。标注点为浮游 动物出现的水深及计算的丰度；灰度区是基于瞬时观测的丰度值使用 Savitzky-Golay卷积平滑算法估算临近水层的丰度，是5点移动平滑算法的改 进。A线为全剖面的叶绿素的分布情况。As shown in Figure 5, it is a multi-environmental factor parameter analysis-section biological abundance distribution map. The marked points are the water depth and the calculated abundance of zooplankton; the gray area is based on the instantaneous observed abundance value using the Savitzky-Golay convolution smoothing algorithm to estimate the abundance of the adjacent water layer, which is an improvement of the 5-point moving smoothing algorithm. Line A is the distribution of chlorophyll in the full section.
如图6所示，为多环境因子参数分析-水文环境因子分布图。C线为全剖面 的叶绿素的分布情况；B线为全剖面的温度的分布情况；D线为全剖面的盐度的 分布情况。As shown in Figure 6, it is a multi-environmental factor parameter analysis-hydrological environment factor distribution diagram. Line C is the distribution of chlorophyll in the whole section; line B is the distribution of temperature in the whole section; line D is the distribution of salinity in the whole section.
5.机器学习及深度学习：利用在步骤3的结果图像中插入一定数量的特征 点，通过对特征点的比较，将符合特征点的部分加以整合和归纳，最终得出图 像的特征进而深度学习。建立并不断补充图像专家库，对提取的特征信息进行 分析，提取与人们对浮游生物主观理解相符合的特征，使识别结果与实际人们 的视觉判断相吻合。通过数据驱动的方式，采用一系列的非线性变换，从原始 数据中提取多层次多角度特征，根据不同的目标形态特征选择具有不同大小、 形状和方向特性的结构元素，从而使获得的特征具有更强的泛化能力和表达能 力，这恰好满足高效图像处理的需求。5. Machine learning and deep learning: insert a certain number of feature points into the result image of step 3, and integrate and summarize the parts that match the feature points by comparing the feature points, and finally obtain the features of the image and then perform deep learning. . Establish and continuously supplement the image expert database, analyze the extracted feature information, and extract the features that are consistent with people's subjective understanding of plankton, so that the identification results are consistent with actual people's visual judgments. In a data-driven way, a series of nonlinear transformations are used to extract multi-level and multi-angle features from the original data, and structural elements with different size, shape and direction characteristics are selected according to different target morphological features, so that the obtained features have Stronger generalization ability and expression ability, which just meet the needs of efficient image processing.
如图7所示，为通过内部的程序-机器学习及深度学习处理后得到的海洋浮 游生物个体的光学影像资料，光学成像-原位采集物种(桡足类)。As shown in Figure 7, the optical imaging data of individual marine plankton obtained after processing by the internal program-machine learning and deep learning, optical imaging-in situ collection of species (Copepod).
6.种类鉴定：采用机器与人工鉴定相结合，获取生物种类。6. Species identification: The combination of machine and manual identification is used to obtain biological species.
使用机器计算可快速有效的实时识别并分类,通过前文所述的实现了自适 应特征选取的快速线形分类器进行目标分类通过目标一致性匹配,结合建立的 图像专家库及自动统计算法,有效获取生物目标的统计和分布信息,并实时反馈 给观测者。观测者通过人工鉴定的方式对机器分类的图像进行进一步的评估， 查看分类的图像正确与否，将有误的图像删除。Machine computing can be used for fast and effective real-time identification and classification, and the target classification is carried out through the fast linear classifier that realizes the adaptive feature selection mentioned above. Through target consistency matching, combined with the established image expert database and automatic statistical algorithm, the effective The statistics and distribution information of biological targets are fed back to the observer in real time. The observer further evaluates the images classified by the machine through manual identification, checks whether the classified images are correct or not, and deletes the wrong images.
7.数据质控与分析：代表种丰度估算及其分布影响因素。7. Data quality control and analysis: Representative species abundance estimation and its distribution influencing factors.
针对观测的代表种的分类信息，结合前期标记的水深位置信息，拍照时有 效的采集体积等，计算该类种群的丰度，同时与环境因子建立联系，进而分析 相应的科学问题，如影响其分布的主要环境因素。According to the classification information of the observed representative species, combined with the water depth position information marked in the early stage, the effective collection volume when taking pictures, etc., calculate the abundance of this type of population, and establish a relationship with environmental factors, and then analyze the corresponding scientific issues, such as affecting its The main environmental factors of distribution.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910370129.4ACN110057824B (en) | 2019-05-06 | 2019-05-06 | Ocean plankton optical imaging device and imaging processing method |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910370129.4ACN110057824B (en) | 2019-05-06 | 2019-05-06 | Ocean plankton optical imaging device and imaging processing method |
| Publication Number | Publication Date |
|---|---|
| CN110057824Atrue CN110057824A (en) | 2019-07-26 |
| CN110057824B CN110057824B (en) | 2024-05-14 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910370129.4AActiveCN110057824B (en) | 2019-05-06 | 2019-05-06 | Ocean plankton optical imaging device and imaging processing method |
| Country | Link |
|---|---|
| CN (1) | CN110057824B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114298940A (en)* | 2021-12-30 | 2022-04-08 | 声耕智能科技(西安)研究院有限公司 | Hyperspectral image shadow compensation method based on nonlinear pixel analysis |
| CN114299332A (en)* | 2021-12-22 | 2022-04-08 | 苏州热工研究院有限公司 | Cold source marine organism intelligent detection method and system for nuclear power plant |
| CN118794953A (en)* | 2024-09-14 | 2024-10-18 | 浙江清盛检测技术有限公司 | A plankton multi-viewing angle in-situ detection device and detection method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06186607A (en)* | 1992-12-18 | 1994-07-08 | Gokou Eizo Kagaku Kenkyusho:Kk | Camera with zoom lens |
| KR0131680B1 (en)* | 1986-05-12 | 1998-04-15 | 마쓰모도 도오루 | Finder optical system of the camera |
| US20010010760A1 (en)* | 2000-01-28 | 2001-08-02 | Masashi Saito, Hiroyuki Oguni, Yoshiteru Okada And Hiroshi Okada | Intraoral imaging camera system |
| JP2005091061A (en)* | 2003-09-16 | 2005-04-07 | Shin Nippon Air Technol Co Ltd | Handholding particle visualizing apparatus and assembling method therefor |
| JP2005221270A (en)* | 2004-02-03 | 2005-08-18 | Tsunehiro Yoshida | Screw inspection apparatus |
| JP2005249824A (en)* | 2004-03-01 | 2005-09-15 | Casio Comput Co Ltd | Imaging device |
| JP2008009187A (en)* | 2006-06-29 | 2008-01-17 | Olympus Corp | Automatic focusing device |
| CN101592619A (en)* | 2009-07-15 | 2009-12-02 | 长沙楚天科技有限公司 | Adopt the automatic lamp inspector picture synchronization tracking and collecting device of end light source |
| KR20100060503A (en)* | 2008-11-27 | 2010-06-07 | 삼성전기주식회사 | Method for controlling an angle of light radiated by flash and camera having function of adjusting an angle of light radiated by flash |
| US20120327274A1 (en)* | 2011-06-23 | 2012-12-27 | Taguchi Yuzuru | Optical instrument, and method for optical instrument |
| US20130012794A1 (en)* | 2010-03-17 | 2013-01-10 | Haishan Zeng | Rapid multi-spectral imaging methods and apparatus and applications for cancer detection and localization |
| US20140204385A1 (en)* | 2010-04-19 | 2014-07-24 | Florida Atlantic University | Mems microdisplay optical imaging and sensor systems for underwater and other scattering environments |
| JP2015012375A (en)* | 2013-06-27 | 2015-01-19 | 貞彦 寺川 | Multi-angle image photographing system, rotating table device, photographing device, and multi-angle image photographing method |
| CN104872083A (en)* | 2015-05-29 | 2015-09-02 | 中国科学院海洋研究所 | Visual multi-connected sampling system of plankton and use method thereof |
| CN204836320U (en)* | 2015-06-29 | 2015-12-02 | 吴少波 | Can remote control and multi -angle rotating's multi -functional camera |
| US20150356725A1 (en)* | 2013-01-28 | 2015-12-10 | Sinvent As | System and method for counting zooplankton |
| JP2016161597A (en)* | 2015-02-26 | 2016-09-05 | 株式会社ソシオネクスト | Image processing device, integrated circuit, and camera |
| JP2016224335A (en)* | 2015-06-02 | 2016-12-28 | カシオ計算機株式会社 | Light source device and projection device |
| CN206117825U (en)* | 2016-09-21 | 2017-04-19 | 宿迁琛博信息科技有限公司 | But multi -angle rotating has camera of flash light |
| US20170293217A1 (en)* | 2014-09-22 | 2017-10-12 | Woods Hole Oceanographic Institution | Continuous particle imaging and classification system |
| CN107271371A (en)* | 2017-07-28 | 2017-10-20 | 中国科学院海洋研究所 | A kind of planktonic organism polarization imager |
| CN209841715U (en)* | 2019-05-06 | 2019-12-24 | 中国科学院海洋研究所 | Optical imaging device for marine plankton |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR0131680B1 (en)* | 1986-05-12 | 1998-04-15 | 마쓰모도 도오루 | Finder optical system of the camera |
| JPH06186607A (en)* | 1992-12-18 | 1994-07-08 | Gokou Eizo Kagaku Kenkyusho:Kk | Camera with zoom lens |
| US20010010760A1 (en)* | 2000-01-28 | 2001-08-02 | Masashi Saito, Hiroyuki Oguni, Yoshiteru Okada And Hiroshi Okada | Intraoral imaging camera system |
| JP2005091061A (en)* | 2003-09-16 | 2005-04-07 | Shin Nippon Air Technol Co Ltd | Handholding particle visualizing apparatus and assembling method therefor |
| JP2005221270A (en)* | 2004-02-03 | 2005-08-18 | Tsunehiro Yoshida | Screw inspection apparatus |
| JP2005249824A (en)* | 2004-03-01 | 2005-09-15 | Casio Comput Co Ltd | Imaging device |
| JP2008009187A (en)* | 2006-06-29 | 2008-01-17 | Olympus Corp | Automatic focusing device |
| KR20100060503A (en)* | 2008-11-27 | 2010-06-07 | 삼성전기주식회사 | Method for controlling an angle of light radiated by flash and camera having function of adjusting an angle of light radiated by flash |
| CN101592619A (en)* | 2009-07-15 | 2009-12-02 | 长沙楚天科技有限公司 | Adopt the automatic lamp inspector picture synchronization tracking and collecting device of end light source |
| US20130012794A1 (en)* | 2010-03-17 | 2013-01-10 | Haishan Zeng | Rapid multi-spectral imaging methods and apparatus and applications for cancer detection and localization |
| US20140204385A1 (en)* | 2010-04-19 | 2014-07-24 | Florida Atlantic University | Mems microdisplay optical imaging and sensor systems for underwater and other scattering environments |
| US20120327274A1 (en)* | 2011-06-23 | 2012-12-27 | Taguchi Yuzuru | Optical instrument, and method for optical instrument |
| US20150356725A1 (en)* | 2013-01-28 | 2015-12-10 | Sinvent As | System and method for counting zooplankton |
| JP2015012375A (en)* | 2013-06-27 | 2015-01-19 | 貞彦 寺川 | Multi-angle image photographing system, rotating table device, photographing device, and multi-angle image photographing method |
| US20170293217A1 (en)* | 2014-09-22 | 2017-10-12 | Woods Hole Oceanographic Institution | Continuous particle imaging and classification system |
| JP2016161597A (en)* | 2015-02-26 | 2016-09-05 | 株式会社ソシオネクスト | Image processing device, integrated circuit, and camera |
| CN104872083A (en)* | 2015-05-29 | 2015-09-02 | 中国科学院海洋研究所 | Visual multi-connected sampling system of plankton and use method thereof |
| JP2016224335A (en)* | 2015-06-02 | 2016-12-28 | カシオ計算機株式会社 | Light source device and projection device |
| CN204836320U (en)* | 2015-06-29 | 2015-12-02 | 吴少波 | Can remote control and multi -angle rotating's multi -functional camera |
| CN206117825U (en)* | 2016-09-21 | 2017-04-19 | 宿迁琛博信息科技有限公司 | But multi -angle rotating has camera of flash light |
| CN107271371A (en)* | 2017-07-28 | 2017-10-20 | 中国科学院海洋研究所 | A kind of planktonic organism polarization imager |
| CN209841715U (en)* | 2019-05-06 | 2019-12-24 | 中国科学院海洋研究所 | Optical imaging device for marine plankton |
| Title |
|---|
| RIESBECK, T. 等: "Generation of tailored pulse trains for efficient material processing by a high power MOPA system with birefringence compensation", 《LASER PHYSICS LETTERS》, vol. 5, no. 3, 1 January 2008 (2008-01-01), pages 240 - 245* |
| 丁志华 等: "快速扫频光源及其在光学频域成像中的应用", 《中国激光》, vol. 36, no. 10, pages 2469 - 2476* |
| 代鲁平 等: "基于ZooScan图像技术的南黄海夏季浮游动物群落结构分析", 《海洋与湖沼》, vol. 47, no. 4, 15 July 2016 (2016-07-15), pages 764 - 773* |
| 王惠斌 等: "环形闪光灯在现场勘查中的运用", 《数字与缩微影像》, no. 03, 15 September 2007 (2007-09-15), pages 44 - 46* |
| 王昊鹏 等: "激光三维成像技术及其主要应用", 《电子设计工程》, vol. 20, no. 12, 30 June 2012 (2012-06-30), pages 160 - 163* |
| 董仲华: "《摄影技术基础》", vol. 1, 30 June 1990, 北京广播学院出版社, pages: 1 - 10* |
| 陈磊: "基于分子标记的刺苦草在长江中下游湖泊的种群动态与维持研究", 《中国博士学位论文全文数据库 基础科学辑》, no. 5, 15 November 2007 (2007-11-15), pages 006 - 44* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114299332A (en)* | 2021-12-22 | 2022-04-08 | 苏州热工研究院有限公司 | Cold source marine organism intelligent detection method and system for nuclear power plant |
| CN114298940A (en)* | 2021-12-30 | 2022-04-08 | 声耕智能科技(西安)研究院有限公司 | Hyperspectral image shadow compensation method based on nonlinear pixel analysis |
| CN114298940B (en)* | 2021-12-30 | 2025-04-15 | 声耕智能科技(西安)研究院有限公司 | A shadow compensation method for hyperspectral images based on nonlinear pixel analysis |
| CN118794953A (en)* | 2024-09-14 | 2024-10-18 | 浙江清盛检测技术有限公司 | A plankton multi-viewing angle in-situ detection device and detection method |
| Publication number | Publication date |
|---|---|
| CN110057824B (en) | 2024-05-14 |
| Publication | Publication Date | Title |
|---|---|---|
| Ma et al. | A state-of-the-art survey of object detection techniques in microorganism image analysis: from classical methods to deep learning approaches | |
| Chen et al. | Multiscale attention networks for pavement defect detection | |
| Lian et al. | An overview of image segmentation based on pulse-coupled neural network | |
| CN104112370B (en) | Parking lot based on monitoring image intelligent car position recognition methods and system | |
| Fan et al. | Automate fry counting using computer vision and multi-class least squares support vector machine | |
| CN108765402B (en) | Non-woven fabric defect detection and classification method | |
| US10304188B1 (en) | Apparatus and method for automated cell analysis | |
| CN101701906B (en) | Method and device for detecting stored-grain insects based on near infrared super-spectral imaging technology | |
| CN114119526A (en) | A system and method for detecting and identifying surface defects of steel plates based on machine vision | |
| CN112883969B (en) | Rainfall intensity detection method based on convolutional neural network | |
| CN110057824A (en) | A kind of halomereid optical imaging device and image processing method | |
| CN106874929B (en) | Pearl classification method based on deep learning | |
| US12051254B2 (en) | Typing biological cells | |
| CN111768402A (en) | A freshness assessment method of chilled pomfret based on MU-SVM | |
| KR100889997B1 (en) | Ship Ballast Water Inspection System Using Image Processing and Its Method | |
| CN119245745B (en) | Device and method for determining multiple qualities of hatching eggs | |
| CN118351100A (en) | Image definition detection and processing method based on deep learning and gradient analysis | |
| CN109886170A (en) | An intelligent detection, recognition and statistics system for snails | |
| Fan et al. | Estimating the aquatic-plant area on a pond surface using a hue-saturation-component combination and an improved Otsu method | |
| CN106991441A (en) | Merge the plant specimen sorting technique and system of multiple dimensioned direction textural characteristics | |
| Peng et al. | An adaptive coarse-fine semantic segmentation method for the attachment recognition on marine current turbines | |
| CN107330472A (en) | A kind of automatic identifying method of unmarked model animal individual | |
| Zhang et al. | Fabric defect detection based on visual saliency map and SVM | |
| CN209841715U (en) | Optical imaging device for marine plankton | |
| CN108073873A (en) | Human face detection and tracing system based on high-definition intelligent video camera |
| 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 |