Aquatic organism health degree monitoring method in aquatic organism isolation field and related device

Abstract

The invention discloses an aquatic organism health degree monitoring method in an aquatic organism isolation field and a related device, and the method comprises the steps: enabling an underwater robot to control an underwater camera to carry out the image collection processing of a target aquatic organism based on an image collection instruction after the underwater robot moves to a preset position relative to the target aquatic organism in water, transmitting the obtained target image data corresponding to the target aquatic organism back to the far-end server; and after receiving the target image data corresponding to the target aquatic organism, the far-end server performs health degree analysis processing on the target aquatic organism by using the target image data, and pushes a health degree analysis result to a monitoring management end. In the embodiment of the invention, the image acquisition of the target aquatic organism is realized through remote control, the health degree analysis of the target aquatic organism is realized through the acquired target image data, regular patrol of corresponding workers is not needed, the patrol cost is reduced, and the health condition of the target aquatic organism can be mastered in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly relates to a method for monitoring the health of aquatic organisms in an aquatic organism isolation field and related devices. Background Art

[0002] For imported aquatic organisms (which can specifically refer to designated imported ornamental fish), they need to be isolated and cultured in a designated area for a period of time before they can safely enter the market for trading operations; during the isolation and culture process, the health status of the aquatic organisms cultured in the designated area of the isolation field cannot be effectively evaluated through traditional video monitoring methods, and corresponding management personnel need to regularly patrol and record in the isolation farm; in this way, the cost of isolation and culture needs to be increased, and the workload of management personnel for patrol and record also needs to be increased; currently, there is an urgent need for a method to realize online monitoring of the health of target aquatic organisms in the isolation farm to solve the above problems. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides a method for monitoring the health of aquatic organisms in an aquatic organism isolation field and related devices, which realizes image acquisition of target aquatic organisms through remote control, and realizes health analysis of the target aquatic organisms through the acquired target image data, without the need for corresponding staff to regularly patrol, reducing the patrol cost and enabling real-time mastery of the health status of the target aquatic organisms.

[0004] To solve the above technical problems, an embodiment of the present invention provides a method for monitoring the health of aquatic organisms in an aquatic organism isolation field, which is applied to an aquatic organism isolation field. An underwater robot carrying an underwater camera is arranged in the aquatic organism isolation field, and the underwater robot is communicatively connected to a remote server; the method includes:

[0005] After the underwater robot moves in the water to a preset position relative to the target aquatic organism, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on an image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server;

[0006] After receiving the target image data corresponding to the target aquatic organism, the remote server performs health analysis processing on the target aquatic organism using the target image data, and pushes the health analysis result to the monitoring and management terminal.

[0007] Optionally, before the step of after the underwater robot moves in the water to a preset position relative to the target aquatic organism, it further includes:

[0008] The remote server sends the image acquisition instruction to the underwater robot, and when the underwater robot receives the image acquisition instruction, it performs positioning processing on the area where the target aquatic organism is located to obtain positioning coordinate data, where the positioning coordinate data is data corresponding to a three-dimensional coordinate system established with a given position in the aquatic organism isolation field as the origin;

[0009] The preset position of the underwater robot relative to the target aquatic organism is determined based on the positioning coordinate data, and the underwater robot is controlled to move to the preset position.

[0010] Optionally, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server, including:

[0011] The underwater robot adjusts the focus of the underwater camera to align with the target aquatic organism, and controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction to obtain target image data corresponding to the target aquatic organism;

[0012] The underwater robot transmits the target image data corresponding to the target aquatic organism back to the remote server based on the communication connection.

[0013] Optionally, the using the target image data to perform health analysis on the target aquatic organism includes:

[0014] Recognize and process the posture of the target aquatic organism based on the target image data to obtain posture data corresponding to the target aquatic organism;

[0015] Performing a first health analysis based on the posture data corresponding to the target aquatic organism to obtain a first health analysis result;

[0016] Performing image target segmentation processing on the target aquatic organism in the target image data on the remote server to obtain segmented image data of the target aquatic organism in the target image data;

[0017] A second health analysis is performed on the target aquatic organism based on the target aquatic organism segmented image data to obtain a second health analysis result.

[0018] Optionally, the identifying and processing the posture of the target aquatic organism based on the target image data to obtain posture data corresponding to the target aquatic organism includes:

[0019] Input the target image data into the MaskR-CNN instance segmentation model in the remote server for image target extraction processing to obtain the target bounding box, the initial mask image, and the pixel background probability images in the target image data;

[0020] Based on the Grabcut image segmentation algorithm, use the target bounding box, the initial mask image, and the pixel background probability images to perform target aquatic organism segmentation processing on the target image data to obtain the segmentation image data of the target aquatic organism;

[0021] In the remote server, retrieve the camera parameters of the underwater camera, and based on the camera parameters, convert the segmentation image data of the target aquatic organism into depth image data corresponding to the segmentation image data. The camera parameters include the focal length parameter and the pose of the camera at the current viewpoint;

[0022] Generate three-dimensional point cloud data corresponding to the target aquatic organism based on the combination of the depth image data and the camera parameters, and perform three-dimensional reconstruction processing using the three-dimensional point cloud data to obtain a three-dimensional target model corresponding to the target aquatic organism;

[0023] Based on the three-dimensional target model corresponding to the target aquatic organism, perform pose recognition processing on the target aquatic organism to obtain pose data corresponding to the target aquatic organism.

[0024] Optionally, the performing image target segmentation processing on the target aquatic organism in the target image data on the remote server to obtain the segmentation image data of the target aquatic organism in the target image data includes:

[0025] On the remote server, call the level set segmentation algorithm for multi-gray targets to perform image target segmentation processing on the target aquatic organism in the target image data to obtain the initial segmentation image data of the target aquatic organism in the target image data;

[0026] On the remote server, call the edge detection algorithm to perform image target edge detection processing on the target aquatic organism in the target image data to obtain the edge detection data of the target aquatic organism in the target image data;

[0027] Use the edge detection data of the target aquatic organism in the target image data to perform edge fitting processing on the initial segmentation image data of the target aquatic organism in the target image data to form the segmentation image data of the target aquatic organism in the target image data.

[0028] Optionally, the performing second health degree analysis processing on the target aquatic organism based on the segmentation image data of the target aquatic organism to obtain a second health degree analysis result includes:

[0029] Perform binary feature extraction processing on the target aquatic organism segmentation image data based on the iterative image optimal threshold method to obtain the binary feature data corresponding to the target aquatic organism segmentation image data;

[0030] Perform a second health analysis process on the target aquatic organism by using the binary feature data corresponding to the target aquatic organism segmentation image data to obtain a second health analysis result.

[0031] In addition, an embodiment of the present invention further provides an aquatic organism health monitoring device in an aquatic organism isolation field, which is applied to the aquatic organism isolation field. An underwater robot carrying an underwater camera is arranged in the aquatic organism isolation field, and the underwater robot is communicatively connected to a remote server; the device includes:

[0032] An image acquisition module: configured to, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on an image acquisition instruction, and transmit the obtained target image data corresponding to the target aquatic organism back to the remote server;

[0033] A health analysis module: configured to, after the remote server receives the target image data corresponding to the target aquatic organism, perform health analysis processing on the target aquatic organism by using the target image data, and push the health analysis result to a monitoring and management terminal.

[0034] In addition, an embodiment of the present invention further provides a remote server, which includes a processor and a memory. The processor runs a computer program or code stored in the memory to implement the aquatic organism health monitoring method as described in any one of the above.

[0035] In addition, an embodiment of the present invention further provides a computer-readable storage medium for storing a computer program or code, which, when executed by a processor, implements the aquatic organism health monitoring method as described in any one of the above.

[0036] In an embodiment of the present invention, after the underwater robot moves in water to a preset position relative to the target aquatic organism, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on an image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; after receiving the target image data corresponding to the target aquatic organism, the remote server uses the target image data to perform health analysis processing on the target aquatic organism, and pushes the health analysis result to the monitoring and management terminal; realizing image acquisition of the target aquatic organism through remote control, and realizing the health analysis of the target aquatic organism through the acquired target image data, without the need for corresponding staff to conduct regular inspections, reducing inspection costs and being able to grasp the health status of the target aquatic organism in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0038] Figure 1 It is a schematic flowchart of the method for monitoring the health of aquatic organisms in the aquatic organism isolation field in the embodiment of the present invention;

[0039] Figure 2 It is a schematic flowchart of the health analysis processing step in the embodiment of the present invention;

[0040] Figure 3 It is a schematic flowchart of the process of recognizing the posture of the target aquatic organism in the embodiment of the present invention;

[0041] Figure 4 It is a schematic flowchart of the image target segmentation processing in the embodiment of the present invention;

[0042] Figure 5 It is a schematic diagram of the structural composition of the device for monitoring the health of aquatic organisms in the aquatic organism isolation field in the embodiment of the present invention;

[0043] Figure 6 It is a schematic diagram of the structural composition of the remote server in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] For example, see Figure 1 , Figure 1 It is a flow chart of a method for monitoring the health of aquatic organisms in an aquatic organism isolation field according to an embodiment of the present invention.

[0046] like Figure 1 As shown, a method for monitoring the health of aquatic organisms in an aquatic organism isolation field is applied to the aquatic organism isolation field, wherein an underwater robot equipped with an underwater camera is arranged in the aquatic organism isolation field, and the underwater robot is communicatively connected with a remote server; the method comprises:

[0047] S101: After the underwater robot moves to a preset position relative to a target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on an image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server;

[0048] In the specific implementation process of the present invention, before the step after the underwater robot moves in the water to a preset position relative to the target aquatic organism, it also includes: the remote server sends the image acquisition instruction to the underwater robot, and when the underwater robot receives the image acquisition instruction, it performs positioning processing on the area where the target aquatic organism is located to obtain positioning coordinate data, and the positioning coordinate data is data corresponding to a three-dimensional coordinate system established with a given position in the aquatic organism isolation field as the origin; based on the positioning coordinate data, the preset position of the underwater robot relative to the target aquatic organism is determined, and the underwater robot is controlled to move to the preset position.

[0049] Furthermore, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server, including: the underwater robot adjusts the focal length of the underwater camera to aim at the target aquatic organism, and controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction to obtain the target image data corresponding to the target aquatic organism; the underwater robot transmits the target image data corresponding to the target aquatic organism back to the remote server based on the communication connection.

[0050] Specifically, this aquatic organism health monitoring is applied to an aquatic organism quarantine facility. That is, after imported ornamental fish are imported, they need to be quarantined and cultured for a period of time before they can safely enter the market, to prevent imported ornamental fish from carrying viruses or bacteria from their place of origin and entering, causing serious losses to local fish. Therefore, an underwater robot is installed in the aquatic organism quarantine facility, and an underwater camera is carried on the underwater robot. This underwater camera can take pictures of the target aquatic organism in the water, that is, perform image acquisition and processing. At the same time, the underwater robot is communicatively connected to a remote server and receives the control of the remote server.

[0051] That is, after the remote server sends an image acquisition instruction to the underwater robot, when the underwater robot receives the image acquisition instruction, it first needs to perform a positioning process on the area where the target aquatic organism is located, and then obtain the corresponding position coordinate data. The positioning coordinate data is the data corresponding to a three-dimensional coordinate system established with a given position in the aquatic organism quarantine facility as the origin. The image acquisition instruction clearly specifies the target aquatic organism (i.e., the imported ornamental fish cultured in this aquatic organism quarantine facility); and the preset position of the underwater robot relative to the target aquatic organism is determined through this positioning coordinate data. The preset position is the position where the underwater robot needs to stay when controlling the underwater camera to perform image acquisition on the target aquatic organism. At this time, it is necessary to control the underwater robot to move to this preset position.

[0052] After the underwater robot moves to the preset position, it will adjust the angle and focal length of the underwater camera, so that the underwater camera can be aligned with the target aquatic organism, and when image acquisition is performed, an image that meets the processing requirements can be acquired. After adjusting the angle and focal length of the underwater camera lens, the underwater robot will control the underwater camera to perform image acquisition processing on the target aquatic organism through the image acquisition instruction, so as to acquire the target image data corresponding to the target aquatic organism; and after acquiring the target image data, the underwater robot will transmit the target image data corresponding to the target aquatic organism back to the remote server through the communication connection; in this way, subsequent health analysis of the target aquatic organism can be carried out on the remote server, so that the powerful computing power of the remote server can be utilized to analyze the image, and thus the health of the target aquatic organism can be analyzed.

[0053] S102: After the remote server receives the target image data corresponding to the target aquatic organism, it uses the target image data to perform a health analysis process on the target aquatic organism and pushes the health analysis result to the monitoring and management terminal.

[0054] In the specific implementation process of the present invention, after the remote server receives the target image data corresponding to the target aquatic organism, it is necessary to call the corresponding algorithm to perform a health analysis process on the target aquatic organism using the target image data, and after analyzing the health analysis result of the target aquatic organism, push the health analysis result to the monitoring and management end; so that the regulatory user can understand the health status of the target aquatic organism cultured in the aquatic organism isolation field in real time.

[0055] For details, reference can be made to Figure 2 , Figure 2 which is a schematic flow chart of the health analysis process steps in an embodiment of the present invention.

[0056] As Figure 2 shown, the process of performing a health analysis process on the target aquatic organism using the target image data includes:

[0057] S1021: Identify the posture of the target aquatic organism based on the target image data to obtain the posture data corresponding to the target aquatic organism;

[0058] In the specific implementation process of the present invention, first, it is necessary to perform three-dimensional reconstruction on the target aquatic organism using the target image data, and then determine the corresponding posture data through the three-dimensionally reconstructed target aquatic organism.

[0059] For details, reference can be made to Figure 3 , Figure 3 which is a schematic flow chart of the process of identifying the posture of the target aquatic organism in an embodiment of the present invention.

[0060] As Figure 3 shown, the process of identifying the posture of the target aquatic organism based on the target image data to obtain the posture data corresponding to the target aquatic organism includes:

[0061] S10211: Input the target image data into the MaskR-CNN instance segmentation model in the remote server for image target extraction processing to obtain the target bounding box, initial mask image, and each pixel background probability image in the target image data;

[0062] Specifically, call the MaskR-CNN instance segmentation model in the remote server, that is, input the target image data into the MaskR-CNN instance segmentation model for image target extraction processing, so as to obtain the target bounding box, initial mask image, and each pixel background probability image in the target image data.

[0063] Among them, the MaskR-CNN instance segmentation model is a powerful object detection and instance segmentation model that combines the object detection ability of Faster R-CNN and the instance segmentation ability of Mask R-CNN; it mainly includes a backbone network, a region proposal network, RoI Align, a classifier, a bounding box regressor, and a mask branch.

[0064] S10212: Based on the Grabcut image segmentation algorithm, use the target bounding box, the initial mask image, and the background probability image of each pixel to perform target aquatic organism segmentation processing on the target image data to obtain the segmentation image data of the target aquatic organism;

[0065] Specifically, after obtaining the target bounding box, the initial mask image, and the background probability image of each pixel in the target image data, it is necessary to call the Grabcut image segmentation algorithm in the remote server to perform the image segmentation operation, that is, use the Grabcut image segmentation algorithm to perform target aquatic organism segmentation on the target image data according to the target bounding box, the initial mask image, and the background probability image of each pixel in the target image data, so as to obtain the segmentation image data of the target aquatic organism; specifically, first use the target bounding box and the initial mask image to perform an initial segmentation on the target image data to obtain an initial segmentation result; then use the corresponding formula combined with the background probability image of each pixel for fine segmentation to obtain the segmentation image data of the target aquatic organism; the formula is as follows:

[0066] U(p) = {p i |V(p i ) + W(p i ) - μ > 0, i = 1, 2,..., N}

[0067] Among them, U(p) is the set of pixels determined as the background; μ is the threshold for background division, default set to 0.6; V(p i ) is the proportion of p i in the gray histogram statistics; W(p i ) is the probability that the pixel is predicted as a background pixel after sigmod activation in the MaskR-CNN instance segmentation model; then the pixel set U(ρ) is used as the background to continue to optimize in the Grabcut image segmentation algorithm, so as to avoid the problem of incomplete segmentation caused by factors such as shadows during the processing.

[0068] S10213: Retrieve the camera parameters of the underwater camera in the remote server, and based on the camera parameters, convert the segmentation image data of the target aquatic organism into depth image data corresponding to the segmentation image data, where the camera parameters include the focal length parameter and the pose of the camera at the current viewing point;

[0069] Specifically, the camera parameters of the underwater camera are retrieved from the remote server, where the camera parameters include the focal length parameter and the pose of the camera at the current viewpoint. Then, a perspective camera is configured in Blender software (3D graphics image software) using the focal length parameter and the pose of the camera at the current viewpoint in the camera parameters, and the segmented image data of the target aquatic organism is converted into depth image data corresponding to the segmented image data.

[0070] S10214: Generate three-dimensional point cloud data corresponding to the target aquatic organism based on the combination of the depth image data and the camera parameters, and perform three-dimensional reconstruction processing using the three-dimensional point cloud data to obtain a three-dimensional target model corresponding to the target aquatic organism;

[0071] Specifically, after obtaining the depth image data, the external parameter matrix T in the camera projection transformation will be calculated by combining the camera parameters with the camera transformation formula. The depth image data is adjusted from the X-Y plane of the world coordinate space to a position perpendicular to the connection line of (x, y, z) and (0, 0, 0) and at a distance of the focal length from (x, y, z) by combining the external parameter matrix T. Then, rays are drawn through the pixels with non-zero depth values in the depth image data from (x, y, z), and three-dimensional space points with the distance from the ray to (x, y, z) equal to the current depth value are added to the target point cloud at the current viewpoint, thereby obtaining the corresponding three-dimensional point cloud data.

[0072] After obtaining the three-dimensional point cloud data, use the three-dimensional point cloud data to perform three-dimensional reconstruction processing, thereby obtaining a three-dimensional target model corresponding to the target aquatic organism.

[0073] S10215: Perform pose recognition processing on the pose of the target aquatic organism based on the three-dimensional target model corresponding to the target aquatic organism to obtain pose data corresponding to the target aquatic organism.

[0074] Specifically, after obtaining the three-dimensional target model corresponding to the target aquatic organism, perform similarity matching between the three-dimensional target model and the three-dimensional models corresponding to each pose data stored in the database, select the three-dimensional model with the highest similarity as the matching model, and use the pose data corresponding to the three-dimensional model with the highest similarity as the pose data corresponding to the target aquatic organism.

[0075] S1022: Perform the first health analysis processing based on the pose data corresponding to the target aquatic organism to obtain the first health analysis result;

[0076] Specifically, after obtaining the pose data corresponding to the target aquatic organism, perform the first health analysis processing through this pose data, that is, retrieve in the first health database through this pose data, and retrieve the health data corresponding to this pose data as the first health analysis result.

[0077] S1023: performing image target segmentation processing on the target aquatic organism in the target image data on the remote server to obtain segmented image data of the target aquatic organism in the target image data;

[0078] Specifically, it is necessary to perform image target segmentation processing on the target aquatic organisms in the target image data on the remote server to obtain the target aquatic organism segmentation image data in the target image data; in the image target segmentation processing in this step, the main purpose is to be able to accurately and completely segment the target image corresponding to the target aquatic organisms in the target image data.

[0079] See also Figure 4 , Figure 4 It is a flowchart of the image target segmentation process in an embodiment of the present invention.

[0080] like Figure 4 As shown, the target aquatic organisms in the target image data are subjected to image target segmentation processing on the remote server to obtain the target aquatic organism segmented image data in the target image data, including:

[0081] S10231: Calling a level set segmentation algorithm for multiple grayscale targets on the remote server to perform image target segmentation processing on the target aquatic organisms in the target image data, and obtaining initial segmented image data of the target aquatic organisms in the target image data;

[0082] Specifically, in this embodiment, the remote server calls the level set segmentation algorithm of multiple grayscale targets to implement image target segmentation processing on the target aquatic organisms in the target image data, thereby obtaining initial segmentation image data of the target aquatic organisms in the target image data.

[0083] In the target image data, there may be multiple targets with different gray values ​​and shapes; and the target of interest is only one or several of these multiple targets; therefore, an external energy term containing image gradient information is established as follows:

[0084] E g =λ 1 ∫∫ inside(C) |gg 0 | 2 dx dy+λ 2 ∫∫ outside(C) |gg 1 | 2 dx dy;

[0085] in,

[0086] When segmenting an image by taking into account both local and overall information, the energy function is constructed as follows:

[0087] E = μE p (C) + E out (C) = μE p (C) + λL g (C) + νA g (C) + E g ;

[0088] Introducing the level set method, the expression of the level set segmentation algorithm for multi - gray - level objects is:

[0089]

[0090] where λ, ν, λ 1 , λ 2 , μ are integer constants; inside(C) is the inside of the target curve C; outside(C) is the outside of the target curve C; Ω is the entire image domain of the target image data; H ∈ (φ) is the Heaviside function; L g (C) is the length of the contour enclosed by the curve C; δ ε (φ) is the Dirac function; g is the edge - localization function; A g (C) is the area of the region enclosed by the curve C; E out (C) is the energy term outside the curve C; E p (C) is the internal energy term of the curve C; The value of

[0091] Performing an initial segmentation process on the image through the above - formed level set segmentation algorithm for multi - gray - level objects. This method can accurately segment the edge of the target, and the segmentation effect is good; however, although all gray - level objects can be segmented, there will be a phenomenon of edge blurring, so further processing is required.

[0092] S10232: Invoke an edge detection algorithm on the remote server to perform image target edge detection processing on the target aquatic organisms in the target image data, and obtain the target aquatic organisms edge detection data in the target image data;

[0093] Specifically, it is necessary to invoke an edge detection algorithm on the remote server to perform image target edge detection processing on the target aquatic organisms in the target image data, so as to accurately obtain the target aquatic organisms edge detection data in the target image data.

[0094] S10233: Use the target aquatic organisms edge detection data in the target image data to perform edge fitting processing on the target aquatic organisms initial segmentation image data in the target image data, and form the target aquatic organisms segmentation image data in the target image data.

[0095] Specifically, in order to more accurately segment the target aquatic organism segmentation image data from the target image data, it is necessary to perform edge fitting processing on the initial segmentation image data of the target aquatic organism in the target image data using the edge detection data of the target aquatic organism in the target image data, so as to solve the problem of edge blurring that may occur during initial segmentation, and thus more accurately obtain the segmentation edge, and then the target aquatic organism segmentation image data in the target image data can be accurately formed according to this.

[0096] S1024: Perform a second health analysis process on the target aquatic organism based on the target aquatic organism segmentation image data to obtain a second health analysis result.

[0097] In the specific implementation process of the present invention, the performing a second health analysis process on the target aquatic organism based on the target aquatic organism segmentation image data to obtain a second health analysis result includes: performing binary feature extraction processing on the target aquatic organism segmentation image data based on the iterative image optimal threshold method to obtain binary feature data corresponding to the target aquatic organism segmentation image data; using the binary feature data corresponding to the target aquatic organism segmentation image data to perform a second health analysis process on the target aquatic organism to obtain a second health analysis result.

[0098] Specifically, in this embodiment, the iterative image optimal threshold algorithm needs to divide the grayscale image corresponding to the target aquatic organism segmentation image data into several sub-images, and each sub-image needs to be further segmented until it cannot be segmented any further; the algorithm of the iterative image optimal threshold is as follows:

[0099] It is necessary to extract the minimum gray value Z in the image min and the maximum gray value Z max , and then set the initial threshold value to: T 0 =(Z min +Z max ) / 2; then according to the threshold T k the image is segmented into two parts: the target and the background, and the average gray values Z 1 and Z 2 of the two parts are calculated; that is:

[0100]

[0101] where Z(i,j) is the gray value of the point (i,j) on the image, and N(i,j) is the weight coefficient of the point (i,j); generally, N(i,j) is assigned a value of 1.

[0102] Finally, calculate the updated threshold: T k+1 =(Z 1 +Z2 ) / 2; at T k = T k+1 When it is, then end, otherwise k ← k + 1, return to the above steps and repeat, then the optimal threshold of the iterative image can be obtained.

[0103] After obtaining the optimal threshold of the iterative image, perform binary feature extraction processing on the target aquatic organism segmentation image data through the optimal threshold of the iterative image, so as to obtain the binary feature data corresponding to the target aquatic organism segmentation image data; finally, perform a second health analysis process on the target aquatic organism by using the binary feature data corresponding to the target aquatic organism segmentation image data to obtain the second health analysis result; the second health analysis process here is to use the binary feature data to perform index matching in the feature database, and when the corresponding similar features are indexed, take the health degree indexed by the similar features as the second health analysis result.

[0104] Then the remote server pushes the first health analysis result and the second health analysis result to the monitoring and management terminal.

[0105] In the embodiment of the present invention, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; after receiving the target image data corresponding to the target aquatic organism, the remote server performs health analysis processing on the target aquatic organism by using the target image data, and pushes the health analysis result to the monitoring and management terminal; it realizes image acquisition of the target aquatic organism through remote control, and realizes the health analysis of the target aquatic organism through the acquired target image data, without the need for corresponding staff to conduct regular inspections, reducing the inspection cost and being able to grasp the health status of the target aquatic organism in real time.

[0106] Embodiment 2, please refer to Figure 5 , Figure 5 is a schematic structural composition diagram of the aquatic organism health monitoring device in the aquatic organism isolation field in the embodiment of the present invention.

[0107] As Figure 5 shown, an aquatic organism health monitoring device in an aquatic organism isolation field is applied to an aquatic organism isolation field. An underwater robot carrying an underwater camera is arranged in the aquatic organism isolation field, and the underwater robot is communicatively connected to a remote server; the device includes:

[0108] Image acquisition module 301: used for, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server;

[0109] In the specific implementation process of the present invention, before the step after the underwater robot moves in the water to a preset position relative to the target aquatic organism, it also includes: the remote server sends the image acquisition instruction to the underwater robot, and when the underwater robot receives the image acquisition instruction, it performs positioning processing on the area where the target aquatic organism is located to obtain positioning coordinate data, and the positioning coordinate data is data corresponding to a three-dimensional coordinate system established with a given position in the aquatic organism isolation field as the origin; based on the positioning coordinate data, the preset position of the underwater robot relative to the target aquatic organism is determined, and the underwater robot is controlled to move to the preset position.

[0110] Furthermore, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server, including: the underwater robot adjusts the focal length of the underwater camera to aim at the target aquatic organism, and controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction to obtain the target image data corresponding to the target aquatic organism; the underwater robot transmits the target image data corresponding to the target aquatic organism back to the remote server based on the communication connection.

[0111] Specifically, the aquatic organism health monitoring is applied to aquatic organism isolation fields, that is, after imported ornamental fish are imported, they need to be isolated and cultivated for a period of time before they can safely enter the market to prevent the imported ornamental fish from carrying viruses or bacteria from their place of origin, which may cause local fish to be infected and cause serious losses. Therefore, an underwater robot needs to be set up in the aquatic organism isolation field, and an underwater camera is mounted on the underwater robot. The underwater camera can perform photo processing, that is, image acquisition processing, on the target aquatic organisms in the water. At the same time, the underwater robot is communicated with a remote server, and the underwater robot is controlled by the remote server.

[0112] That is, after the remote server sends an image acquisition instruction to the underwater robot, when the underwater robot receives the image acquisition instruction, it first needs to locate the area where the target aquatic organism is located, and then obtain the corresponding position coordinate data. The positioning coordinate data is the data corresponding to the three-dimensional coordinate system established with a given position in the aquatic organism isolation field as the origin; the image acquisition instruction clearly states the target aquatic organism (i.e., the imported ornamental fish cultured in the aquatic organism isolation field); and the preset position of the underwater robot relative to the target aquatic organism is determined through the positioning coordinate data. The preset position is the position where the underwater robot needs to stay when controlling the underwater camera to perform image acquisition on the target aquatic organism; at this time, it is necessary to control the underwater robot to move to the preset position.

[0113] After the underwater robot moves to the preset position, it will adjust the angle and focal length of the underwater camera so that the underwater camera can be aligned with the target aquatic organism, and when image acquisition is performed, an image that meets the processing requirements can be acquired. After adjusting the angle and focal length of the underwater camera lens, the underwater robot will control the underwater camera to perform image acquisition processing on the target aquatic organism through the image acquisition instruction, so as to acquire the target image data corresponding to the target aquatic organism; and after acquiring the target image data, the underwater robot will transmit the target image data corresponding to the target aquatic organism back to the remote server through the communication connection; in this way, subsequent health analysis of the target aquatic organism can be carried out on the remote server, and the powerful computing power of the remote server can be utilized to analyze the image, so as to analyze the health of the target aquatic organism.

[0114] Health analysis module 302: It is used for the remote server to perform health analysis processing on the target aquatic organism by using the target image data after receiving the target image data corresponding to the target aquatic organism, and push the health analysis result to the monitoring and management terminal.

[0115] In the specific implementation process of the present invention, after the remote server receives the target image data corresponding to the target aquatic organism, it is necessary to call the corresponding algorithm to perform health analysis processing on the target aquatic organism by using the target image data, and after analyzing the health analysis result of the target aquatic organism, push the health analysis result to the monitoring and management terminal; so that the supervision user can understand the health status of the target aquatic organism cultured in the aquatic organism isolation field in real time.

[0116] Specifically, reference can be made to Figure 2 , Figure 2 which is a schematic flowchart of the health analysis processing steps in the embodiment of the present invention.

[0117] As Figure 2 shown, the using the target image data to perform health analysis processing on the target aquatic organism includes:

[0118] S1021: Identify and process the pose of the target aquatic organism based on the target image data to obtain the pose data corresponding to the target aquatic organism;

[0119] In the specific implementation process of the present invention, first, it is necessary to use the target image data to perform three-dimensional reconstruction on the target aquatic organism, and then determine the corresponding pose data through the three-dimensionally reconstructed target aquatic organism.

[0120] Specifically, reference can be made to Figure 3 , Figure 3 which is a schematic flowchart of the identification and processing of the pose of the target aquatic organism in the embodiments of the present invention.

[0121] As Figure 3 shown, the identifying and processing of the pose of the target aquatic organism based on the target image data to obtain the pose data corresponding to the target aquatic organism includes:

[0122] S10211: Input the target image data into the MaskR-CNN instance segmentation model in the remote server for image target extraction processing to obtain the target bounding box, initial mask image, and each pixel background probability image in the target image data;

[0123] Specifically, call the MaskR-CNN instance segmentation model in the remote server, that is, input the target image data into the MaskR-CNN instance segmentation model for image target extraction processing, so as to obtain the target bounding box, initial mask image, and each pixel background probability image in the target image data.

[0124] Among them, the MaskR-CNN instance segmentation model is a powerful object detection and instance segmentation model, which combines the object detection ability of Faster R-CNN and the instance segmentation ability of Mask R-CNN; it mainly includes a backbone network, a region proposal network, RoI Align, a classifier, a bounding box regressor, and a mask branch.

[0125] S10212: Use the Grabcut image segmentation algorithm to perform target aquatic organism segmentation processing on the target image data based on the target bounding box, initial mask image, and each pixel background probability image to obtain the segmented image data of the target aquatic organism;

[0126] Specifically, after obtaining the target bounding box, the initial mask image, and the pixel background probability images in the target image data, it is necessary to call the Grabcut image segmentation algorithm in the remote server to perform image segmentation operations. That is, use the Grabcut image segmentation algorithm to perform target aquatic organism segmentation on the target image data based on the target bounding box, the initial mask image, and the pixel background probability images in the target image data, so as to obtain the segmented image data of the target aquatic organism. Specifically, first use the target bounding box and the initial mask image to perform an initial segmentation on the target image data to obtain an initial segmentation result. Then, use the corresponding formula in combination with the pixel background probability images for fine segmentation to obtain the segmented image data of the target aquatic organism. The formula is as follows:

[0127] U(p) = {p i |V(p i ) + W(p i ) - μ > 0, i = 1, 2, …, N}

[0128] Among them, U(p) is the set of determined background pixels; μ is the threshold for background division, and the default setting is 0.6; V(p i ) is the proportion of p i in the grayscale histogram statistics; W(ρ i ) is the probability that the pixel is predicted as a background pixel after being activated by sigmod in the MaskR-CNN instance segmentation model; then the pixel set U(ρ) is used as the background to continue optimization in the Grabcut image segmentation algorithm, so as to avoid the problem of incomplete segmentation caused by factors such as shadows during the processing.

[0129] S10213: Retrieve the camera parameters of the underwater camera in the remote server, and based on the camera parameters, convert the segmented image data of the target aquatic organism into depth image data corresponding to the segmented image data. The camera parameters include the focal length parameter and the pose of the camera at the current viewing point;

[0130] Specifically, retrieve the camera parameters of the underwater camera in the remote server, where the camera parameters include the focal length parameter and the pose of the camera at the current viewing point; then configure a perspective camera in the Blender software (3D graphics image software) using the focal length parameter and the pose of the camera at the current viewing point in the camera parameters, and convert the segmented image data of the target aquatic organism into depth image data corresponding to the segmented image data.

[0131] S10214: Generate three-dimensional point cloud data corresponding to the target aquatic organism based on the combination of the depth image data and the camera parameters, and perform three-dimensional reconstruction processing using the three-dimensional point cloud data to obtain a three-dimensional target model corresponding to the target aquatic organism;

[0132] Specifically, after obtaining the depth image data, the external parameter matrix T in the camera projection transformation will be calculated by combining the camera parameters with the camera transformation formula. The depth image data will be adjusted from the X-Y plane of the world coordinate space to a position perpendicular to the connection line of (x, y, z) and (0, 0, 0) and at a distance equal to the focal length from (x, y, z). Then, a ray passing through the pixels with non-zero depth values in the depth image data will be drawn through (x, y, z), and a three-dimensional space point with a distance from the ray to (x, y, z) equal to the current depth value will be added to the target point cloud under the current viewpoint, thereby obtaining the corresponding three-dimensional point cloud data.

[0133] After obtaining the three-dimensional point cloud data, the three-dimensional point cloud data is used for three-dimensional reconstruction processing to obtain the three-dimensional target model corresponding to the target aquatic organism.

[0134] S10215: Based on the three-dimensional target model corresponding to the target aquatic organism, identify the posture of the target aquatic organism to obtain the posture data corresponding to the target aquatic organism.

[0135] Specifically, after obtaining the three-dimensional target model corresponding to the target aquatic organism, the three-dimensional model is used to perform similarity matching with the three-dimensional models corresponding to each posture data stored in the database. The three-dimensional model with the highest similarity is selected as the matching model, and the posture data corresponding to the three-dimensional model with the highest similarity of this pixel point is used as the posture data corresponding to the target aquatic organism.

[0136] S1022: Based on the posture data corresponding to the target aquatic organism, perform the first health analysis to obtain the first health analysis result;

[0137] Specifically, after obtaining the posture data corresponding to the target aquatic organism, the first health analysis is performed through this posture data, that is, the posture data is retrieved in the first health database, and the health data corresponding to this posture data is retrieved as the first health analysis result.

[0138] S1023: On the remote server, perform image target segmentation on the target aquatic organism in the target image data to obtain the target aquatic organism segmentation image data in the target image data;

[0139] Specifically, on the remote server, it is necessary to perform image target segmentation on the target aquatic organism in the target image data to obtain the target aquatic organism segmentation image data in the target image data; in the image target segmentation process of this step, the main purpose is to accurately and completely segment the target image corresponding to the target aquatic organism in the target image data.

[0140] Please refer to Figure 4 , Figure 4It is a flowchart of the image target segmentation process in an embodiment of the present invention.

[0141] like Figure 4 As shown, the target aquatic organisms in the target image data are subjected to image target segmentation processing on the remote server to obtain the target aquatic organism segmented image data in the target image data, including:

[0142] S10231: Calling a level set segmentation algorithm for multiple grayscale targets on the remote server to perform image target segmentation processing on the target aquatic organisms in the target image data, and obtaining initial segmented image data of the target aquatic organisms in the target image data;

[0143] Specifically, in this embodiment, the remote server calls the level set segmentation algorithm of multiple grayscale targets to implement image target segmentation processing on the target aquatic organisms in the target image data, thereby obtaining initial segmentation image data of the target aquatic organisms in the target image data.

[0144] In the target image data, there may be multiple targets with different gray values ​​and shapes; and the target of interest is only one or several of these multiple targets; therefore, an external energy term containing image gradient information is established as follows:

[0145] E g =λ 1 ∫∫ inside(C) |gg 0 | 2 dx dy+λ 2 ∫∫ outside(C) |gg 1 | 2 dx dy;

[0146] in,

[0147] When segmenting an image by taking into account both local and overall information, the energy function is constructed as follows:

[0148] E=μE p (C)+E out (C) = μE p (C)+λL g (C)+vA g (C)+E g ;

[0149] The level set method is introduced to form the expression of the level set segmentation algorithm for multi-grayscale targets:

[0150]

[0151] Among them, λ, v, λ1 , λ 2 , μ are integer constants; inside(C) is the interior of the target curve C; outside(C) is the exterior of the target curve C; Ω is the entire image domain of the target image data; H ∈ (φ) is the Heaviside function; L g (C) is the length of the contour enclosed by the curve C; δ ε (φ) is the Dirac function; g is the edge localization function; A g (C) is the area of the region enclosed by the curve C; e out (c) is the energy term outside the curve C; E p (C) is the internal energy term of the curve C; The value approaches 1.

[0152] The initial segmentation process of the image is performed by the above-mentioned level set segmentation algorithm for multi-gray targets. This method can accurately segment the edges of the target, and the segmentation effect is good; however, although all gray targets can be segmented, there will be a phenomenon of edge blurring, so further processing is required.

[0153] S10232: Invoke an edge detection algorithm on the remote server to perform image target edge detection processing on the target aquatic organisms in the target image data, and obtain the target aquatic organisms edge detection data in the target image data;

[0154] Specifically, it is necessary to invoke an edge detection algorithm on the remote server to perform image target edge detection processing on the target aquatic organisms in the target image data, so as to accurately obtain the target aquatic organisms edge detection data in the target image data.

[0155] S10233: Use the target aquatic organisms edge detection data in the target image data to perform edge fitting processing on the target aquatic organisms initial segmentation image data in the target image data, and form the target aquatic organisms segmentation image data in the target image data.

[0156] Specifically, in order to more accurately segment the target aquatic organisms segmentation image data in the target image data, it is necessary to use the target aquatic organisms edge detection data in the target image data to perform edge fitting processing on the target aquatic organisms initial segmentation image data in the target image data, so as to solve the possible edge blurring phenomenon during the initial segmentation, and thus more accurately obtain the segmentation edge, that is, the target aquatic organisms segmentation image data in the target image data can be accurately formed.

[0157] S1024: Based on the target aquatic organisms segmentation image data, perform a second health analysis process on the target aquatic organisms to obtain a second health analysis result.

[0158] In the specific implementation process of the present invention, the second health degree analysis process is performed on the target aquatic organism based on the target aquatic organism segmentation image data to obtain a second health degree analysis result, including: performing binary feature extraction processing on the target aquatic organism segmentation image data based on the iterative image optimal threshold method to obtain binary feature data corresponding to the target aquatic organism segmentation image data; using the binary feature data corresponding to the target aquatic organism segmentation image data to perform a second health degree analysis process on the target aquatic organism to obtain a second health degree analysis result.

[0159] Specifically, in this embodiment, the iterative image optimal threshold algorithm needs to divide the grayscale image corresponding to the target aquatic organism segmentation image data into several sub-images, and each sub-image needs to be further segmented until it cannot be segmented any further; the algorithm of the iterative image optimal threshold is as follows:

[0160] It is necessary to extract the minimum grayscale value Z min and the maximum grayscale value Z max in the image, and then set the initial threshold value to: T 0 =(Z min +Z max ) / 2; then, according to the threshold T k , the image is segmented into two parts: the target and the background, and the average grayscale values Z 1 and Z 2 of the two parts are calculated; that is:

[0161]

[0162] where Z(i,j) is the grayscale value of the point (i,j) on the image, and N(i,j) is the weight coefficient of the point (i,j); generally, N(i,j) is assigned a value of 1.

[0163] Finally, the updated threshold is calculated: T k+1 =(Z 1 +Z 2 ) / 2; when T k =T k+1 , the process ends; otherwise, k←k + 1, and the above steps are repeated to obtain the iterative image optimal threshold.

[0164] After obtaining the optimal threshold of the iterative image, the binary feature extraction process is performed on the target aquatic organism segmentation image data through the optimal threshold of the iterative image, so as to obtain the binary feature data corresponding to the target aquatic organism segmentation image data; finally, the binary feature data corresponding to the target aquatic organism segmentation image data is used to perform the second health analysis process on the target aquatic organism, and the second health analysis result is obtained; the second health analysis process here is to use the binary feature data to perform index matching in the feature database, and when the corresponding approximate feature is indexed, the health degree indexed by the approximate feature is used as the second health analysis result.

[0165] Then the remote server pushes the first health analysis result and the second health analysis result to the monitoring and management terminal.

[0166] In the embodiment of the present invention, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; after receiving the target image data corresponding to the target aquatic organism, the remote server performs health analysis processing on the target aquatic organism by using the target image data, and pushes the health analysis result to the monitoring and management terminal; it realizes image acquisition of the target aquatic organism through remote control, and realizes the health analysis of the target aquatic organism through the acquired target image data, without corresponding staff to conduct regular inspections, reducing the inspection cost and being able to grasp the health status of the target aquatic organism in real time.

[0167] A computer-readable storage medium provided by an embodiment of the present invention, on which a computer program is stored, and when the program is executed by a processor, it implements the aquatic organism health monitoring method of any one of the above embodiments. Among them, the computer-readable storage medium includes, but is not limited to, any type of disk (including floppy disk, hard disk, optical disk, CD-ROM, and magneto-optical disk), ROM (Read-Only Memory), RAM (Random Access Memory), EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic card or optical card. That is, the storage device includes any medium that can store or transmit information in a readable form by a device (such as a computer, mobile phone), and can be a read-only memory, a magnetic disk or an optical disk, etc.

[0168] An embodiment of the present invention further provides a computer application program that runs on a computer and is used to execute the aquatic organism health monitoring method according to any one of the above embodiments.

[0169] In addition, Figure 6 It is a schematic diagram of the structural composition of the remote server in the embodiment of the present invention.

[0170] An embodiment of the present invention further provides a remote server, as Figure 6 shown. The remote server includes devices such as a processor 402, a memory 403, an input unit 404, and a display unit 405. Those skilled in the art can understand that Figure 6 the structural devices of the remote server shown do not constitute a limitation on all devices, and may include more or fewer components than shown, or combine certain components. The memory 403 can be used to store the application program 401 and each functional module. The processor 402 runs the application program 401 stored in the memory 403 to execute various functional applications and data processing of the device. The memory can be an internal memory or an external memory, or include both an internal memory and an external memory. The internal memory can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, or a random access memory. The external memory can include a hard disk, a floppy disk, a ZIP disk, a USB flash drive, a magnetic tape, etc. The memory disclosed in the present invention includes but is not limited to these types of memories. The memory disclosed in the present invention is only an example rather than a limitation.

[0171] The input unit 404 is used to receive the input of signals and the keywords input by the user. The input unit 404 can include a touch panel and other input devices. The touch panel can collect the touch operations of the user on or near it (such as the operations of the user using a finger, a stylus, or any suitable object or accessory on or near the touch panel), and drive the corresponding connection device according to a pre-set program; the other input devices can include, but are not limited to, a physical keyboard, function keys (such as play control keys, switch keys, etc.), a trackball, a mouse, a joystick, etc. The display unit 405 can be used to display the information input by the user or the information provided to the user and various menus of the terminal device. The display unit 405 can be in the form of a liquid crystal display, an organic light-emitting diode, etc. The processor 402 is the control center of the terminal device, connects various parts of the entire device through various interfaces and lines, and executes various functions and processes data by running or executing the software programs and / or modules stored in the memory 403, and calling the data stored in the memory.

[0172] As an embodiment, the remote server includes: one or more processors 402, a memory 403, and one or more applications 401, wherein the one or more applications 401 are stored in the memory 403 and configured to be executed by the one or more processors 402, and the one or more applications 401 are configured to execute the corresponding aquatic organism health monitoring method in any one of the above embodiments.

[0173] In an embodiment of the present invention, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on an image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; after receiving the target image data corresponding to the target aquatic organism, the remote server performs health analysis processing on the target aquatic organism using the target image data, and pushes the health analysis result to the monitoring and management terminal; realizing image acquisition of the target aquatic organism through remote control, and realizing the health analysis of the target aquatic organism through the acquired target image data, without the need for corresponding staff to conduct regular inspections, reducing inspection costs and being able to grasp the health status of the target aquatic organism in real time.

[0174] In addition, the above has introduced in detail a method and related device for monitoring the health of aquatic organisms in an aquatic organism isolation field provided by the embodiments of the present invention. Specific examples should have been used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for monitoring the health of aquatic organisms in an aquatic organism isolation field, characterized in that: The method is applied to an aquatic biological isolation field, in which an underwater robot carrying an underwater camera is arranged, and the underwater robot is communicatively connected with a remote server; the method comprises: After the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; After receiving the target image data corresponding to the target aquatic organism, the remote server uses the target image data to perform health analysis on the target aquatic organism and pushes the health analysis result to the monitoring management end.

2. The method for monitoring the health of aquatic organisms according to claim 1, characterized in that: Before the step of the underwater robot moving in water to a preset position relative to the target aquatic organism, the method further includes: The remote server sends the image acquisition instruction to the underwater robot, and when the underwater robot receives the image acquisition instruction, it performs positioning processing on the area where the target aquatic organism is located to obtain positioning coordinate data, where the positioning coordinate data is data corresponding to a three-dimensional coordinate system established with a given position in the aquatic organism isolation field as the origin; The preset position of the underwater robot relative to the target aquatic organism is determined based on the positioning coordinate data, and the underwater robot is controlled to move to the preset position.

3. The method for monitoring the health of aquatic organisms according to claim 1, characterized in that: The underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server, including: The underwater robot adjusts the focus of the underwater camera to aim at the target aquatic organism, and controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction to obtain target image data corresponding to the target aquatic organism; The underwater robot transmits the target image data corresponding to the target aquatic organism back to the remote server based on the communication connection.

4. The method for monitoring the health of aquatic organisms according to claim 1, characterized in that: The step of analyzing the health of the target aquatic organism using the target image data includes: Recognize and process the posture of the target aquatic organism based on the target image data to obtain posture data corresponding to the target aquatic organism; Performing a first health analysis based on the posture data corresponding to the target aquatic organism to obtain a first health analysis result; Performing image target segmentation processing on the target aquatic organism in the target image data on the remote server to obtain segmented image data of the target aquatic organism in the target image data; A second health analysis is performed on the target aquatic organism based on the target aquatic organism segmented image data to obtain a second health analysis result.

5. The method for monitoring the health of aquatic organisms according to claim 4, characterized in that: The step of performing recognition processing on the posture of the target aquatic organism based on the target image data to obtain posture data corresponding to the target aquatic organism includes: In the remote server, the target image data is input into the Mask R-CNN instance segmentation model for image target extraction processing to obtain a target bounding box, an initial mask image, and a background probability image of each pixel in the target image data; Based on the Grabcut image segmentation algorithm, the target image data is segmented into target aquatic organisms using the target bounding box, the initial mask image and the background probability image of each pixel to obtain the segmented image data of the target aquatic organisms; Retrieving camera parameters of the underwater camera on the remote server, and converting the segmented image data of the target aquatic organism into depth image data corresponding to the segmented image data based on the camera parameters, wherein the camera parameters include focal length parameters and the posture of the camera at the current viewpoint; Generate three-dimensional point cloud data corresponding to the target aquatic organism based on the combination of the depth image data and the camera parameters, and perform three-dimensional reconstruction processing using the three-dimensional point cloud data to obtain a three-dimensional target model corresponding to the target aquatic organism; The posture of the target aquatic organism is recognized based on the three-dimensional target model corresponding to the target aquatic organism to obtain posture data corresponding to the target aquatic organism.

6. The method for monitoring the health of aquatic organisms according to claim 4, characterized in that: The performing image target segmentation processing on the target aquatic organism in the target image data on the remote server to obtain the target aquatic organism segmentation image data in the target image data comprises: Calling a level set segmentation algorithm for multiple grayscale targets on the remote server to perform image target segmentation processing on the target aquatic organisms in the target image data to obtain initial segmented image data of the target aquatic organisms in the target image data; Calling an edge detection algorithm on the remote server to perform image target edge detection processing on the target aquatic organism in the target image data to obtain edge detection data of the target aquatic organism in the target image data; The target aquatic organism edge detection data in the target image data is used to perform edge fitting processing on the target aquatic organism initial segmentation image data in the target image data to form the target aquatic organism segmentation image data in the target image data.

7. The method for monitoring the health of aquatic organisms according to claim 4, characterized in that: The performing a second health analysis on the target aquatic organism based on the target aquatic organism segmented image data to obtain a second health analysis result includes: Performing binary feature extraction processing on the target aquatic organism segmentation image data based on an iterative image optimal threshold method to obtain binary feature data corresponding to the target aquatic organism segmentation image data; A second health analysis is performed on the target aquatic organism using the binary feature data corresponding to the target aquatic organism segmentation image data to obtain a second health analysis result.

8. A device for monitoring the health of aquatic organisms in an aquatic organism isolation field, characterized in that: The device is applied to an aquatic biological isolation field, in which an underwater robot carrying an underwater camera is arranged, and the underwater robot is connected to a remote server for communication; the device comprises: Image acquisition module: used for, after the underwater robot moves to a preset position relative to the target aquatic organism in the water, the underwater robot controls the underwater camera to perform image acquisition processing on the target aquatic organism based on the image acquisition instruction, and transmits the obtained target image data corresponding to the target aquatic organism back to the remote server; Health analysis module: After the remote server receives the target image data corresponding to the target aquatic organism, it uses the target image data to perform health analysis on the target aquatic organism and pushes the health analysis result to the monitoring management end.

9. A remote server, comprising a processor and a memory, characterized in that: The processor runs the computer program or code stored in the memory to implement the aquatic organism health monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium for storing a computer program or code, characterized in that: When the computer program or code is executed by a processor, the method for monitoring the health of aquatic organisms according to any one of claims 1 to 7 is implemented.

Images