Intelligent test system and method for disease resistance of shrimps based on automatic image acquisition and analysis

Abstract

This invention relates to the field of image analysis and automated testing technology, and particularly to an intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis. The system includes an automatic inspection module, a data transmission module, and an information processing module. This invention achieves a fundamental shift in shrimp disease resistance trait testing from manual observation and recording to automated and intelligent judgment. By integrating an autonomous navigation inspection robot, a multi-view image acquisition device, and a machine learning-based intelligent analysis system, a fully automated and continuously operating testing system is constructed. This system can automatically complete timed image acquisition of a large number of test individuals and use a trained YOLO model to objectively analyze shrimp posture, accurately determine survival status and mortality time, effectively overcoming the inherent defects of manual monitoring, such as high labor intensity, high subjective judgment error, and easy recording errors.

Description

Technical Field

[0001] This invention relates to the field of image analysis and automated testing technology, and in particular to an intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis. Background Technology

[0002] In the breeding of new varieties of aquatic species such as shrimp, disease resistance is a key economic trait. Regardless of whether family-based or genomic breeding techniques are used, accurate acquisition of individual or population resistance phenotypic data is essential as the basis for estimating breeding value and making mating decisions. Therefore, the reliability, efficiency, and scale of resistance testing methods directly affect the effectiveness of the breeding process. Currently, the resistance testing methods commonly used in the industry are still primarily manual. Typically, test individuals are induced to become ill through methods such as feeding poisoned bait, soaking in bacterial solutions, or injection infection. Subsequently, continuous manual observation and recording of mortality times are required as a quantitative indicator of resistance levels. This method has the following significant limitations in practical applications: heavy reliance on manpower and low work efficiency. To record mortality events promptly, experimental personnel need to be on duty for extended periods, observing all test individuals individually every 1-2 hours. For large-scale testing of hundreds to thousands of shrimp, the manpower investment is high, the workload is intense, and high-frequency monitoring is difficult to achieve. Secondly, the judgment of condition is subjective, and errors are difficult to avoid. The large number of shrimp tested and their subtle changes in condition make manual observation prone to misjudgment due to factors such as visual fatigue and differences in subjective judgment standards among researchers. For example, individuals that are near death or lying on their side may be mistakenly identified as dead, or dead individuals may be missed. This judgment error directly reduces the accuracy of resistance phenotypic data. Furthermore, manual recording is prone to errors. During batch recording, information such as individual numbers, pedigree numbers, and times of death must be manually recorded and entered, which is susceptible to confusion, omissions, or errors, affecting the reliability of subsequent data analysis and breeding evaluation.

[0003] The aforementioned problems make it difficult for traditional resistance testing methods to meet the demands of modern breeding for high-throughput, high-precision resistance phenotypic determination. Therefore, there is an urgent need to establish an automated and intelligent disease resistance trait testing system to achieve continuous, standardized, and accurate identification and recording of individual shrimp survival status, providing reliable data support for resistance breeding. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides an intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis.

[0005] To achieve the above objectives, on the one hand, the present invention provides an intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis, including an automatic inspection function module, a data transmission function module, and an information processing function module.

[0006] The automatic inspection function module includes an inspection robot module, an image acquisition module, and an individual labeling and recognition module. The image acquisition module is installed on the inspection robot module, the individual labeling and recognition module is installed at the test container, and the data transmission function module is connected to the automatic inspection function module wirelessly and to the information processing function module via a wired connection.

[0007] The inspection robot module is used for autonomous navigation and movement within the test area.

[0008] The image acquisition module is used to collect image data of individual shrimp in the test container during the inspection process.

[0009] The individual labeling and identification module is used to uniquely identify each test individual and its location.

[0010] The data transmission module is used to transmit the image data acquired by the image acquisition module to the information processing module.

[0011] The information processing module is used to receive and process the image data, identify the information of the individual labeling and identification module to determine the identity of the test individual, and determine the survival status of the shrimp individual through a machine learning model based on the image of the shrimp individual, record the time of death and statistically analyze the test results.

[0012] Furthermore, the inspection robot module includes an autonomous mobile chassis, a navigation and mapping unit, a collision avoidance unit, and an automatic charging unit.

[0013] The navigation and mapping unit is used to construct maps and plan routes for the test area.

[0014] The anti-collision unit is used to avoid obstacles during movement.

[0015] The automatic charging unit is used to enable the robot to automatically return to the charging station for charging during inspection breaks.

[0016] Furthermore, the image acquisition module includes multiple industrial cameras, a protective housing, and an internal temperature and humidity control unit.

[0017] The multiple industrial cameras are positioned at different heights to capture images of test containers at different levels of the test rack.

[0018] The protective shell is a closed cylindrical structure used for moisture and corrosion protection.

[0019] The internal temperature and humidity control unit is used to prevent lens fogging.

[0020] Furthermore, the individual labeling and identification module is a QR code label affixed above each test container compartment.

[0021] The information encoded by the QR code includes at least the test rack number, layer number, container number, and grid number; the QR code encoding is used to achieve unique identification and location of each test individual.

[0022] Furthermore, the information processing module includes an image decoding and matching unit and a state intelligent determination unit.

[0023] The image decoding and matching unit is used to decode the received video or image data, identify the QR code information, and match the individual shrimp image with its corresponding grid position.

[0024] The intelligent state determination unit is used to run a pre-trained machine learning model to analyze the matched individual shrimp images and determine their survival or death status based on their body posture.

[0025] Furthermore, the intelligent state determination unit uses the YOLO model as its machine learning model. The YOLO model is trained using a large amount of labeled live and dead shrimp image data and can determine mortality based on the shrimp's side-lying or belly-up posture.

[0026] Furthermore, the information processing function module also includes a multi-round logic judgment unit.

[0027] The multi-round logic judgment unit is used to cross-verify and finally confirm the status judgment results of the same test individual in multiple consecutive inspections, so as to eliminate possible misjudgments of single death status and further improve the accuracy of death time.

[0028] Based on the same inventive concept, this invention also provides an intelligent testing method for shrimp disease resistance traits based on automatic image acquisition and analysis, including: Step S1: By placing the automatic inspection robot in the test area and activating its control software, the robot moves between test frames to complete the environmental modeling of the walkable area and the labeling of critical path points.

[0029] Step S2: Import the QR code ID, corresponding family number, and test start time information of all test grids through the information processing function module to complete the initialization of test data.

[0030] Step S3: In the inspection robot module, the walking route is planned based on the movement sequence of the marked points, and the parameters of the inspection strategy are configured by setting the number of walks, the patrol time interval and the judgment round.

[0031] Step S4: By activating the image acquisition function of the information processing module and running the inspection program, the robot begins to perform automatic inspection according to the set strategy, and supports manual pause and resume during the inspection process.

[0032] Step S5: After each round of inspection, the results data are automatically saved and exported, and the status and death time information of each test individual are displayed on the human-machine interface to achieve real-time feedback and recording of the monitoring results.

[0033] Step S6: Through the manual review interface, the system judgment result is reviewed and corrected. After confirmation, the "clear" operation is performed to finally confirm the status of the deceased individual and block its QR code identifier, making it invalid in subsequent detection.

[0034] Step S7: The information processing module automatically calculates and dynamically updates the overall mortality curve and family survival rate statistics charts to achieve comprehensive visualization and continuous tracking of the adversarial test results.

[0035] Furthermore, in step S3, when the number of judgment rounds is set to be greater than 1, the system will logically compare the judgment results of the same test individual in consecutive rounds. Only when the individual is confirmed to be in a dead state in multiple consecutive judgments will the individual's death time be finally recorded. The final time is the time when the individual is first judged to be in a dead state.

[0036] Compared with existing technologies, this invention achieves a fundamental shift in shrimp disease resistance testing from manual observation to automated and intelligent judgment. By integrating an autonomous navigation inspection robot, a multi-view image acquisition device, and a machine learning-based intelligent analysis system, a fully automated and continuously operating testing system is constructed. This system can automatically complete timed image acquisition of a large number of test individuals and use a trained YOLO model to objectively analyze shrimp posture, accurately determining the survival status and mortality time of each test individual, effectively overcoming the inherent defects of manual monitoring, such as high labor intensity, high subjective judgment error, and easy recording errors. At the same time, the system has high-throughput detection capabilities, supports single-batch testing of large numbers of individuals, and enhances data consistency and comparability. All process data, images, and videos can be recorded, traced, and visualized in real time, and a manual review interface is provided. While improving the efficiency and accuracy of measurement, it ensures the reliability and integrity of the original data, providing an efficient and accurate phenotypic data acquisition solution for disease-resistant breeding. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1This is a block diagram of the intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis, according to an embodiment of the present invention. Figure 2 This is a flowchart of an intelligent testing method for shrimp disease resistance traits based on automatic image acquisition and analysis, according to an embodiment of the present invention. Figure 3 This invention relates to a design scheme for an automated and intelligent shrimp disease resistance trait testing system. Figure 4 This is a design drawing of the automatic inspection robot and image acquisition module of the present invention; Figure 5 The QR code and its recognition in the individual annotation and recognition module of this invention; Figure 6 This invention uses the YOLO model to determine the state of shrimp based on their body shape. Figure 7 This invention tests the logical flow of individual death determination; Figure 8 This is the interface of the intelligent determination system for shrimp resistance phenotypes of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] Throughout this specification, references to "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "an embodiment," "an example," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0041] like Figure 1 As shown, the present invention provides an intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis, including an automatic inspection function module, a data transmission function module, and an information processing function module.

[0042] The automatic inspection function module includes an inspection robot module, an image acquisition module, and an individual labeling and recognition module. The image acquisition module is installed on the inspection robot module, the individual labeling and recognition module is installed at the test container, and the data transmission function module is connected to the automatic inspection function module wirelessly and to the information processing function module via a wired connection.

[0043] The inspection robot module is used for autonomous navigation and movement within the test area.

[0044] The image acquisition module is used to collect image data of individual shrimp in the test container during the inspection process.

[0045] The individual labeling and identification module is used to uniquely identify each test individual and its location.

[0046] The data transmission module is used to transmit the image data acquired by the image acquisition module to the information processing module.

[0047] The information processing module is used to receive and process the image data, identify the information of the individual labeling and identification module to determine the identity of the test individual, and determine the survival status of the shrimp individual through a machine learning model based on the image of the shrimp individual, record the time of death and statistically analyze the test results.

[0048] Furthermore, the inspection robot module includes an autonomous mobile chassis, a navigation and mapping unit, a collision avoidance unit, and an automatic charging unit.

[0049] The navigation and mapping unit is used to construct maps and plan routes for the test area.

[0050] The anti-collision unit is used to avoid obstacles during movement.

[0051] The automatic charging unit is used to enable the robot to automatically return to the charging station for charging during inspection breaks.

[0052] Furthermore, the image acquisition module includes multiple industrial cameras, a protective housing, and an internal temperature and humidity control unit.

[0053] The multiple industrial cameras are positioned at different heights to capture images of test containers at different levels of the test rack.

[0054] The protective shell is a closed cylindrical structure used for moisture and corrosion protection.

[0055] The internal temperature and humidity control unit is used to prevent lens fogging.

[0056] Furthermore, the individual labeling and identification module is a QR code label affixed above each test container compartment.

[0057] The information encoded by the QR code includes at least the test rack number, layer number, container number, and grid number; the QR code encoding is used to achieve unique identification and location of each test individual.

[0058] Furthermore, the information processing module includes an image decoding and matching unit and a state intelligent determination unit.

[0059] The image decoding and matching unit is used to decode the received video or image data, identify the QR code information, and match the individual shrimp image with its corresponding grid position.

[0060] The intelligent state determination unit is used to run a pre-trained machine learning model to analyze the matched individual shrimp images and determine their survival or death status based on their body posture.

[0061] Furthermore, the intelligent state determination unit uses the YOLO model as its machine learning model. The YOLO model is trained using a large amount of labeled live and dead shrimp image data and can determine mortality based on the shrimp's side-lying or belly-up posture.

[0062] Furthermore, the information processing function module also includes a multi-round logic judgment unit.

[0063] The multi-round logic judgment unit is used to cross-verify and finally confirm the status judgment results of the same test individual in multiple consecutive inspections, so as to eliminate possible misjudgments of single death status and further improve the accuracy of death time.

[0064] Based on the same inventive concept, such as Figure 2 As shown, this invention also provides an intelligent testing method for shrimp disease resistance traits based on automatic image acquisition and analysis, including: Step S1: By placing the automatic inspection robot in the test area and activating its control software, the robot moves between test frames to complete the environmental modeling of the walkable area and the labeling of critical path points.

[0065] Step S2: Import the QR code ID, corresponding family number, and test start time information of all test grids through the information processing function module to complete the initialization of test data.

[0066] Step S3: In the inspection robot module, the walking route is planned based on the movement sequence of the marked points, and the parameters of the inspection strategy are configured by setting the number of walks, the patrol time interval and the judgment round.

[0067] Step S4: By activating the image acquisition function of the information processing module and running the inspection program, the robot begins to perform automatic inspection according to the set strategy, and supports manual pause and resume during the inspection process.

[0068] Step S5: After each round of inspection, the results data are automatically saved and exported, and the status and death time information of each test individual are displayed on the human-machine interface to achieve real-time feedback and recording of the monitoring results.

[0069] Step S6: Through the manual review interface, the system judgment result is reviewed and corrected. After confirmation, the "clear" operation is performed to finally confirm the status of the deceased individual and block its QR code identifier, making it invalid in subsequent detection.

[0070] Step S7: The information processing module automatically calculates and dynamically updates the overall mortality curve and family survival rate statistics charts to achieve comprehensive visualization and continuous tracking of the adversarial test results.

[0071] Furthermore, in step S3, when the number of judgment rounds is set to be greater than 1, the system will logically compare the judgment results of the same test individual in consecutive rounds. Only when the individual is confirmed to be in a dead state in multiple consecutive judgments will the individual's death time be finally recorded. The final time is the time when the individual is first judged to be in a dead state.

[0072] To address the shortcomings of current manual-based shrimp resistance testing methods, such as being time-consuming, labor-intensive, and inaccurate, this invention aims to establish an automated and intelligent testing system and method based on existing resistance testing systems. Through the development and combination of multiple functional and technical modules, including automated robot inspection, QR code labeling for individual identification, industrial camera image data acquisition, and machine learning models for determining individual survival status, the system achieves automatic and accurate determination and recording of the survival time of infected individuals. Figure 3 This system aims to address the issues of labor-intensive and inaccurate resistance testing. Simultaneously, it can increase testing speed to a high-throughput level, allowing for the testing of more individuals at once. This merges and simultaneously performs tests that previously required multiple batches, further enhancing the consistency and comparability of resistance data.

[0073] To more clearly illustrate the present invention, the following embodiment 1 further explains the solution of the present invention: The testing system's frame and basic method are based on the authorized patent "Large-scale determination method of resistance to acute hepatopancreatic necrosis disease in Litopenaeus vannamei" (ZL202310391807.1) and the utility model patent "A separation device for preventing escape in shrimp disease resistance testing" (ZL202320811114.9). Specifically, all shrimp are placed individually in transparent acrylic boxes, each box containing 10 compartments, with one shrimp per compartment. The test boxes are then placed on a three-tiered testing rack, each tier having a water-filled glass tank, with 6-8 boxes on each side.

[0074] The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis consists of three main functions: automatic inspection, data transmission, and information processing.

[0075] The automatic inspection function consists of three hardware modules: an inspection robot module, an image acquisition module, and an individual labeling and recognition module, along with corresponding software.

[0076] The inspection robot module consists of a robot that can model the walking area of ​​the test room and walk automatically according to the planned route (see Table 1 for specific parameters).

[0077]

[0078] Table 1. Main Parameters of the Robot The inspection robot module is equipped with a corresponding automatic charging station, which can charge autonomously and maintain continuous operation for more than 6 hours. After completing each inspection task, the inspection robot will automatically walk to the charging station to charge and have enough power to complete the next inspection task.

[0079] The inspection robot module is equipped with robot control software, which can control the robot to perform functions such as map modeling, key point marking, autonomous navigation, and movement according to a planned route. The maximum mapping area should be larger than the area of ​​the test laboratory, and the mapping resolution should be less than 5cm.

[0080] The walking speed of the inspection robot module is set to a range of 0.1m / s–1m / s. The low-speed movement corresponds to the movement mode when shooting test individuals, so that the images are clear during shooting; the high-speed movement corresponds to the movement mode when not shooting test individuals, so as to reduce the inspection time.

[0081] The inspection robot module has functions such as laser anti-collision and contact anti-collision, and a red stop button on the body that can stop movement in an emergency.

[0082] The image acquisition module is located above the inspection robot module and is a closed cylindrical structure to prevent excessive humidity from causing corrosion or damage to internal components such as cameras. It is equipped with three sets of industrial cameras, which are fixed at three different heights by support brackets, corresponding to the three layers of the test rack. The three sets of cameras can be flexibly adjusted up and down to achieve the most suitable shooting angle.

[0083] The image acquisition module has an internal heating function to prevent excessive temperature and humidity differences from causing the lens to fog up and become unclear. The control button is located at the rear of the camera.

[0084] The image acquisition module can continuously capture photos of shrimp postures using three video cameras and merge them into a single continuous video file for easy transmission.

[0085] The individual labeling and recognition module consists of a set of square QR codes affixed above the glass slot of each test box, with the same width as the box / grid. Figure 5 ).

[0086] The individual labeling and recognition module's QR code information consists of four numbers representing individual information (abcd), including the test rack number, layer number, box number, and grid number.

[0087] The data transmission function refers to the information exchange between the inspection robot and the data processing module achieved through a wireless router. The router is connected to the inspection robot wirelessly and to the information processing module via a wired connection, ensuring unrestricted bandwidth for receiving video files sent by the robot in real time.

[0088] The router with the data transmission function can transmit the images of all test individuals collected by the image acquisition module during the inspection process to the information processing module in real time in the form of video files.

[0089] The information processing module has an image recognition function. Based on the received test shrimp video images from each round of inspection, it can decode the QR code in each grid through model recognition and match the test individuals and QR codes one by one according to the principle of the closest vertical distance.

[0090] The information processing module processes test shrimp video images and uses a trained YOLO model to determine the shrimp's survival or death status through a decision algorithm based on the shrimp's body posture (lying on its side / belly-up, etc.). Figure 6 ).

[0091] The YOLO model in the information processing module has been trained using 16,000 images of test individuals collected in the testing system, including 11,000 images of live shrimp and 5,000 images of dead shrimp. The trained model can quickly identify the survival or death status of individuals. In a validation group of 500 samples, the individual recognition rate of the image acquisition system exceeded 99.5% (502 / 504). Compared with the results of manual judgment, the accuracy rate of single individual survival status determination exceeded 99.0% (497 / 502), and the cumulative accuracy rate was 98.5%.

[0092] The information processing module is implemented using AI machine learning algorithms and requires a regular computer with an 8-core or higher CPU, 16GB or more of RAM, 100GB or more of hard disk space, a GPU with more than 3000 CUDA cores, and hardware interfaces such as USB serial ports and RJ45 network ports.

[0093] The information processing module can further confirm or correct test individuals judged as dead in a single round based on the results of multiple rounds of patrols, and determine the final time of death. Figure 7 ).

[0094] The information processing module displays the status and death time of each test shrimp in real time in a graphical and list interface.

[0095] The information processing module can calculate the survival rate of different families at a specified time of death based on the statistical information on individual survival status and time of death, and plot the total mortality curve of the test group.

[0096] The control and parameter settings of the inspection robot module, the machine learning YOLO model of the information processing module, the code for single-round and multi-round logical judgments, and the output of results ultimately form the interface of the shrimp resistance phenotype intelligent determination system. Figure 8 ).

[0097] The method for using the intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis includes the following steps: Step 1: Place the automatic inspection robot in the experimental workshop, turn it on and open the robot control software in the inspection robot module, so that the robot can move between the test racks, model the walkable area of ​​the test laboratory, and mark the key locations.

[0098] Step 2: In the information processing module, select and import all the grid QR code IDs of the individuals placed in the test, the family lineage information of each box, and the test start time, etc.

[0099] Step 3: Configure the following parameters in the inspection robot module: Set the movement order by filling in the marked points of the model to control the robot's walking path, ensuring that the walking path of the marked points covers all test individuals; Set the number of walks, which controls the number of times the robot repeats the specified walking path within a patrol time point; Set the patrol time, which controls how often the robot patrols; Set the number of judgment rounds, which controls the number of comparisons the system uses to arrive at the final judgment result. 1 round means that the final result is given after one patrol round, 2 rounds means that the final result is obtained after two patrol rounds by combining the contents of the first two rounds. The default value is 1.

[0100] Step 4: After clicking "Camera Open" in the information processing module, click the run button; the robot will start the inspection according to the set parameters. During the inspection, it supports the function of pausing and restarting the inspection at any time.

[0101] Step 5: After each round of patrol and inspection, the results table will be automatically saved and exported; simultaneously, the status of all individuals will be displayed in the graphical and list interfaces of the information processing module, showing "Live," "Dead and Not Retrieved," and "Dead and Retrieved." Test individuals determined to be dead and not retrieved will have their time of death recorded. This will also be displayed in a dedicated list area in the lower right corner. Figure 8 (A separate reminder.)

[0102] Step Six: Staff can conduct periodic reviews as needed. Within this module's interface, the status of any missed individuals can be manually corrected to "Dead, Not Caught," and incorrectly identified individuals can be corrected back to "Live." After confirming the status, click the "Clear" button to change all shrimp in the "Dead, Not Caught" status to "Dead, Caught" and confirm the time of death. Simultaneously, the QR code tags for these individuals are disabled in this test and will no longer be recognized.

[0103] Step 7: The information processing module will automatically calculate and display the overall mortality rate line graph for all individuals, as well as the survival rate calculated by family line, and update it each time the "Clear" button is clicked.

[0104] like Figure 3-8 As shown, the present invention relates to an intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis.

[0105] To enable examiners to better understand the technical solution of the present invention, the implementation steps and effects of the present invention are further illustrated below through Example 2.

[0106] Example 2: The AHPND resistance of a whole-sib family of Litopenaeus vannamei was tested using the intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis.

[0107] Step 1: Randomly select 30 individuals from each family lineage and place them in three 10-slot test boxes. Place 18 test boxes on one side of each shelf, for a total of 36 test boxes on both sides. Fill all three test shelves, resulting in a total of 36 full-sibling families and 1080 individuals. Align the 10 slots of each test box with the 10 QR codes.

[0108] Step 2: Prepare and feed Vibrio parahaemolyticus poison bait according to the "Large-scale determination method for the resistance of Litopenaeus vannamei to acute hepatopancreatic necrosis disease" in ZL202310391807.1, and record the feeding time as the start time of the experiment.

[0109] Step 3: Open the robot control software in the inspection robot module and import the laboratory modeling file you created; open the information processing function module and import the fixed template Excel spreadsheet. The spreadsheet contains the QR code ID of all the individuals placed in the test, the pedigree information of each corresponding box, and the test start time, etc. After importing, the information processing function module automatically recognizes and displays that all shrimp are "alive".

[0110] Step 4: Open the parameter settings in the information processing function module, set the movement order to the outer and inner marker points of each shelf, so that the planned route covers 3 shelves; set the number of walks to 1, which means that the planned route will be walked once during each inspection time; set the patrol time to 1, which means that the inspection will be carried out once every 1 hour; set the judgment round to 1, which means that the result will be output directly after each round of patrol, without the need for multiple rounds of result comparison; select the default settings for the remaining parameters.

[0111] Step 5: Click "Open Camera", then click "Run". The robot will begin automatic inspection every hour.

[0112] Step 6: Every 12 hours, staff will check the individuals identified as "dead and not retrieved" in the special list area in the lower right corner of the information processing module, retrieve the dead individuals, correct the incorrectly identified individuals, and change their status to "alive"; click "clear" to change the status of all dead individuals to "dead and retrieved" in batches.

[0113] Step 7: When the cumulative mortality rate in the information processing module's display interface exceeds 60%, click "End All". The system will automatically generate the experiment end time and export the final results table. The table allows you to query information such as the start time, end status, death time, survival time, and experiment termination time for each family and each shrimp, which can be used for breeding evaluation and mating plan formulation.

[0114] During this resistance test, the robot inspected both sides of one shelf, i.e., 36 boxes and 360 shrimp, in 1.5 minutes. Based on this, the detection speed is no less than 12,000 shrimp / hour. The first 600 shrimp were selected at three random time points for manual verification of the system's judgment results. The accuracy rate for judging 14 dead shrimp was 100%, and the accuracy rate for judging 586 live shrimp was 99.5%. The overall accuracy rate of the system's judgment for all individuals was 99.5%.

[0115] Analysis of the embodiments shows that the intelligent testing system and method for shrimp disease resistance traits based on automatic image acquisition and analysis of the present invention can replace the manual inspection process, automatically and intelligently completing the determination of dead test individuals and the recording of death time in the traditional resistance testing process, greatly reducing labor costs and labor intensity; at the same time, it can increase the observation frequency to 1 hour or half an hour, more accurately determine the shrimp's death time, and increase the accuracy of the survival time phenotypic; the observation and inspection results have high accuracy, with an error rate lower than manual inspection, and can automatically save phenotypic results, reducing the introduction of recording errors; the original video data can be saved, providing direct experimental evidence; the detection speed reaches a high throughput level, and test batches can be merged and reduced, increasing the comparability of data between different batches.

[0116] It should be noted that the specific methods by which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0117] Finally, it should be noted that although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis, characterized in that, The system includes an automatic inspection module, a data transmission module, and an information processing module. The automatic inspection function module includes an inspection robot module, an image acquisition module, and an individual labeling and recognition module. The image acquisition module is installed on the inspection robot module, the individual labeling and recognition module is installed at the test container, and the data transmission function module is connected to the automatic inspection function module wirelessly and to the information processing function module via a wired connection. The inspection robot module is used for autonomous navigation and movement within the test area; The image acquisition module is used to acquire image data of individual shrimp in the test container during the inspection process; The individual labeling and identification module is used to uniquely identify each test individual and its location; The data transmission module is used to transmit the image data acquired by the image acquisition module to the information processing module. The information processing module is used to receive and process the image data, identify the information of the individual labeling and identification module to determine the identity of the test individual, and determine the survival status of the shrimp individual through a machine learning model based on the image of the shrimp individual, record the time of death and statistically analyze the test results.

2. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 1, characterized in that, The inspection robot module includes an autonomous mobile chassis, a navigation and mapping unit, a collision avoidance unit, and an automatic charging unit. The navigation and mapping unit is used to construct maps and plan routes for the test area; The anti-collision unit is used to avoid obstacles during movement; The automatic charging unit is used to enable the robot to automatically return to the charging station for charging during inspection breaks.

3. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 2, characterized in that, The image acquisition module includes multiple industrial cameras, a protective housing, and an internal temperature and humidity control unit. The multiple industrial cameras are positioned at different heights to capture images of test containers at different levels of the test rack. The protective outer shell is a closed cylindrical structure for moisture and corrosion protection; The internal temperature and humidity control unit is used to prevent lens fogging.

4. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 3, characterized in that, The individual labeling and recognition module is a QR code label attached above each test container compartment; The information on the QR code label includes at least the test rack number, layer number, container number, and grid number; it is used to achieve unique identification and location of each test individual.

5. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 4, characterized in that, The information processing module includes an image decoding and matching unit and a state intelligent determination unit; The image decoding and matching unit is used to decode the received video or image data, identify QR code information, and match individual shrimp images with their respective grid positions. The intelligent state determination unit is used to run a pre-trained machine learning model to analyze the matched individual shrimp images and determine their survival or death status based on their body posture.

6. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 5, characterized in that, The intelligent state determination unit uses the YOLO model as its machine learning model. The YOLO model is trained using a large amount of labeled live and dead shrimp image data and can determine mortality based on the shrimp's side-lying or belly-up posture.

7. The intelligent testing system for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 6, characterized in that, The information processing module also includes a multi-round logic judgment unit; The multi-round logic judgment unit is used to cross-verify and finally confirm the status judgment results of the same test individual in multiple consecutive inspections, so as to eliminate possible misjudgments of single death status and further improve the accuracy of death time.

8. An intelligent testing method for shrimp disease resistance traits based on automatic image acquisition and analysis, characterized in that, The method employs the testing system as described in any one of claims 1 to 7, and includes the following steps: Step S1: By placing the automatic inspection robot in the test area and starting its control software, the robot moves between test frames to complete the environmental modeling of the walkable area and the labeling of key path points. Step S2: Import the QR code ID, corresponding family number and test start time information of all test grids through the information processing function module to complete the initialization of test data; Step S3: In the inspection robot module, the movement sequence based on the marked points is set to plan the walking route, and the parameters of the inspection strategy are configured by setting the number of walks, the patrol time interval and the judgment round. Step S4: By activating the image acquisition function of the information processing module and running the inspection program, the robot begins to perform automatic inspection according to the set strategy, and supports manual pause and resume during the inspection process. Step S5: After each round of inspection, the results data are automatically saved and exported, and the status and death time information of each test individual are displayed on the human-machine interface to achieve real-time feedback and recording of the monitoring results. Step S6: Through the manual review interface, the system judgment result is reviewed and corrected. After confirmation, the "clear" operation is performed to finally confirm the status of the deceased individual and block its QR code identifier, making it invalid in subsequent detection. Step S7: The information processing module automatically calculates and dynamically updates the overall mortality curve and family survival rate statistics charts to achieve comprehensive visualization and continuous tracking of the adversarial test results.

9. The intelligent testing method for shrimp disease resistance traits based on automatic image acquisition and analysis according to claim 8, characterized in that, In step S3, when the number of judgment rounds is set to be greater than 1, the system will logically compare the judgment results of the same test individual in consecutive rounds. Only when the individual is confirmed to be in a dead state in multiple consecutive judgments will the death time of the individual be finally recorded. The final time is the time when the individual is first judged to be in a dead state.

Images