Unmanned ship for killing ampullaria gigas eggs and control method thereof

Abstract

The invention discloses an unmanned ship for killing ampullaria gigas eggs and a control method thereof.The unmanned ship comprises a ship body, the ship body is provided with a visual recognition module, a Beidou positioning module, a main control unit, an ultrasonic unit, a high-pressure water gun and a holder, and the main control unit collects distance information in real time through the ultrasonic unit and controls the ship body to cruise along the bank; the visual recognition module collects images around the ship body in real time, recognizes the collected images through a preset ampullaria gigas egg image recognition model, and sends target coordinates to the main control unit if ampullaria gigas eggs are recognized; after receiving the target coordinates, the main control unit controls the steering engine angle of the holder according to the coordinates, so that the high-pressure water gun aims at the target and is controlled to shoot the target; and after killing is completed, the ship body is controlled to continue cruising, and the next target is found. The pomacea canaliculata eggs are analyzed through the visual identification technology, the pomacea canaliculata eggs are beaten off through the water gun, so that the pomacea canaliculata cannot breed, and the purpose of killing the pomacea canaliculata is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of apple snail prevention and control, and in particular to an unmanned boat for disinfecting apple snail eggs and a control method thereof. Background Art

[0002] The golden apple snail, an invasive species with a strong reproductive capacity and no natural enemies in China, feeds on rice seedlings, severely harming rice production in some parts of China. Its strong adaptability and rapid reproduction allow it to spread rapidly throughout rivers, lakes, and fields, wreaking havoc on crops and the plant ecosystems of rivers and lakes.

[0003] Chemical control is the most widely used and effective method for the control of golden apple snails. However, all chemical control methods have certain environmental impacts, and the effects of the drugs can cause persistent pollution. Chemical agents can cause serious environmental pollution, affecting biodiversity and the ecological environment. They also have certain impacts on fish and shrimp aquaculture, potentially affecting production, and can further harm humans along the food chain. Furthermore, if the frequency and dosage of use are not controlled, the golden apple snails may develop drug resistance, making it more difficult to kill them.

[0004] Besides chemical control, the most widely used method is manual control. For example, barriers such as wire mesh can be placed at irrigation canals or at the inlets and outlets of rice paddies in areas with severe golden apple snail infestations to prevent them from entering the fields. Alternatively, long strips of bamboo, such as bamboo strips, can be inserted into rice paddies at an appropriate height to entice the snails to climb onto them and lay their eggs in clusters. Regular removal and disposal of the egg masses can be highly effective. However, physical control methods have the disadvantages of high labor costs and low efficiency, and the snails can easily escape and spread with water during collection.

[0005] Manual pest control is inefficient and costly, and can easily spread along waterways, making it difficult to effectively remove the snails. Traditional chemical pest control methods, regardless of their effectiveness, are environmentally friendly and have an unavoidable impact on the ecosystem. Therefore, an environmentally friendly and harmless method for disinfecting apple snail eggs is urgently needed. Summary of the Invention

[0006] The technical problem to be solved by the embodiments of the present invention is to provide an unmanned boat for disinfecting apple snail eggs and a control method thereof, so as to achieve environmentally friendly and harmless disinfecting of apple snail eggs.

[0007] In order to solve the above technical problems, the embodiment of the present invention proposes an unmanned boat for disinfecting apple snail eggs, including a hull, a visual recognition module, a Beidou positioning module, a main control unit, an ultrasonic unit, a high-pressure water gun for disinfecting apple snail eggs, and a pan-tilt system for driving the high-pressure water gun to adjust the shooting direction, wherein: The main control unit uses the Beidou positioning module to locate the ship and the ultrasonic unit to collect distance information in real time to control the ship's coastal cruising. The visual recognition module collects images around the hull in real time and uses a preset golden apple snail egg image recognition model to identify the collected images. If golden apple snail eggs are identified, the target coordinates are sent to the main control unit; After receiving the target coordinates, the main control unit controls the servo angle of the gimbal according to the coordinates, aims the high-pressure water gun at the target, and controls the high-pressure water gun to shoot the target; after completing the disinfection, it controls the hull to continue cruising and look for the next target.

[0008] Accordingly, an embodiment of the present invention further provides a method for controlling an unmanned boat for disinfecting golden apple snail eggs, comprising: Automatic cruising steps: Activate the ultrasonic unit to collect distance data, keep the hull cruising along the coast, and simultaneously obtain and record the position information in real time through the Beidou positioning module; scan the coastline through the visual recognition module, and identify the apple snail eggs in real time through the preset apple snail egg image. After identifying the apple snail eggs, control the pan-tilt system to aim the high-pressure water gun at the target, and then activate the high-pressure water gun to shoot the target; after completing the disinfection, control the hull to continue cruising and search for the next target; Manual cruising steps: Control the navigation of the hull according to the user's remote control instructions, collect images in real time, obtain and record location information, and upload the collected images at the same time; control the pan-tilt head according to the user's corresponding remote control disinfection instructions to aim the high-pressure water gun at the target, and then start the high-pressure water gun to shoot the target to complete the disinfection.

[0009] The beneficial effects of the present invention are as follows: the present invention analyzes the eggs of the golden apple snail through visual recognition technology, and destroys the eggs of the golden apple snail through the impact of a high-pressure water gun, making it impossible for the golden apple snail to reproduce, thereby achieving the purpose of eliminating the golden apple snail, and can effectively solve the environmental pollution and ecological balance problems caused by traditional chemical control. In particular, it will not affect the use scene during and after use, and will not affect production and life. The present invention can effectively simplify the processing flow. The current manual disinfection mode consumes a lot of manpower, but the use and maintenance operation of the present invention are convenient, and one person can manage multiple machines, reducing labor costs. The present invention can adjust the dynamic cleaning strategy by marking the cleaning area, and carry out targeted cleaning according to the reproduction speed and cycle for the area that needs to be cleaned. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 The present invention is a flowchart of an unmanned boat for disinfecting golden apple snail eggs according to an embodiment of the present invention. DETAILED DESCRIPTION

[0011] It should be noted that, unless there is a conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The present invention is further described in detail below with reference to the drawings and specific embodiments.

[0012] In the embodiments of the present invention, if there are directional indications (such as up, down, left, right, front, back, etc.), they are only used to explain the relative position relationship and movement status of the various components under a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0013] In addition, the terms "first," "second," and so on, used in this disclosure are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Therefore, features specified as "first" or "second" may explicitly or implicitly include at least one of these features.

[0014] The unmanned boat for disinfecting golden apple snail eggs according to the embodiment of the present invention comprises a hull equipped with a visual recognition module, a Beidou positioning module, a main control unit, an ultrasonic unit, a high-pressure water gun, and a pan / tilt platform.

[0015] The present invention can utilize existing unmanned surface vehicles (USVs), fully autonomous surface robots that utilize precise satellite positioning and self-sensing capabilities to navigate the water according to pre-defined missions. For example, the most common monohulls with electric propulsion (lithium batteries, solar power) can be used. These are simple, suitable for use in calm waters (such as lakes), and environmentally friendly. Propulsion types include propellers, water jets, and sail power (such as wind-powered saildrones).

[0016] As an implementation method, the hull is provided with a peripheral expansion module, a communication module for communicating with the host computer, and a relay for switching between automatic cruise mode and manual remote control mode. In the manual remote control mode, the main control unit uploads the image collected in real time by the visual recognition module, and controls the navigation of the hull, controls the pan-tilt head to adjust the shooting direction of the high-pressure water gun, and whether to start the high-pressure water gun shooting according to the user's remote control instructions.

[0017] In specific implementation, the hull can be based on a fully 3D printed catamaran wind-powered boat, with a 2212 1400kv motor and a 30A electronic regulator as the power source, steering controlled by a wind rudder, and dual PID manual-automatic control achieved through relays, ESP32, and receivers.

[0018] A high-pressure water gun is used to disinfect golden apple snail eggs. It consists of a high-pressure pump, a water tank, a water gun, a solenoid valve, and a water pipe. The high-pressure pump draws water from the tank through the pipe and ejects it through the gun. The main control unit controls the on / off function of the high-pressure water gun via the solenoid valve. During implementation, the water in the tank can also be treated with chemicals.

[0019] The gimbal is used to drive the high-pressure water gun to adjust the shooting direction. The gimbal is built using two 360° servos.

[0020] The visual recognition module captures images of the ship's surroundings in real time and uses a pre-set golden apple snail egg image recognition model to identify the captured images. If a golden apple snail egg is identified, the target coordinates (the model outputs an identification box containing four coordinates: x, y, w, h) are sent to the main control unit. The visual recognition module can be a K210 visual recognition module. In specific implementations, the visual recognition module and high-pressure water gun are mounted on the gimbal.

[0021] Please refer to Figure 1 The main control unit uses the Beidou positioning module to locate the target and the ultrasonic unit to collect real-time distance information, controlling the vessel's coastal cruising. After receiving the target coordinates, the main control unit controls the PTZ's steering gear angles based on the coordinates, aiming the high-pressure water cannon at the target and firing the water cannon at the target (preferably, the water pressure and firing time are adjusted according to the target distance. If there are multiple targets, they are fired sequentially after sorting by coordinates). After completing the disinfection, the vessel is controlled to continue cruising and search for the next target. In specific implementation, Zhang Zhengyou's calibration method is first used to obtain the internal parameters of the visual recognition module, and a mapping relationship is established between the pixel coordinate system and the PTZ's steering gear angles. This mapping relationship is then used to control the PTZ according to the coordinates.

[0022] During implementation, the ultrasonic unit is mounted on the left side of the ship, receiving real-time information about the ship's position relative to the shore, enabling coastal disinfection. Finally, the Beidou positioning module records the disinfection location and patrol trajectory and sends the data to a host PC.

[0023] The Beidou positioning module (ATGM336H is available) receives Beidou satellite signals through an antenna and outputs real-time NMEA-0183 data such as latitude, longitude, altitude, and time. This data is then transmitted via a serial communication port (UART). The Beidou positioning module has an accuracy of less than 2.5 meters and supports the NMEA-0183 standard protocol. It is compatible with GPS, Beidou, QZSS, and GLONASS positioning, ensuring stability in complex environments.

[0024] The main control unit, an ESP32, serves as the "brain" of the unmanned vessel. It receives BeiDou data from the BeiDou NMEA-0183 positioning module (BeiDou NMEA-0183), runs filtering algorithms, and transmits it remotely via Wi-Fi. The receiving unit uses serial communication (UART) to support seamless communication with the BeiDou positioning module, and its dual-core processor ensures real-time data processing. The processing unit uses the TinyGPSPlus library to parse NMEA-0183 data, including the current latitude, longitude, time, altitude, and number of satellites, and transmits this information to a host PC via Wi-Fi. The storage unit stores the NMEA-0183 data parsed by the ESP32 every 60 seconds, including location information (including latitude, longitude, time, altitude, and number of satellites). The reading unit connects to the ESP32 and uses the Arduino IDE to print the stored data via serial communication (UART) to the serial monitor in the Arduino IDE if network communication is disrupted. Sending part: ESP32 connects to the LAN created by the PC. After the connection is successful, it creates a TCP server and prints the IP address of ESP32. After that, it checks every 500ms whether there is a Wi-Fi connection and continuously prints the connection status. If the connection is successful, it continuously sends the parsed file and the original file to the TCP server.

[0025] The ESP32 main control unit has a built-in Wi-Fi module and a local area network created by the PC.

[0026] Peripheral expansion module: A Python-based PC host computer records information by accepting the TCP server and port created by the ESP32. The Python program is also connected to the ATK-GPS program via a simulated serial port (UART). This allows the current location to be printed on the AutoNavi map, along with the latitude, longitude, time, and altitude. GPS logs can also be saved to record the path taken.

[0027] The development software for the main control unit can be: Python·Ardunio IDE·Launch Virtual SerialPort Driver (VSPD).

[0028] As an embodiment, the golden apple snail egg image recognition model is constructed and trained according to the following steps: Construct an image recognition model for golden apple snail eggs; Collect images of apple snail eggs and apple snail egg-like objects under different lighting, angles, and backgrounds to construct a dataset. The apple snail egg-like objects include granular plastic and red flowers. The images in the dataset were annotated with the bounding boxes of the egg mass area and individual eggs. The dataset was divided into training, validation, and test sets. The images of the golden apple snail egg mass, golden apple snail eggs, plastic pellets, and red flowers were annotated with different recognition boxes. The golden apple snail egg image recognition model is trained based on the data in the data set, and the model parameters are iteratively optimized based on the false detection and missed detection situations; After the training is completed, a preset golden apple snail egg image recognition model is obtained, and the model is deployed in the visual recognition module.

[0029] As an embodiment, after identifying the golden apple snail eggs, the visual recognition module determines whether the size of the identified unit frame is within a preset range. If so, the target coordinates are sent to the main control unit. If not, the main control unit controls the hull to continue sailing.

[0030] The K210 visual recognition module accurately identifies golden apple snail eggs. This is achieved through four steps: feature analysis, image acquisition, model training, and software implementation. Software development can be done using the Canaan AI training platform / CanMV IDE. The following are the specific implementation methods and key technical points: 1. Feature Analysis (1) Identify the typical characteristics of golden apple snail eggs: Golden apple snail eggs are round, about 2 mm in diameter, pink or bright red when first laid, and covered with a layer of white powder. The egg masses are oval and arranged neatly, often attached to plant stems, ditch walls, or ridges above the water surface. These color and morphological characteristics are the core basis for identification.

[0031] (2) Identify similar features to differentiate. Red flowers, granular plastics, etc. may be identified as similar objects, and model training and differentiation are required.

[0032] 2. Image Acquisition (1) Considering environmental adaptability, high-definition images of egg masses were taken under different lighting and angles to ensure that the dataset covers actual application scenarios. (2) Consider similar characteristics: Take photos of objects that easily appear together with apple snail eggs in water, which may affect the recognition efficiency.

[0033] (3) Image screening Deduplication: Use clustering algorithms to remove duplicate images.

[0034] Clarity: Use image gradients to detect blurry images and remove them.

[0035] Eliminate irrelevant content: Manually filter out non-target objects.

[0036] 3. Model Training (1) Dataset annotation and training platform Training platform: Use the Canaan AI training platform to annotate the collected egg mass images, with the annotated targets being the egg mass area and the bounding boxes of individual eggs.

[0037] Model selection and parameter adjustment: When labeling, the egg mass, granular plastic, and red flowers need to be marked with different identification frames to enhance model specificity.

[0038] Training parameters: training times 240; Batch size 8; Learning rate 0.001; Callout box limit 5.

[0039] Object detection algorithm: The YOLO lightweight model is used to adapt to the computing power limitations of the K210.

[0040] (2) Shape matching: Using the circular features of the egg, the target is further screened through edge detection (Canny algorithm).

[0041] (3) Model deployment and verification: Deploy the trained model file (kmodel) to the K210 module and load it through the SD card.

[0042] Test the recognition accuracy in a real environment and iteratively optimize the model parameters based on false detections and missed detections.

[0043] 4. Software Implementation Load the YOLO anchor box parameters in the CanMV IDE software, run YOLO inference (input image to obtain detection results), obtain the detection result list (feature map height, feature map width, feature map color feature kpu.init_yolo2(anchor, anchor_num=(int)(len(anchor) / 2), img_w=320, img_h=240, net_w=320 ,net_h=240 ,layer_w=10 ,layer_h=8classes=len(labels))), if there is a detection target (img = sensor.snapshot), perform similarity screening (dect = kpu.regionlayer_yolo2), select the target with a similarity greater than 30% (threshold=0.6, nms_value=0.3,) draw the bounding box on the image (the (x, y, z) coordinates of the feature box will also be output at the same time a = img.draw_rectangle(l[0],l[1],l[2],l[3],color=(0,255,0))) and create label information (info = "%s %.3f" % (labels[l[4]], l[5]))). The label information is the name "Apple Snail Egg" defined in the model. Finally, it is transmitted through the serial port (UART) (adjust the baud rate to 9600: Serial.begin(9600).

[0044] As an implementation method, the main control unit records and uploads the ship's navigation track, disinfection location, and date information.

[0045] The main control unit consists of three parts: coastal cruising, automatic aiming, automatic marking, and manual and automatic integration. It controls the hull cruising and automatic aiming and carries out the disinfection of egg masses. The main control unit has encapsulated functions for direction control, start and stop of the pneumatic motor, water gun control, ultrasonic ranging, and K210 gimbal two-way calibration. Among them, the three encapsulated functions of direction control, start and stop of the pneumatic motor, and ultrasonic ranging are called and then encapsulated into the main movement function. At the beginning, the address is sent, and then the status is determined to see if it is machine control (i.e. automatic cruise mode). If it is automatic movement, the golden apple snail egg mass is detected (received the signal sent back by K210). After detection, automatic aiming is performed and the water gun is turned on for disinfection. The following are the specific details: 1. Coastal Cruise (1) By using the HC-SR04 ultrasonic unit to collect distance data, and through serial communication (UART), after testing, the wind rudder servo and dual generator output PWM signals are controlled at different return thresholds to keep the hull cruising about 50 cm away from the shore.

[0046] (2) Receive the recognition signal of K210 visual recognition through serial communication (UART). When the recognition signal appears, ESP32 sends a PWM signal to control the motor to stop and enter the next step of automatic aiming.

[0047] 2. Automatic aiming (1) After K210 recognizes the target, it sends the (x, y, z) coordinates of the positioning identification frame on the K210 display screen through serial communication (UART), and outputs the corresponding PWM signals to the lateral and longitudinal servos to achieve automatic aiming of the pan-tilt control water gun.

[0048] (2) ESP32 outputs a PWM signal to the high-pressure water gun 2 seconds after identifying the target, controlling the high-pressure water gun to shoot at the golden apple snail eggs. After 10 seconds of disinfection, K210 visual recognition again determines whether the disinfection is complete.

[0049] 3. Automatic Tagging (1) The Beidou positioning module receives the current location information in real time. ESP32 stores the current data and sends it to the PC host computer through the wife module to mark the date, time, coordinates, and date.

[0050] 4. Integration of hand and self (1) Dual PID switching is achieved through relays to manually implement the above operations, and at the same time, analog image transmission is connected to view the actual situation in real time.

[0051] ESP32 program development: The main program is developed in C / C++ based on the Arduino IDE 2.3.4 platform, and is implemented by calling the K210 and TinyGPSPlus function libraries and communicating through the serial port (UART).

[0052] The control method of the unmanned boat for disinfecting golden apple snail eggs of the present invention includes: Automatic cruising steps: Activate the ultrasonic unit to collect distance data, maintain the hull cruising along the coast, and simultaneously obtain and record the position information in real time through the Beidou positioning module; scan the coastline through the visual recognition module, and identify the apple snail eggs in real time through the preset apple snail egg image. After identifying the apple snail eggs, control the pan-tilt to aim the high-pressure water gun at the target, and then activate the high-pressure water gun to shoot the target (preferably, adjust the water pressure and shooting time according to the target distance. If there are multiple targets, sort them by coordinates and then hit them in sequence); after completing the disinfection, control the hull to continue cruising and search for the next target; Manual cruising steps: Control the navigation of the hull according to the user's remote control instructions, collect images in real time, obtain and record location information, and upload the collected images at the same time; control the pan-tilt head according to the user's corresponding remote control disinfection instructions to aim the high-pressure water gun at the target, and then start the high-pressure water gun to shoot the target to complete the disinfection.

[0053] As an embodiment, the golden apple snail egg image recognition model is obtained according to the following steps: Construct an image recognition model for golden apple snail eggs; Collect images of apple snail eggs and apple snail egg-like objects under different lighting, angles, and backgrounds to construct a dataset. The apple snail egg-like objects include granular plastic and red flowers. The images in the dataset were annotated with the bounding boxes of the egg mass area and individual eggs. The dataset was divided into training, validation, and test sets. The images of the golden apple snail egg mass, golden apple snail eggs, plastic pellets, and red flowers were annotated with different recognition boxes. The golden apple snail egg image recognition model is trained based on the data in the data set, and the model parameters are iteratively optimized based on the false detection and missed detection situations; After the training is completed, the preset golden apple snail egg image recognition model is obtained.

[0054] The specific implementation process is as follows: load the YOLO anchor box parameters in the CanMV IDE software, run YOLO inference (input image to obtain detection results), obtain the detection result list (feature map height, feature map width, feature map color feature kpu.init_yolo2(anchor, anchor_num=(int)(len(anchor) / 2), img_w=320,img_h=240,net_w=320,net_h=240,layer_w=10,layer_h=8classes=len(labels))), if there is a detection target (img = sensor.snapshot), perform similarity screening (dect = kpu.regionlayer_yolo2), select the target with a similarity greater than 30% (threshold=0.6, nms_value=0.3,) draw the bounding box on the image (the (x, y, W, h) coordinates of the feature box will also be output at the same time a = img.draw_rectangle(l[0],l[1],l[2],l[3],color=(0,255,0))) and create label information (info = "%s %.3f" % (labels[l[4]], l[5]))). The label information is the name "Apple Snail Egg" defined in the model. Finally, it is transmitted through the serial port (UART) (adjust the baud rate to 9600: Serial.begin(9600).

[0055] The (x, y, W, h) coordinates of the feature box are also output at the same time a = img.draw_rectangle(l[0],l[1],l[2],l[3],color=(0,255,0))) and label information is created (info = "%s %.3f" % (labels[l[4]],l[5]))).

[0056] As an embodiment, in the automatic cruising step, after the visual recognition module identifies the golden apple snail eggs, it determines whether the size of the identified unit frame is within a preset range. If so, the pan-tilt platform is controlled to aim the high-pressure water gun at the target; if not, the hull is controlled to continue sailing.

[0057] As an implementation method, in the automatic cruising step, after the disinfection is completed, the ship's navigation track, disinfection location (including GPS location + image), and date information are recorded and uploaded.

[0058] As an implementation method, in the automatic cruising step, after the high-pressure water gun is started for a preset time, it is recognized again through the visual recognition module. If the golden apple snail eggs are not recognized, it is determined that the disinfection is completed; if the golden apple snail eggs are recognized, the pan-tilt control is used to make the high-pressure water gun re-aim at the target, and the high-pressure water gun is started again to shoot the target.

[0059] The specific implementation of the automatic cruise step is as follows: first send the address, then judge the status (digitalRead(36)==1), it is in automatic cruise mode, automatically move (call move() function), detect the golden apple snail egg mass (receive the signal Serial.available() returned by K210 to be true), and after detection, automatically aim (call w_jz(k210_msg.x,k210_msg.w); and h_jz(k210_msg.y,k210_msg.h);), turn on the water gun for disinfection (call gun() function).

[0060] void gun() function: Use the global variable gunflag(bool) to determine the current state of the water gun, and use the for loop to control the servogun.write() function to control the water gun to enter another state.

[0061] float dis() function: Sends a sound wave signal through digitalWrite(26, HIGH), receives the signal through pulseIn(27, HIGH), and returns the detected distance after processing.

[0062] void dir_change(): changes the rudder direction via the servodir.write() function.

[0063] void zj(): Checks the status of the global variable servoflag(bool). If it is false, initializes and calibrates the gimbal servo via servoh.write() and servow.write(), and marks the variable record as true.

[0064] Void w_jz(): Maintains now_w through the two parameters passed in, and changes the angle through servow.write(now_w) to achieve automatic aiming.

[0065] Void h_zj(): Refer to the definition of w_jz.

[0066] Void qianjing(): Operate the motor (motor ESC) through servomotor.write() and maintain the status of the global variable motorflag(bool).

[0067] / / Before calling this function, you need to determine the status of the global variable motorflag.

[0068] Void stopping(): For its definition, please refer to the qinajng() function.

[0069] Viod move(): Keep the motors on through the qingjian() function, call the dis() function to detect the distance, and based on the detected distance, if it deviates from the preset track, correct the direction through the dir_change() function to keep it on the correct track.

[0070] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An unmanned boat for disinfecting golden apple snail eggs, comprising a hull equipped with a visual recognition module and a Beidou positioning module, characterized in that: The hull is also equipped with a main control unit, an ultrasonic unit, a high-pressure water gun for disinfecting apple snail eggs, and a pan-tilt system for driving the high-pressure water gun to adjust the shooting direction. The main control unit uses the Beidou positioning module to locate the ship and the ultrasonic unit to collect distance information in real time to control the ship's coastal cruising. The visual recognition module collects images around the hull in real time and uses a preset golden apple snail egg image recognition model to identify the collected images. If golden apple snail eggs are identified, the target coordinates are sent to the main control unit; After receiving the target coordinates, the main control unit controls the servo angle of the gimbal according to the coordinates, aims the high-pressure water gun at the target, and controls the high-pressure water gun to shoot the target; after completing the disinfection, it controls the hull to continue cruising and look for the next target.

2. The unmanned boat for disinfecting golden apple snail eggs according to claim 1, characterized in that: The golden apple snail egg image recognition model is constructed and trained according to the following steps: Construct an image recognition model for golden apple snail eggs; Collect images of apple snail eggs and apple snail egg-like objects under different lighting, angles, and backgrounds to construct a dataset. The apple snail egg-like objects include granular plastic and red flowers. The images in the dataset were annotated with the bounding boxes of the egg mass area and individual eggs. The dataset was divided into training, validation, and test sets. The images of the golden apple snail egg mass, golden apple snail eggs, plastic pellets, and red flowers were annotated with different recognition boxes. The golden apple snail egg image recognition model is trained based on the data in the data set, and the model parameters are iteratively optimized based on the false detection and missed detection situations; After the training is completed, a preset golden apple snail egg image recognition model is obtained, and the model is deployed in the visual recognition module.

3. The unmanned boat for disinfecting golden apple snail eggs according to claim 2, characterized in that: After identifying the golden apple snail egg, the visual recognition module determines whether the size of the identified unit frame is within a preset range. If so, the target coordinates are sent to the main control unit. If not, the main control unit controls the hull to continue sailing.

4. The unmanned boat for disinfecting golden apple snail eggs according to claim 1, characterized in that: The hull is provided with a communication module for communicating with the host computer and a relay for switching between automatic cruise mode and manual remote control mode. In the manual remote control mode, the main control unit uploads the images collected in real time by the visual recognition module, and controls the navigation of the hull, controls the pan-tilt head to adjust the shooting direction of the high-pressure water gun, and determines whether to start the high-pressure water gun shooting according to the user's remote control instructions.

5. The unmanned boat for disinfecting golden apple snail eggs according to claim 1, characterized in that: The main control unit records and uploads the ship's navigation track, disinfection location, and date information.

6. A control method for an unmanned boat for disinfecting apple snail eggs, characterized in that: include: Automatic cruising steps: Start the ultrasonic unit to collect distance data, keep the hull cruising along the coast, and at the same time use the Beidou positioning module to obtain and record the position information in real time; The visual recognition module scans the coastline and identifies apple snail eggs in real time using preset apple snail egg images. After identifying the apple snail eggs, the gimbal controls the high-pressure water gun to aim at the target and then activates the high-pressure water gun to shoot at the target. After completing the disinfection, the ship controls the hull to continue cruising and search for the next target. Manual cruising steps: Control the navigation of the hull according to the user's remote control instructions, collect images in real time, obtain and record location information, and upload the collected images at the same time; control the pan-tilt head according to the user's corresponding remote control disinfection instructions to aim the high-pressure water gun at the target, and then start the high-pressure water gun to shoot the target to complete the disinfection.

7. The control method for an unmanned boat for disinfecting golden apple snail eggs according to claim 6, characterized in that: The golden apple snail egg image recognition model is obtained according to the following steps: Construct an image recognition model for golden apple snail eggs; Collect images of apple snail eggs and apple snail egg-like objects under different lighting, angles, and backgrounds to construct a dataset. The apple snail egg-like objects include granular plastic and red flowers. The images in the dataset were annotated with the bounding boxes of the egg mass area and individual eggs. The dataset was divided into training, validation, and test sets. The images of the golden apple snail egg mass, golden apple snail eggs, plastic pellets, and red flowers were annotated with different recognition boxes. The golden apple snail egg image recognition model is trained based on the data in the data set, and the model parameters are iteratively optimized based on the false detection and missed detection situations; After the training is completed, the preset golden apple snail egg image recognition model is obtained.

8. The control method for an unmanned boat for disinfecting golden apple snail eggs according to claim 7, characterized in that: In the automatic cruising step, after identifying the golden apple snail eggs, the visual recognition module determines whether the size of the identified unit frame is within a preset range. If so, the pan-tilt platform is controlled to aim the high-pressure water gun at the target. If not, the hull is controlled to continue sailing.

9. The control method for an unmanned boat for disinfecting golden apple snail eggs according to claim 6, characterized in that: During the automatic cruise step, after the disinfection is completed, the ship's navigation track, disinfection location, and date information are recorded and uploaded.

10. The control method for an unmanned boat for disinfecting golden apple snail eggs according to claim 6, characterized in that: In the automatic cruise step, after the high-pressure water gun is started for a preset time, it is recognized again through the visual recognition module. If no apple snail eggs are recognized, it is judged that the disinfection is completed; if apple snail eggs are recognized, the pan-tilt control is used to make the high-pressure water gun re-aim at the target, and the high-pressure water gun is started again to shoot the target.

Images