Breeding house decontamination control device and decontamination strategy generation control method

Abstract

The invention provides a breeding house decontamination control device and a decontamination strategy generation method, and the method comprises the steps: constructing initial walking path information according to environment perception data; acquiring multispectral data of each to-be-cleaned area; inputting the multispectral data into a stain prediction model to obtain a stain identification result; according to the stain degree corresponding to the stain identification result, cleaning priority information of each to-be-cleaned area is obtained, and according to the cleaning priority information, the initial walking path information is adjusted to obtain target walking path information; based on the target walking path information, determining a current target to-be-cleaned area, and obtaining target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target to-be-cleaned area; and spraying cleaning water to the target to-be-cleaned area according to the target walking path information and the target cleaning area information. The cleaning efficiency of the breeding house is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent breeding technology, and in particular to a breeding house cleaning and disinfection control device and a cleaning and disinfection strategy generation method. Background Art

[0002] In modern livestock and poultry farming environments, disinfection of breeding facilities such as pig houses is extremely important and requires a high frequency of disinfection. Routine disinfection is usually required every two or three days, and a comprehensive deep disinfection operation is also required every week.

[0003] However, in current farming practices, the cleaning of farm houses is mostly done manually, which not only means a huge workload and a high-intensity work pace, but also requires cleaning staff to work for a long time in a harsh environment filled with high humidity and harmful gases such as ammonia, which in turn affects the quality and efficiency of cleaning and disinfection of farm houses.

[0004] Therefore, there is an urgent need for a breeding house cleaning and disinfection control device and a cleaning and disinfection strategy generation method to solve the above problems. Summary of the invention

[0005] In view of the problems existing in the prior art, the present invention provides a breeding house cleaning and disinfection control device and a cleaning and disinfection strategy generation method.

[0006] The present invention provides a cleaning and disinfection control device for a breeding house, comprising a sensor module, a multi-spectral acquisition module, a data processing module, a walking mechanism, a nozzle and a lifting module, wherein the nozzle is arranged on the cross arm side of the lifting module, wherein: The sensor module is used to sense the scene environment of the target breeding house and obtain the environmental perception data of the target breeding house; The multispectral acquisition module is used to collect multispectral data of each area to be cleaned in the target breeding house; The data processing module is used to generate corresponding target walking path information and target cleaning area information according to the environmental perception data and the multi-spectral data, wherein the target cleaning area information includes nozzle control data and lifting height data; The walking mechanism is used to drive to the target area to be cleaned in the target breeding house according to the target walking path information; The nozzle is used to spray the target area to be cleaned with cleaning water according to the cleaning water spraying mode corresponding to the nozzle control data; The lifting module is used to adjust the cleaning water spraying height according to the lifting height data.

[0007] According to a breeding house cleaning and disinfection control device provided by the present invention, the lifting module includes a lifting rod and a cross arm, and the top end of the lifting rod is connected to the middle end of the cross arm by a first rotating mechanism, and the first rotating mechanism is used to adjust the height of the lifting module according to the lifting height data; the nozzle is provided with a second rotating mechanism, and the second rotating mechanism is used to adjust the spraying angle and spraying direction of the nozzle according to the cleaning water spraying mode.

[0008] According to a farm house cleaning and disinfection control device provided by the present invention, the farm house cleaning and disinfection control device further comprises an image acquisition module, and the image acquisition module is used to acquire images of the target cleaning area to obtain an image of the area before cleaning and an image of the area after cleaning; The data processing module is also used to compare the area image before cleaning and the area image after cleaning, and if the comparison result does not meet the preset cleaning conditions, generate a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray cleaning water on the target area to be cleaned again.

[0009] The present invention also provides a method for generating a disinfection strategy based on the above-mentioned disinfection control device for aquaculture houses, comprising: Obtain environmental perception data of the target breeding house; Constructing initial walking path information according to the environmental perception data; Acquire multispectral data of each area to be cleaned in the target breeding house; Inputting the multispectral data into a stain prediction model to obtain a stain recognition result of the area to be cleaned output by the stain prediction model, wherein the stain prediction model is trained by sample spectral data marked with stain category labels; According to the degree of stain corresponding to the stain identification result, the cleaning priority information of each of the to-be-cleaned areas is obtained, and according to the cleaning priority information, the initial walking path information is adjusted to obtain the target walking path information; Based on the target walking path information, determine the current target area to be cleaned, and obtain the target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned; According to the target walking path information and the target cleaning area information, cleaning water is sprayed on the target area to be cleaned.

[0010] The present invention also provides a method for generating a disinfection strategy, wherein the stain prediction model is trained by the following steps: Obtain sample multispectral data from each area of ​​the breeding house; According to the stain category, stain degree and stain distribution in the sample multispectral data, marking the sample multispectral data with the corresponding stain category label to obtain a multispectral data sample set; The multi-task learning model is trained by using the multi-spectral data sample set, and after determining that the training results meet the preset training conditions, the stain prediction model is obtained.

[0011] The present invention also provides a method for generating a cleaning strategy, wherein the method obtains target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned, including: According to the mapping relationship between the degree of stain and the water jet parameters, the target water jet parameters are obtained, wherein the target water jet parameters at least include nozzle water pressure parameters and cleaning water flow parameters; Determine nozzle configuration information according to the stain distribution state data, wherein the nozzle configuration information at least includes nozzle type information, nozzle angle data and nozzle position data; Obtaining nozzle control data according to the target water jet parameters and the nozzle configuration information; Determining lifting height data according to the positional relationship between the target area to be cleaned and the nozzle; The target cleaning area information is obtained according to the nozzle control data and the lifting height data.

[0012] The present invention also provides a method for generating a disinfection strategy. After spraying the target area to be cleaned with cleaning water according to the target walking path information and the target cleaning area information, the method further includes: Respectively acquiring a region image before cleaning and a region image after cleaning of the target region to be cleaned; Comparing the image of the area before cleaning and the image of the area after cleaning, if the comparison result does not meet the preset cleaning condition, generating a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray the cleaning water on the target area to be cleaned again; If the comparison result meets the preset cleaning condition, the walking mechanism is driven to the next area to be cleaned according to the target walking path information.

[0013] The present invention also provides a method for generating a decontamination strategy, wherein the comparison of the area image before the decontamination and the area image after the decontamination comprises: Acquire a residual stain region image according to a cleaning region image corresponding to the target cleaning region, wherein the cleaning region image includes a region image before cleaning and a region image after cleaning; Obtaining the number of stain pixels corresponding to each residual stain area in the residual stain area image; In the case where it is determined that the number of stain pixels is greater than a preset stain pixel threshold, the corresponding residual stain area is marked as a residual stain area to be cleaned, and if the number of the residual stain areas to be cleaned is greater than the preset residual stain number, the secondary cleaning instruction corresponding to the target area to be cleaned is generated; When it is determined that the number of the stain pixels is less than or equal to the preset stain pixel threshold, the corresponding residual stain area is marked as a non-cleaned area; The method further comprises: Based on piecewise linear grayscale transformation, the cleaning area image is subjected to image enhancement processing according to the grayscale information and grayscale interval information of the cleaning area image to obtain the cleaning area image after image enhancement processing.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for generating a disinfection strategy as described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for generating a disinfection strategy as described in any one of the above is implemented.

[0016] The cleaning and disinfection control device and cleaning and disinfection strategy generation method for the breeding house provided by the present invention perform environmental perception and stain detection through hybrid sensing technology, design and integrate lidar, infrared sensor and multispectral camera, generate a detailed three-dimensional map of the breeding house, and combine historical cleaning data, use stain prediction algorithm to predict potential high-stain areas, thereby comprehensively analyzing the surface characteristics and stain distribution of the area to be cleaned in the breeding house, so as to ensure accurate detection and identification of stains and improve the cleaning efficiency of the breeding house. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of the structure of the breeding house control device provided by the present invention; Figure 2 A schematic diagram of the structure of the cleaning robot provided by the present invention; Figure 3A schematic diagram of the structure of the nozzle provided by the present invention; Figure 4 A schematic diagram of the structure of the lifting module provided by the present invention; Figure 5 This is a schematic diagram of the overall structure of the breeding house cleaning and disinfection control device provided by the present invention; Figure 6 A schematic diagram of the control strategy of the cleaning and disinfection control device for a breeding house provided by the present invention; Figure 7 A schematic diagram of a flow chart of a method for generating a decontamination strategy provided by the present invention; Figure 8 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0020] With the rapid growth of the scale and output value of the breeding industry, consumers have been stimulated to have high demands for the quantity and quality of livestock and poultry products. In this context, the livestock and poultry breeding industry is accelerating towards large-scale and highly intensive production. The fully intelligent and mechanized production method has immeasurable value in promoting the further development of the livestock and poultry breeding industry.

[0021] In the modern breeding model, the epidemic prevention work of breeding houses, especially its mechanization and intelligence level, has become a core element to promote the high-quality development of livestock and poultry breeding. Taking pig houses as an example, disinfection is a key link in their daily management. Routine disinfection is usually required every two or three days, and a comprehensive deep disinfection is required once a week. If there is a potential risk of epidemic diseases in the pig house, the frequency of disinfection needs to be further increased to ensure the effectiveness of disease prevention and control.

[0022] In the process of livestock and poultry farming, the emergence of zoonotic diseases is not uncommon, and such diseases pose a serious threat to the health of both humans and livestock. Therefore, daily cleaning and disinfection of breeding houses is particularly important. However, most breeding houses are still cleaned manually, which is not only labor-intensive and intensive, but also requires cleaning staff to work for a long time in a harsh environment filled with high humidity and harmful gases such as ammonia, which can easily lead to a decrease in work enthusiasm. In addition, cleaning staff also face certain personal safety risks when directly operating high-pressure cleaning equipment.

[0023] In view of the problems existing in the existing cleaning technology of breeding houses, the present invention adopts robot technology to replace manual disinfection and cleaning tasks, so as to achieve a safer, more efficient and easy-to-operate cleaning strategy in the livestock and poultry breeding environment. In view of the infrastructure characteristics of the breeding house and the factors such as the flatness of the ground, the present invention provides a cleaning robot chassis travel device based on geomagnetic navigation technology (i.e., a breeding house cleaning control device), uses relevant simulation software to perform a fine finite element analysis on the high-pressure water flow, and deeply studies the key parameters such as the morphology of the water flow after being sprayed from the high-pressure nozzle and the effective coverage area, thereby constructing a cleaning control device with multiple degrees of freedom and integration of multiple motion pairs. At the control level, the present invention adopts the STM32 single-chip microcomputer as the core controller of the cleaning robot, and designs the control logic of the robot.

[0024] At the same time, a control program was written for the robot in combination with the coordinated control requirements of the robot's walking mechanism and the cleaning device. Compared with the traditional manual cleaning method, the cleaning robot based on the breeding house cleaning control device provided by the present invention has shown significant advantages in stain removal. Taking the pig fattening house as an example, pigs are representatives of livestock and poultry. The pig house cleaning robot can intelligently perform cleaning tasks and only needs to be started regularly without frequent manual intervention, thereby greatly reducing labor costs and improving cleaning efficiency.

[0025] In addition, the breeding house cleaning and disinfection control device provided by the present invention can flexibly adapt to the environment and layout of the pig house, perform autonomous cleaning according to a preset route, and also support manual operation, showing extremely high applicability. It should be noted that breeding houses can currently be roughly divided into two categories: those with sewage channels and those without sewage channels. For example, for pig houses with sewage channels, the cleaning robot can push the stains directly into the sewage channel during the cleaning process without transporting them out. The path control is simple and only requires reciprocating motion. Therefore, the present invention is mainly optimized for pig houses without sewage channels.

[0026] Figure 1 The structural diagram of the breeding house control device provided by the present invention is as follows: Figure 1 As shown, the present invention provides a cleaning and disinfection control device for a breeding house, comprising a sensor module 101, a multi-spectral acquisition module 102, a data processing module 103, a walking mechanism 104, a nozzle 105 and a lifting module 106, wherein the nozzle 105 is arranged on the cross arm side of the lifting module 106, wherein: The sensor module 101 is used to sense the scene environment of the target breeding house and obtain the environmental perception data of the target breeding house; The multispectral acquisition module 102 is used to collect multispectral data of each area to be cleaned in the target breeding house; The data processing module 103 is used to generate corresponding target walking path information and target cleaning area information according to the environmental perception data and the multi-spectral data, wherein the target cleaning area information includes nozzle control data and lifting height data; The walking mechanism 104 is used to drive to the target area to be cleaned in the target breeding house according to the target walking path information; The nozzle 105 is used to spray the target area to be cleaned with cleaning water according to the cleaning water spraying mode corresponding to the nozzle control data; The lifting module 106 is used to adjust the height of the cleaning water spray according to the lifting height data.

[0027] In the present invention, a robot for cleaning a breeding house (hereinafter referred to as a cleaning robot or robot) constructed based on a breeding house cleaning and disinfection control device is described. Figure 2 The schematic diagram of the cleaning robot provided by the present invention can be referred to Figure 2 As shown, the present invention provides a cleaning robot based on a cleaning control device for a breeding house, the cleaning robot comprises a hose reel 1, wheels 2, a front camera 3, a screen 4, a nozzle 5, a multi-spectral camera 6, a lifting rod 7, a microcontroller 8, a lifting motor 9, a laser radar integrated infrared device 10 and a counterweight layer plate 11, wherein a sensor module 101 is constructed based on the laser radar integrated infrared device 10; a multi-spectral acquisition module 102 is composed of a multi-spectral camera 6; a main component in a data processing module 103 is a microprocessor 8; a walking mechanism 104 is mainly composed of wheels 2, and it should be noted that the wheels 2 are described as a wheel-shaped structure, but in actual scenarios, the walking mechanism 104 is not limited to this shape, and can be a crawler-type, Mecanum-type, or other shape structure suitable for complex terrains in a pig house; the lifting rod 7 and the lifting motor 9 constitute a lifting module 106; In the present invention, reference may be made to Figure 2 As mentioned above, the vehicle body bottom plate, support column and support plate provide support, dust and pollution prevention, etc., protect the internal devices and circuits, and avoid contact with poultry animals. The wheels provide support and steering. In addition, the power supply is a 24V rechargeable battery, which provides power for the entire device. For different voltage requirements, corresponding step-down modules can be used to provide the required power for different devices.

[0028] In the present invention, the sensor module 101 is used to sense the scene environment of the target breeding house. The module can integrate multiple sensors, such as laser radar integrated infrared devices. These sensors can obtain real-time information such as the spatial layout, obstacle locations, light conditions, and animal activity range in the breeding house. Through complex algorithm processing, these raw data are converted into environmental perception data to provide basic information for subsequent decision-making.

[0029] The multispectral acquisition module 102 is composed of a multispectral camera and has the ability to capture spectral information in different bands. In the breeding house environment, different materials (such as soil, feces, feed residues, etc.) will show different reflection characteristics under specific spectra. By capturing these subtle differences, the multispectral camera can accurately identify which areas need to be cleaned and the difficulty of cleaning, thereby generating detailed multispectral data. This data is crucial for formulating cleaning strategies.

[0030] The data processing module 103 is mainly composed of a microprocessor. After receiving data from the sensor module 101 and the multi-spectral acquisition module 102, it generates target walking path information and target cleaning area information through algorithm analysis. Among them, the target walking path information is used to guide the walking mechanism on how to avoid obstacles and efficiently reach the target area to be cleaned; and the target cleaning area information includes nozzle control data and lifting height data. The nozzle control data determines the mode of cleaning water spraying (such as spraying intensity, coverage, etc.), and the lifting height data adjusts the position of the nozzle according to the height and shape of the area to be cleaned.

[0031] The walking mechanism 104 is a moving part of the farm house control device. It drives the entire device to move flexibly in the farm house according to the target walking path information generated by the data processing module 103. In practical applications, the walking mechanism 104 can be more diversified, such as crawler type, Mecanum wheel type, etc., to adapt to the complex terrain of the farm house and the environment of animal activities. The walking mechanism 104 ensures the stability and flexibility of the farm house control device so that it can operate efficiently in narrow spaces and irregular terrains.

[0032] The nozzle 105 is a key component for performing the cleaning task. The spraying mode of the cleaning water can be adjusted according to the nozzle control data provided by the data processing module 103. These modes may include different water pressures, flow rates, spraying angles and coverages to meet the cleaning requirements of different materials and pollution levels. In the present invention, the nozzle 105 is usually installed on the cross arm side of the lifting module 106 so as to maintain accurate coverage of the target area while adjusting the spraying height.

[0033] The lifting module 106 is mainly composed of a lifting rod and a lifting motor. The module adjusts the vertical position of the nozzle according to the lifting height data provided by the data processing module 103. This function is essential for cleaning areas at different heights, such as walls, railings, and raised beds. The precise control of the lifting module 106 ensures that the cleaning water can accurately reach the target area while avoiding unnecessary waste and interference with the surrounding environment.

[0034] The cleaning and disinfection control device for aquaculture houses provided by the present invention performs environmental perception and stain detection through hybrid sensing technology, designs and integrates lidar, infrared sensors and multi-spectral cameras, generates a detailed three-dimensional map of the aquaculture houses, and combines historical cleaning data to predict potential high-stain areas using a stain prediction algorithm, thereby comprehensively analyzing the surface characteristics and stain distribution of the area to be cleaned in the aquaculture houses, to ensure accurate detection and identification of stains, and improve the cleaning efficiency of the aquaculture houses.

[0035] Based on the above embodiment, the lifting module includes a lifting rod and a cross arm, and the top end of the lifting rod is connected to the middle end of the cross arm through a first rotating mechanism, and the first rotating mechanism is used to adjust the height of the lifting module according to the lifting height data; the nozzle is provided with a second rotating mechanism, and the second rotating mechanism is used to adjust the spraying angle and spraying direction of the nozzle according to the cleaning water spraying mode.

[0036] Figure 3 A schematic diagram of the structure of the nozzle provided by the present invention, Figure 4 For the structural diagram of the lifting module provided by the present invention, please refer to Figure 3 and Figure 4 As shown, in the present invention, two nozzles are arranged on both sides of the cross arm of the lifting module, and the two nozzles are rotated by a rotating mechanism (i.e., a first rotating mechanism, such as Figure 3 Each nozzle can be adjusted by the rotating pair shown in the figure, and its spraying angle and direction can be controlled separately, thereby achieving the following functions: All-round cleaning: The nozzle can rotate and adjust the spraying direction to cover all areas in the breeding house (such as pig house), whether it is the floor or the wall, to ensure that there is no dead corner cleaning.

[0037] Focused cleaning: When a dirty area is detected, the nozzle spraying time and water pressure can be increased to focus on cleaning the area to ensure that the stains are completely removed.

[0038] The lifting module is an important component of the pig house cleaning control device, which is used to adjust the cleaning height. When adjusting the height, the lifting rod can adjust the height of the nozzle as needed, so that the nozzle can reach areas at different heights for cleaning. For example, when cleaning the walls and ceilings of the pig house, the lifting rod will raise the nozzle; when cleaning the floor, the lifting rod will lower the nozzle. Figure 4 As shown, the lifting rod and the cross arm are connected by a second rotating mechanism (i.e. Figure 4 In the present invention, the height adjustment of the lifting rod is automatically completed by the cleaning control of the breeding house, and combined with the sensor data, it ensures that the nozzle is always in the best cleaning position to improve the cleaning efficiency.

[0039] On the basis of the above embodiment, the breeding house cleaning and disinfection control device further includes an image acquisition module, which is used to acquire images of the target cleaning area to obtain images of the area before cleaning and images of the area after cleaning; The data processing module is also used to compare the area image before cleaning and the area image after cleaning, and if the comparison result does not meet the preset cleaning conditions, generate a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray cleaning water on the target area to be cleaned again.

[0040] In the present invention, the cleaning and disinfection control device for the breeding house further includes an image acquisition module (i.e. Figure 1 The image acquisition module is composed of a high-resolution camera or image sensor, which is specially designed to capture image information of the target cleaning area. Before the cleaning operation begins, the image acquisition module will take pictures of the area to be cleaned and generate images of the area before cleaning. These images not only record the degree and distribution of dirt before cleaning, but also provide an intuitive comparison benchmark for subsequent cleaning effect evaluation.

[0041] After the cleaning operation is completed, the image acquisition module will take pictures of the same area again to generate images of the cleaned area. These images show the actual effect after the cleaning operation, including the degree of cleanliness and whether there are any dirty areas missed.

[0042] In the present invention, the data processing module is the core part of the cleaning and disinfection control device of the breeding house, which is used to process and analyze data from various modules, including the images of the area before and after cleaning provided by the image acquisition module. When processing these images, the data processing module will use image recognition and comparison algorithms to make detailed comparisons of the images before and after cleaning. In the present invention, the images before and after cleaning are compared for the purpose of evaluating whether the effect of the cleaning operation has reached the preset cleaning conditions, which include but are not limited to the cleanliness reaching a certain specific standard, the dirty area being completely covered and effectively cleared, etc. If the comparison result shows that the cleaning effect does not meet the preset conditions, that is, there is still obvious dirt or uneven cleaning in the image of the cleaned area, the data processing module will immediately generate a secondary cleaning instruction.

[0043] In the present invention, the secondary cleaning instruction contains all the information required for re-cleaning, such as the spraying pattern of the nozzle, the flow rate and pressure of the cleaning water, and the specific location of the target area to be cleaned, etc. This instruction will be sent to the nozzle and the related driving mechanism to drive them to spray the cleaning water again on the target area to be cleaned to ensure the expected cleaning effect.

[0044] The present invention introduces the secondary cleaning function of the image acquisition module and the data processing module, so that the cleaning control device of the breeding house can achieve more accurate and efficient cleaning operations, which not only improves the cleaning quality, but also saves water resources and energy and reduces operating costs by reducing unnecessary repeated cleaning. At the same time, it also provides a strong guarantee for the sanitation management of the breeding house and the health of the animals.

[0045] Figure 5 This is a schematic diagram of the overall structure of the breeding house cleaning and disinfection control device provided by the present invention, which can be referred to Figure 5 As shown in the figure, the cleaning and disinfection control device of the breeding house contains multiple key components and modules, such as valve control, RFID tags, multi-point potentiometers and magnetic stripe sensors. These components and modules each have different functions and together constitute the core part of the control system.

[0046] Specifically, valve control modules can be used to control the flow of fluids, such as water spraying or gas discharge. RFID tags are used to identify and track objects or equipment in the system to achieve automated management and monitoring. Multi-point potentiometers can be used to measure or adjust the voltage or current in the device to ensure stable operation of the device. Magnetic stripe sensors are used to detect the position or direction of the magnetic stripe, providing key information for navigation and positioning of the device.

[0047] Furthermore, the sensor layer is an important component of the device, which contains multiple sensors for real-time monitoring of various parameters in the system, such as temperature, humidity, pressure, etc. The water spray rotation module can be used for operations such as cleaning or disinfection, and achieves full coverage through rotating nozzles. The remote control device allows the operator to control the operation of the system from a remote location, improving the flexibility and safety of the operation.

[0048] In the present invention, the high-performance microcontroller STM32F4 is used as the central controller of the device to receive data from various sensors and modules, process and analyze them, and issue corresponding control instructions based on the analysis results. These instructions are transmitted to various actuators or functional modules through the data flow path, thereby realizing the automatic control of the system.

[0049] In the present invention, the collision detection module is used to detect whether objects or equipment in the device collide to avoid potential safety hazards. The server layer can be used to store and process large amounts of data to provide support for the device's decision-making. The display screen provides an intuitive user interface that allows the operator to monitor the status and performance of the system in real time. In the present invention, two common communication interfaces, RS232 and RS485, are used to achieve data exchange and communication between devices. The disinfection system module can be used for cleaning and disinfection operations to ensure the hygiene and safety of the device. The power supply module provides a stable power supply for the entire device and is the basis for the normal operation of the system.

[0050] The magnetic stripe navigation module is used to provide navigation and positioning information for the device, ensuring that the device can move and operate according to the predetermined path. The human-computer interaction module provides an interactive interface between the user and the system, allowing the user to input instructions and view system status information.

[0051] The chassis, peripherals and running gear modules constitute the mobile part of the device, which is used to enable the system device to move and operate in different environments. The design and configuration of these modules depend on the specific application scenarios and requirements of the device.

[0052] In the present invention, the cleaning and disinfection control device of the breeding house mainly includes a sensor system and a control system. The sensor system is mainly used to collect various information data of the cleaning robot during its working process. The main data include the robot position, path status, water pressure level, cleaning area, environmental humidity and water temperature, etc. The above data can extract the real-time operation status of the cleaning robot by setting the frequency, and some data can also be temporarily extracted by user operation.

[0053] The control system is mainly responsible for controlling some key operating devices, including the start and stop of the high-pressure water pump, nozzle position and angle adjustment, speed and direction control of the travel motor, etc. The remote monitoring control system is judged by the user or the host. When some device operations are required during the cleaning process, the control system of the cleaning robot will control the corresponding equipment to operate and realize timely adjustment of the cleaning process.

[0054] The cleaning robot in the present invention is equipped with a high-precision depth camera and a variety of sensors, which can detect and accurately determine the dirty areas in the breeding house in real time. During the cleaning process, the robot automatically increases the cleaning time and water pressure of key points, and dynamically adjusts the path through an intelligent algorithm to ensure comprehensive cleaning coverage. After each step is completed, the robot takes a photo inspection through the front camera of the pole (i.e., the front camera), evaluates the cleaning effect through image comparison, and automatically performs secondary cleaning for areas that do not meet the standards. At the same time, the control system manages the high-pressure water pump, nozzle and travel motor, and combines remote monitoring to achieve closed-loop negative feedback control. When the sensor system detects that there is a lot of dirt in a certain area, the control system will automatically increase the cleaning time and water pressure in that area to ensure the cleaning effect. The waterproof design of the present invention meets the IP66 standard to ensure the stable operation of the device in a high humidity environment.

[0055] Figure 6 This is a schematic diagram of the control strategy of the farm house cleaning and disinfection control device provided by the present invention, which can be referred to Figure 6 As shown, in the present invention, the 48V DC power supply provides a stable DC power supply for the entire device, and the lithium battery serves as a backup or auxiliary power supply to ensure that the device can still operate for a period of time when the main power supply of the device fails.

[0056] Two rotary servos are used to drive related components to perform rotational movements to achieve specific operations or adjustments. The traveling motor drives the entire device or equipment to walk or move on a specified path.

[0057] The site reading and writing device is used to read and write site information, which is related to the location or status record of the device. The RFID reader identifies the identity information of different devices or items by reading RFID tags. Among them, the data communication within the system or with other devices is realized through the RS485 communication interface.

[0058] As the core processing unit of the device, the MCU controller is responsible for receiving input signals, executing control algorithms, and outputting control instructions.

[0059] Magnetic navigation DI detects the presence and position of the magnetic stripe through a magnetic stripe sensor, and uses the magnetic field information to guide the device along a predetermined path.

[0060] Furthermore, the switch state of the valve is controlled by the relay to achieve fluid or gas control. The collision monitoring DI can monitor whether the equipment has collided based on the external collision sensor to ensure safe operation.

[0061] For some other auxiliary components, such as water valve, used to control the flow or stop of water; 4G module, used to realize long-distance wireless communication, used for data transmission or remote control, etc., together ensure the stable operation and efficient operation of the system.

[0062] In the present invention, when the cleaning robot receives the cleaning instruction, it immediately activates its sensor system to scan the environment. In this step, the laser radar (LiDAR) and infrared sensor work together to generate a detailed three-dimensional map of the livestock and poultry house and accurately identify all areas to be cleaned.

[0063] Next, the cleaning robot uses the generated 3D map and historical cleaning data to build an environmental model and predict potential high-stain areas through a stain prediction algorithm, thereby developing a reasonable cleaning plan. In order to have a more comprehensive understanding of the area to be cleaned, the cleaning robot activates the multispectral camera system to collect images under the visible light, infrared and ultraviolet spectra. These image data will be used to analyze the surface characteristics of the area and the distribution of stains.

[0064] The cleaning robot then inputs the multispectral image data into a multi-task learning model to identify and classify stains. The model can accurately identify different types of stains and assess their severity, providing strong support for subsequent cleaning work.

[0065] In the present invention, based on the results of stain identification and classification, the cleaning robot sets a dynamic cleaning priority for each stain area to ensure that resources are optimally allocated. In terms of path planning, the cleaning robot uses an adaptive grid generation algorithm to dynamically generate grids of different sizes according to the stain type and cleaning requirements, thereby optimizing the cleaning path and strategy. At the same time, the cleaning robot also uses a path generation and simulation system to evaluate the virtual path and select the optimal cleaning path.

[0066] During the cleaning process, the two nozzles equipped with the cleaning robot can dynamically adjust the configuration and water jet parameters according to the type and distribution of stains to ensure the best cleaning effect; at the same time, the cleaning robot also monitors the cleaning effect in real time through sensors, and evaluates and adjusts the cleaning effect after each stage.

[0067] After the cleaning task is completed, the cleaning robot uses multi-spectral cameras and chemical sensors to verify the area in multiple dimensions to ensure that all stains have been completely removed. If residual stains are found, the cleaning robot will automatically adjust the cleaning strategy and clean again.

[0068] Finally, when the cleaning robot detects that all areas have been cleaned, it will automatically generate a cleaning report containing comparison images before and after cleaning, cleaning path, time and energy consumption data, and send it to the user terminal via the wireless network, thereby achieving efficient and thorough cleaning tasks in the complex fattening house environment, providing strong guarantees for the hygienic management of livestock and poultry houses.

[0069] Figure 7 A schematic diagram of the flow chart of the method for generating a disinfection strategy provided by the present invention, such as Figure 7 As shown, the present invention provides a method for generating a disinfection strategy for the disinfection control device for a breeding house according to the above embodiments, comprising: Step 701, obtaining environmental perception data of a target breeding house; Step 702, constructing initial walking path information according to the environmental perception data; Step 703, obtaining multispectral data of each area to be cleaned in the target breeding house; Step 704, inputting the multispectral data into a stain prediction model to obtain a stain recognition result of the area to be cleaned output by the stain prediction model, wherein the stain prediction model is trained by sample spectral data marked with stain category labels; Step 705, obtaining cleaning priority information of each of the to-be-cleaned areas according to the degree of stain corresponding to the stain identification result, and adjusting the initial walking path information according to the cleaning priority information to obtain target walking path information; Step 706, based on the target walking path information, determine the current target area to be cleaned, and obtain the target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned; Step 707: spray cleaning water on the target area to be cleaned according to the target walking path information and the target cleaning area information.

[0070] In the present invention, multiple sensors in the sensor module (such as laser radar integrated infrared device) start working, providing real-time information such as the spatial layout, obstacle location, light conditions, and animal activity range in the breeding house. The laser radar can accurately measure distance and draw three-dimensional maps, while the infrared device can detect heat distribution and help identify the animal's activity area. These raw data are processed by algorithms and converted into environmental perception data, which include but are not limited to the spatial structure in the breeding house, the precise location of obstacles, the distribution of light intensity, and the possible activity trajectories of animals, providing basic information for the subsequent construction of walking paths and the formulation of cleaning strategies.

[0071] Furthermore, the initial walking path information is constructed based on the environmental perception data analysis, and one or more feasible paths from the starting point to the end point are generated based on the obstacle positions and spatial layout in the environmental perception data. The initial walking path information is intended to ensure that the cleaning device can avoid obstacles and efficiently and safely reach each area to be cleaned in the breeding house.

[0072] Furthermore, the multispectral acquisition module collects spectral information of each area to be cleaned in the breeding house in different bands. This spectral information can reflect the reflection characteristics of different materials (such as soil, feces, feed residues, etc.), thereby helping to identify which areas need to be cleaned and how difficult it is to clean them. The multispectral camera captures these subtle differences and generates detailed multispectral data, providing key information for subsequent stain identification and cleaning strategy formulation.

[0073] In the present invention, multispectral data is input into a pre-trained stain prediction model. This model is trained based on a large amount of sample spectral data marked with stain category labels, and can identify the type, distribution and degree of stains. The stain prediction model analyzes the difference in reflective characteristics in the multispectral data and outputs the stain recognition results of the area to be cleaned, including information such as the type of stain, distribution range and possible cleaning difficulty.

[0074] Based on the stain recognition results, the data processing module calculates the cleaning priority of each area to be cleaned, which involves a comprehensive assessment of the degree of stain, distribution range, and cleaning difficulty. Then, based on this priority information, the initial walking path is adjusted, including replanning the path to give priority to areas with higher stain levels, or adjusting the access order according to the cleaning difficulty. Finally, the target walking path information is obtained to ensure that the cleaning device can visit each area to be cleaned in the optimal order to improve cleaning efficiency and effect.

[0075] Furthermore, the current target area to be cleaned is determined according to the target walking path information. Then, the multispectral acquisition module collects multispectral data of the area again to obtain more accurate data on the degree of stains and the distribution of stains. These data are input into the data processing module, and the target cleaning area information is generated through algorithm analysis. In the present invention, the target cleaning area information includes nozzle control data and lifting height data, which are used to instruct the cleaning device on how to clean the target area to be cleaned.

[0076] Finally, the cleaning task is started according to the target walking path information and the target cleaning area information. The walking mechanism drives the cleaning device to move to the target area to be cleaned according to the target walking path information. At the same time, the nozzle adjusts the spraying mode of the cleaning water (such as spraying intensity, coverage, etc.) according to the nozzle control data in the target cleaning area information. The lifting module adjusts the vertical position of the nozzle according to the lifting height data to ensure that the cleaning water can accurately reach the target area. Through this series of operations, the cleaning device can efficiently and accurately complete the cleaning task of the target area to be cleaned.

[0077] The cleaning and disinfection strategy generation method provided by the present invention uses hybrid sensing technology to perform environmental perception and stain detection, designs and integrates lidar, infrared sensors and multispectral cameras, generates a detailed three-dimensional map of the breeding house, and combines historical cleaning data to use a stain prediction algorithm to predict potential high-stain areas, thereby comprehensively analyzing the surface characteristics and stain distribution of the area to be cleaned in the breeding house, so as to ensure accurate detection and identification of stains and improve the cleaning efficiency of the breeding house.

[0078] Based on the above embodiment, the stain prediction model is trained by the following steps: Obtain sample multispectral data from each area of ​​the breeding house; According to the stain category, stain degree and stain distribution in the sample multispectral data, marking the sample multispectral data with the corresponding stain category label to obtain a multispectral data sample set; The multi-task learning model is trained by using the multi-spectral data sample set, and after determining that the training results meet the preset training conditions, the stain prediction model is obtained.

[0079] In the present invention, in order to collect multispectral data of samples in various areas of the breeding house, it is necessary to use professional multispectral acquisition equipment (such as multispectral cameras) to comprehensively scan and record the breeding house. These devices can capture spectral information in different bands for subsequent training of the model on the type, degree and distribution of stains.

[0080] During the collection process, it is necessary to ensure that the sample data is representative and diverse, that is, to cover all possible types of stains in the breeding house (such as soil, feces, feed residues, etc.), as well as areas with different stain degrees and distribution states. At the same time, in order to enhance the generalization ability of the model, sample data under different lighting conditions and different material surfaces (such as cement floors, wooden boards, metal railings, etc.) should also be included.

[0081] Furthermore, after obtaining the sample multispectral data, these data are labeled. This step can be automatically labeled using the existing knowledge base. The labeled content includes the type of stain (such as soil, feces, etc.), the degree of stain (such as mild, moderate, severe) and the distribution of stains (such as dots, strips, flakes, etc.). The labeling process requires high accuracy and consistency to ensure the effectiveness of subsequent model training.

[0082] After completing the labeling of the sample multispectral data, a multispectral data sample set is obtained. This sample set contains the stain information of each area in the breeding house and the corresponding stain category labels, which are used for subsequent multi-task learning model training. In the present invention, the multi-task learning model can learn multiple related tasks at the same time to improve the generalization ability and performance of the model, that is, the multi-task learning model can be designed to simultaneously predict the category, degree and distribution of stains.

[0083] During the training process, the multispectral data sample set is input into the multi-task learning model, and the model parameters are adjusted through the back-propagation algorithm to minimize the prediction error. In addition, the model is regularly verified and tested to ensure that its performance meets expectations. This process can use a portion of sample data that is not involved in the training to evaluate the prediction accuracy of the model. If the performance of the model is poor, the model structure, parameters or training strategy need to be adjusted.

[0084] After multiple rounds of training and adjustment, when the performance of the model reaches the preset training conditions (such as the prediction accuracy reaches a certain threshold, the training loss converges, etc.), the model can be considered to have been trained and the final stain prediction model is obtained. The stain prediction model can be used to predict new multispectral data to identify the type, degree and distribution of stains, so as to achieve smarter and more efficient execution of cleaning tasks.

[0085] On the basis of the above embodiment, the step of obtaining the target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned includes: According to the mapping relationship between the degree of stain and the water jet parameters, the target water jet parameters are obtained, wherein the target water jet parameters at least include nozzle water pressure parameters and cleaning water flow parameters; Determine nozzle configuration information according to the stain distribution state data, wherein the nozzle configuration information at least includes nozzle type information, nozzle angle data and nozzle position data; Obtaining nozzle control data according to the target water jet parameters and the nozzle configuration information; Determining lifting height data according to the positional relationship between the target area to be cleaned and the nozzle; The target cleaning area information is obtained according to the nozzle control data and the lifting height data.

[0086] In the present invention, during the cleaning process of the breeding house, the degree of stains (such as light, moderate, and severe) will directly affect the strength required for cleaning. In order to complete the cleaning task efficiently and without excessive waste of resources, it is necessary to adjust the parameters of the water jet according to the degree of stains. These parameters include at least the water pressure parameters of the nozzle and the flow parameters of the cleaning water.

[0087] Specifically, a mapping relationship between the degree of stain and water jet parameters can be established through a large amount of experimental data or experience accumulation. For example, for heavy stains, higher water pressure and greater flow rate may be required to ensure the cleaning effect; while for light stains, the water pressure and flow rate can be appropriately reduced to reduce the waste of water resources.

[0088] After determining the degree of contamination, appropriate target water jet parameters can be selected based on this mapping relationship. These parameters will directly guide the nozzles of the cleaning equipment on how to spray cleaning water to achieve the best cleaning effect.

[0089] In addition to the water jet parameters, the configuration of the nozzle is also an important factor affecting the cleaning effect, including the type, angle and position of the nozzle. The distribution data of the stains (such as dots, strips, flakes, etc.) can provide information about the shape, size and location of the stains. Based on this information, the most appropriate nozzle configuration can be determined. For example, for dot stains, you may need to use a more focused nozzle and aim the nozzle at the center of the stain; for strip or flake stains, you may need to use a nozzle with a wider spray range and adjust the angle of the nozzle to cover the entire stain area.

[0090] The position data of the nozzle involves the installation position of the nozzle on the cleaning device and its positional relationship relative to the area to be cleaned. These position data can ensure that the nozzle can accurately cover the target stain area and avoid unnecessary interference with the surrounding area.

[0091] Furthermore, after determining the target water jet parameters and nozzle configuration information, the nozzle control data can be calculated, including the nozzle opening time, closing time, water pressure adjustment instructions, flow adjustment instructions, etc. These data will be sent directly to the control device of the cleaning equipment to control the nozzle's spraying behavior.

[0092] During the cleaning process, the positional relationship between the nozzle and the target area to be cleaned is also very important, especially when the area to be cleaned is at different heights (such as walls, etc.), the lifting height of the nozzle needs to be adjusted to ensure that the cleaning water can accurately reach the target area. The lifting height data can be obtained by measurement or calculation. For example, a sensor can be used to measure the actual distance between the nozzle and the target area, and the lifting height of the nozzle can be adjusted according to this distance; or, the appropriate lifting height can be determined by calculation based on the shape and height information of the area to be cleaned.

[0093] Finally, by combining the nozzle control data and the lifting height data, we can get the complete target cleaning area information, which is used to guide the cleaning equipment on how to clean the target area to be cleaned. In actual operation, the cleaning equipment will adjust the nozzle's spraying behavior and lifting height according to this information to ensure that the cleaning water can accurately and efficiently cover the entire target area. At the same time, this information can also be used to monitor and record the cleaning process for subsequent analysis and optimization.

[0094] On the basis of the above embodiment, after spraying the target area to be cleaned with cleaning water according to the target walking path information and the target cleaning area information, the method further includes: Respectively acquiring a region image before cleaning and a region image after cleaning of the target region to be cleaned; Comparing the image of the area before cleaning and the image of the area after cleaning, if the comparison result does not meet the preset cleaning condition, generating a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray the cleaning water on the target area to be cleaned again; If the comparison result meets the preset cleaning condition, the walking mechanism is driven to the next area to be cleaned according to the target walking path information.

[0095] In the present invention, during the process of automatically cleaning the breeding house, in order to ensure the cleaning effect and monitor the cleaning progress, it is necessary to obtain images of the target area to be cleaned before and after cleaning, which can be achieved by a camera or image acquisition device installed on the cleaning equipment.

[0096] The image of the area before cleaning is used to record the stain condition before cleaning, including the type, distribution and degree of the stain. The image of the area after cleaning is taken after the cleaning operation is completed and is used to evaluate the cleaning effect. By comparing with the image before cleaning, it can be intuitively seen whether the cleaning has achieved the expected effect.

[0097] In the present invention, after obtaining the regional images before and after cleaning, it is necessary to compare and analyze the two images. The comparison process can be based on pixel-level comparison or can be to identify the stain changes before and after cleaning through an image recognition algorithm. In order to determine whether the cleaning meets the preset cleaning conditions, analysis can be performed from many aspects, including the removal rate of stains, the surface cleanliness after cleaning, whether there are residual stains, etc. In the present invention, corresponding cleaning conditions can be set according to specific cleaning requirements and standards.

[0098] If the comparison between the image of the area after cleaning and the image before cleaning shows that the cleaning effect does not meet the preset cleaning conditions, a secondary cleaning instruction needs to be generated, which will drive the nozzle of the cleaning equipment to spray cleaning water on the target area to be cleaned again. The secondary cleaning instruction includes adjusting the nozzle's water pressure, flow rate, angle and other parameters, as well as determining the traversal path and cleaning time for the second cleaning, to ensure that the target area can meet the cleaning requirements.

[0099] In the present invention, when generating the secondary cleaning instruction, factors such as the operating status of the cleaning equipment, energy consumption, and cleaning efficiency also need to be considered to achieve rational use of resources while ensuring the cleaning effect.

[0100] If the comparison result of the image of the area after cleaning and the image before cleaning shows that the cleaning effect has met the preset cleaning conditions, then it can be considered that the cleaning task of the target area to be cleaned has been completed. At this time, according to the preset target walking path information, the walking mechanism of the cleaning equipment is driven to move to the next area to be cleaned. In the present invention, the target walking path information includes the moving path of the cleaning equipment in the breeding house, the order of each area to be cleaned, and the moving speed.

[0101] Before moving to the next area to be cleaned, some necessary adjustments and maintenance work can be done on the cleaning equipment to ensure that it can continue to operate stably. At the same time, it is also necessary to constantly monitor and evaluate the cleaning effect to ensure that the entire cleaning process can achieve the expected results.

[0102] On the basis of the above embodiment, comparing the region image before cleaning and the region image after cleaning includes: Acquire a residual stain region image according to a cleaning region image corresponding to the target cleaning region, wherein the cleaning region image includes a region image before cleaning and a region image after cleaning; Obtaining the number of stain pixels corresponding to each residual stain area in the residual stain area image; In the case where it is determined that the number of stain pixels is greater than a preset stain pixel threshold, the corresponding residual stain area is marked as a residual stain area to be cleaned, and if the number of the residual stain areas to be cleaned is greater than the preset residual stain number, the secondary cleaning instruction corresponding to the target area to be cleaned is generated; When it is determined that the number of the stain pixels is less than or equal to the preset stain pixel threshold, the corresponding residual stain area is marked as a non-cleaned area; The method further comprises: Based on piecewise linear grayscale transformation, the cleaning area image is subjected to image enhancement processing according to the grayscale information and grayscale interval information of the cleaning area image to obtain the cleaning area image after image enhancement processing.

[0103] In the present invention, when collecting images of the cleaning area, the image processing process is used to further determine whether there are still local stains in the breeding house after the overall scrubbing. Specifically, according to actual data after multiple cleanings, the main stains that still exist in the breeding house area after the overall scrubbing can be roughly divided into three categories, namely, large pieces of feces, dark spots and strips of dust marks.

[0104] Taking pig house stains as an example, there may be some relatively small areas in the pig house stains extracted after stain identification. These small stain areas do not actually affect normal breeding processing. Therefore, the detection accuracy of residual stains in the present invention is set to 40mm², and 2 pixels are used to represent 10mm, which represents 16 pixels in the collected image. The number of pixels is used as the threshold (that is, the preset stain pixel threshold) to determine whether the number of stain pixels contained in the stain connected domain in the image result after cleaning meets n 16. If it is larger, the area is retained and it is determined that it needs to be cleaned again; otherwise, the stain area is considered to be too small and no cleaning is required.

[0105] Further, in order to obtain better features and visual effects, the obtained image is subjected to image enhancement processing. The present invention adopts piecewise linear gray-scale transformation to expand the gray level of the required image details, while enhancing its contrast, compressing the gray level of the unnecessary image details, so as to make a distinction. In one embodiment, it is assumed that the gray level of the input image f(x, y) is from 0 to M levels, and the gray level of the enhanced image g(x, y) is from 0 to N levels. The intervals [a, b] and [c, d] are respectively a certain gray interval of the source image (i.e., the input original image) and the enhanced image. The piecewise linear transformation function is: ; Among them, when different values are taken for a, b, c, and d, different effects can be obtained. If a = c and b = d, the gray-scale transformation function is a straight line with a slope of 1, which means that the enhanced image is exactly the same as the original image; If a > c and b < d, in the intervals [0, a] and [b, M], the dynamic range of the gray-scale values of the source image decreases; in the interval [a, b], the dynamic range of the gray-scale values of the source image increases, so the contrast in the middle range is enhanced.

[0106] If a < c and b > d, in the intervals [0, a] and [b, M], the dynamic range of the source image increases, and at the same time, in the interval [a, b], the dynamic range of the source image decreases.

[0107] From the above three cases, it can be seen that the values of a, b, c, and d can directly affect the slope of the segmentation, and any gray interval can be compressed or expanded by adjusting a, b, c, and d.

[0108] Figure 8 The structural schematic diagram of the electronic device provided by the present invention is shown in Figure 8As shown, the electronic device may include: a processor (Processor) 801, a communication interface (Communications Interface) 802, a memory (Memory) 803 and a communication bus 804, wherein the processor 801, the communication interface 802, and the memory 803 communicate with each other through the communication bus 804. The processor 801 can call the logic instructions in the memory 803 to execute the disinfection strategy generation method, which includes: obtaining environmental perception data of the target breeding house; constructing initial walking path information according to the environmental perception data; obtaining multispectral data of each area to be cleaned in the target breeding house; inputting the multispectral data into the stain prediction model to obtain the stain recognition result of the area to be cleaned output by the stain prediction model, wherein the stain prediction model is trained by sample spectral data marked with a stain category label; according to the stain degree corresponding to the stain recognition result, obtaining the cleaning priority information of each area to be cleaned, and adjusting the initial walking path information according to the cleaning priority information to obtain the target walking path information; based on the target walking path information, determining the current target area to be cleaned, and obtaining the target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned; according to the target walking path information and the target cleaning area information, spraying cleaning water on the target area to be cleaned.

[0109] In addition, the logic instructions in the above-mentioned memory 803 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk.

[0110] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the disinfection strategy generation method provided by the above methods, and the method includes: obtaining environmental perception data of a target breeding house; constructing initial walking path information based on the environmental perception data; obtaining multispectral data of each area to be cleaned in the target breeding house; inputting the multispectral data into a stain prediction model to obtain a stain recognition result of the area to be cleaned output by the stain prediction model, and In the method, the stain prediction model is obtained by training sample spectral data marked with stain category labels; according to the stain degree corresponding to the stain identification result, the cleaning priority information of each of the to-be-cleaned areas is obtained, and according to the cleaning priority information, the initial walking path information is adjusted to obtain the target walking path information; based on the target walking path information, the current target to-be-cleaned area is determined, and according to the stain degree and stain distribution state data corresponding to the multi-spectral data of the target to-be-cleaned area, the target cleaning area information is obtained; according to the target walking path information and the target cleaning area information, the target to-be-cleaned area is sprayed with cleaning water.

[0111] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented by a processor to execute the disinfection strategy generation method provided by the above embodiments, the method comprising: obtaining environmental perception data of a target breeding house; constructing initial walking path information based on the environmental perception data; obtaining multispectral data of each area to be cleaned in the target breeding house; inputting the multispectral data into a stain prediction model to obtain a stain recognition result of the area to be cleaned output by the stain prediction model, wherein the stain prediction model is trained by sample spectral data marked with a stain category label; obtaining cleaning priority information of each area to be cleaned based on the stain degree corresponding to the stain recognition result, and adjusting the initial walking path information based on the cleaning priority information to obtain target walking path information; determining the current target area to be cleaned based on the target walking path information, and obtaining target cleaning area information based on the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned; and spraying cleaning water on the target area to be cleaned based on the target walking path information and the target cleaning area information.

[0112] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0113] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cleaning and disinfection control device for a breeding house, characterized in that: It includes a sensor module, a multi-spectral acquisition module, a data processing module, a walking mechanism, a nozzle and a lifting module, wherein the nozzle is arranged on the cross arm side of the lifting module, wherein: The sensor module is used to sense the scene environment of the target breeding house and obtain the environmental perception data of the target breeding house; The multispectral acquisition module is used to collect multispectral data of each area to be cleaned in the target breeding house; The data processing module is used to generate corresponding target walking path information and target cleaning area information according to the environmental perception data and the multi-spectral data, wherein the target cleaning area information includes nozzle control data and lifting height data; The walking mechanism is used to drive to the target area to be cleaned in the target breeding house according to the target walking path information; The nozzle is used to spray the target area to be cleaned with cleaning water according to the cleaning water spraying mode corresponding to the nozzle control data; The lifting module is used to adjust the cleaning water spraying height according to the lifting height data.

2. The cleaning and disinfection control device for aquaculture houses according to claim 1, characterized in that: The lifting module includes a lifting rod and a cross arm. The top end of the lifting rod is connected to the middle end of the cross arm via a first rotating mechanism. The first rotating mechanism is used to adjust the height of the lifting module according to the lifting height data. The nozzle is provided with a second rotating mechanism. The second rotating mechanism is used to adjust the spraying angle and spraying direction of the nozzle according to the cleaning water spraying mode.

3. The cleaning and disinfection control device for aquaculture houses according to claim 2, characterized in that: The breeding house cleaning and disinfection control device also includes an image acquisition module, which is used to collect images of the target cleaning area to obtain an image of the area before cleaning and an image of the area after cleaning; The data processing module is also used to compare the area image before cleaning and the area image after cleaning, and if the comparison result does not meet the preset cleaning conditions, generate a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray cleaning water on the target area to be cleaned again.

4. A method for generating a disinfection strategy based on the disinfection control device for a breeding house according to any one of claims 1 to 3, characterized in that: include: Obtain environmental perception data of the target breeding house; Constructing initial walking path information according to the environmental perception data; Acquire multispectral data of each area to be cleaned in the target breeding house; Inputting the multispectral data into a stain prediction model to obtain a stain recognition result of the area to be cleaned output by the stain prediction model, wherein the stain prediction model is trained by sample spectral data marked with stain category labels; According to the degree of stain corresponding to the stain identification result, the cleaning priority information of each of the to-be-cleaned areas is obtained, and according to the cleaning priority information, the initial walking path information is adjusted to obtain the target walking path information; Based on the target walking path information, determine the current target area to be cleaned, and obtain the target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned; According to the target walking path information and the target cleaning area information, cleaning water is sprayed on the target area to be cleaned.

5. The method for generating a decontamination strategy according to claim 4, characterized in that: The stain prediction model is trained by the following steps: Obtain sample multispectral data from each area of ​​the breeding house; According to the stain category, stain degree and stain distribution in the sample multispectral data, marking the sample multispectral data with the corresponding stain category label to obtain a multispectral data sample set; The multi-task learning model is trained by using the multi-spectral data sample set, and after determining that the training results meet the preset training conditions, the stain prediction model is obtained.

6. The method for generating a decontamination strategy according to claim 4, characterized in that: The step of obtaining target cleaning area information according to the stain degree and stain distribution state data corresponding to the multispectral data of the target area to be cleaned includes: According to the mapping relationship between the degree of stain and the water jet parameters, the target water jet parameters are obtained, wherein the target water jet parameters at least include nozzle water pressure parameters and cleaning water flow parameters; Determine nozzle configuration information according to the stain distribution state data, wherein the nozzle configuration information at least includes nozzle type information, nozzle angle data and nozzle position data; Obtaining nozzle control data according to the target water jet parameters and the nozzle configuration information; Determining lifting height data according to the positional relationship between the target area to be cleaned and the nozzle; The target cleaning area information is obtained according to the nozzle control data and the lifting height data.

7. The method for generating a decontamination strategy according to claim 4, characterized in that: After spraying the target area to be cleaned with cleaning water according to the target walking path information and the target cleaning area information, the method further includes: Respectively acquiring a region image before cleaning and a region image after cleaning of the target region to be cleaned; Comparing the image of the area before cleaning and the image of the area after cleaning, if the comparison result does not meet the preset cleaning condition, generating a secondary cleaning instruction corresponding to the target area to be cleaned, wherein the secondary cleaning instruction is used to drive the nozzle to spray the cleaning water to the target area to be cleaned again; If the comparison result meets the preset cleaning condition, the walking mechanism is driven to the next area to be cleaned according to the target walking path information.

8. The method for generating a decontamination strategy according to claim 7, characterized in that: The comparing the region image before cleaning and the region image after cleaning includes: Acquire a residual stain region image according to a cleaning region image corresponding to the target cleaning region, wherein the cleaning region image includes a region image before cleaning and a region image after cleaning; Obtaining the number of stain pixels corresponding to each residual stain area in the residual stain area image; In the case where it is determined that the number of stain pixels is greater than a preset stain pixel threshold, the corresponding residual stain area is marked as a residual stain area to be cleaned, and if the number of the residual stain areas to be cleaned is greater than the preset residual stain number, the secondary cleaning instruction corresponding to the target area to be cleaned is generated; When it is determined that the number of the stain pixels is less than or equal to the preset stain pixel threshold, the corresponding residual stain area is marked as a non-cleaned area; The method further comprises: Based on piecewise linear grayscale transformation, the cleaning area image is subjected to image enhancement processing according to the grayscale information and grayscale interval information of the cleaning area image to obtain the cleaning area image after image enhancement processing.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the disinfection strategy generation method as described in any one of claims 4 to 8 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the decontamination strategy generation method as described in any one of claims 4 to 8 is implemented.

Images