A mouse catching system based on a YOLO mouse hole detection model

Abstract

The present application relates to the technical field of image processing, and more particularly to a mouse catching system based on a YOLO mouse hole detection model, which is provided with a feature acquisition module, a footprint analysis module, a mouse hole recognition module, a model determination output module, and a mouse catching regulation module. The feature acquisition module is used to mark suspected mouse holes, perform footprint recognition, and the footprint analysis module is used to screen associated footprint groups, determine the activity track of the suspected mouse hole, determine the activity tendency characteristic quantity through the mouse hole recognition module, screen characteristic footprints, identify characteristic activity tracks, screen characteristic mouse holes, output the division result of the mouse hole system through the model determination output module, and determine the growth tendency coefficient through the mouse catching regulation module to adjust the mouse catching frequency. The present application realizes the judgment of the activity of the mouse hole according to the intermittent footprints around the recognized mouse hole, performs joint analysis on the active mouse hole, adaptively adjusts the mouse catching frequency, and improves the reliability and flexibility of the mouse catching system.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a mouse-catching system based on the YOLO mouse hole detection model. Background Technology

[0002] Grassland rodent infestation is a global problem threatening ecological security and the sustainable development of animal husbandry. In recent years, the number of rodent burrows on grasslands has been increasing. The main species of grassland rodent infestation is Brandt's vole, whose burrow system contains multiple entrances. This not only exacerbates grassland desertification but also affects grassland yield. The increase in the number of rodents also poses a threat to people's lives. Traditional rodent monitoring relies on manual ground surveys, which require a lot of manpower and resources. Furthermore, burrow identification is affected by the experience of the surveyors, resulting in a high error rate. It is also impossible to track rodent activity in real time and monitor dynamic changes such as burrow expansion and migration, leading to delayed control measures. With the development of deep learning technology, YOLO-based target detection models have been gradually applied to burrow identification. However, factors such as changes in lighting and vegetation obstruction can lead to misjudgments. Moreover, the identified burrows cannot determine their activity level, making it impossible to accurately identify core burrows that require priority control. This affects the reliability and flexibility of the rodent trapping system. Therefore, improving the reliability of rodent activity monitoring and the flexibility of rodent trapping decision-making and control are urgent technical problems that need to be solved.

[0003] For example, Chinese patent application publication number CN117746504A discloses a behavior analysis device and method for modeling the skeleton of laboratory mice based on YOLO and ST-GCN algorithms. When the user uses the device in other places such as the breeding box of laboratory mice, the images or videos of laboratory mice are collected by a USB camera module that can be connected to the motherboard and then sent to the behavior analysis system. The system first uses the YOLO algorithm model to realize target detection and posture estimation, automatically captures dynamic skeletal points, and accurately obtains the skeletal point data of laboratory mice. Then, the ST-GCN algorithm model is used to perform dynamic skeleton modeling on the data obtained from the previous model, and obtains behavioral data such as target modification, feeding, and no less than five action modes, as well as centroid data, limb length, and movement trajectory, to realize the behavior analysis of laboratory mice.

[0004] The following problems still exist in the existing technology:

[0005] Existing technologies cannot determine the activity level of mouse holes based on intermittent imprints around the identified holes, nor can they perform joint analysis of active mouse holes or adaptively adjust the frequency of mouse trapping, thus affecting the reliability and flexibility of the mouse trapping system. Summary of the Invention

[0006] To address this, the present invention provides a mouse-catching system based on the YOLO mouse hole detection model, which overcomes the problems of existing technologies that cannot determine the activity of mouse holes based on the intermittent imprints around the identified mouse holes, cannot perform joint analysis of active mouse holes, and cannot adaptively adjust the mouse-catching frequency, thus affecting the reliability and flexibility of the mouse-catching system.

[0007] To achieve the above objectives, this invention provides a mouse-catching system based on the YOLO mouse hole detection model, comprising:

[0008] The feature acquisition module is used to mark suspected mouse holes based on the surface images of each sub-detection area on the grassland, and to perform imprint recognition within a first preset range set based on the suspected mouse holes;

[0009] The imprint analysis module, which is connected to the feature acquisition module, is used to filter several associated imprint groups of the suspected mouse hole based on the identified imprints, and to determine the activity trajectory of the suspected mouse hole based on the location distribution of the associated imprint groups.

[0010] The mouse hole identification module is connected to the feature acquisition module and the imprint analysis module respectively. It is used to determine the activity tendency characterization quantity and screen out the feature imprints based on the imprint distribution of the activity trajectory. It identifies the feature activity trajectory based on the activity tendency characterization quantity and the feature imprints, and screens the feature mouse holes based on the feature activity trajectory and the soil characterization quantity of the suspected mouse hole.

[0011] The model determination output module is connected to the mouse hole recognition module and is used to determine the division result of the mouse hole system based on the location distribution of each characteristic mouse hole;

[0012] The rodent control module is connected to the feature acquisition module and the model determination output module, respectively. It is used to determine the growth tendency coefficient based on the soil characterization of characteristic rodent holes in each rodent hole system, and to adjust the rodent trapping frequency based on the growth tendency coefficient, and push the rodent trapping frequency to the output end.

[0013] Furthermore, the imprint analysis module is used to filter several related imprint groups, wherein,

[0014] The imprint analysis module filters two adjacent imprints into an associated imprint group based on the determination result that the interval between two adjacent imprints meets the conditions for an associated imprint group.

[0015] The condition for the associated imprint group is that the distance between two adjacent imprints does not exceed a preset first distance threshold.

[0016] Furthermore, the imprint analysis module is used to determine the activity trajectory of the suspected mouse burrow, wherein,

[0017] The activity trajectory is determined based on several interconnected imprint vectors. The interconnected imprint vectors are constructed with the imprint with the largest distance between two adjacent imprints and the suspected mouse hole as the vector starting point and the imprint with the smallest distance between two adjacent imprints and the suspected mouse hole as the vector ending point.

[0018] Furthermore, the mouse hole recognition module is used to determine activity tendency representation quantities and feature imprints, wherein,

[0019] The mouse hole identification module is used to calculate the distance between each imprint in the activity trajectory and the suspected mouse hole, determine the maximum distance as the activity tendency characteristic of the activity trajectory, and determine the imprint with the minimum distance as the feature imprint.

[0020] Furthermore, the mouse hole recognition module is used to identify characteristic activity trajectories, wherein,

[0021] The mouse hole identification module is used to identify the activity trajectory as a characteristic activity trajectory based on the activity tendency representation quantity of the activity trajectory and the determination result that the characteristic imprint meets the condition of the characteristic activity trajectory.

[0022] The condition for the characteristic activity trajectory is that the activity tendency representation exceeds a preset activity tendency representation threshold, and the distance between the characteristic imprint and the suspected mouse hole does not exceed a preset second distance threshold.

[0023] Furthermore, the mouse burrow identification module is used to determine the soil characterization quantities of suspected mouse burrows, wherein,

[0024] The mouse hole identification module is used to determine the gray values ​​of each monitoring point in the sub-detection area where the suspected mouse hole is located and the gray values ​​within a second preset range based on the surface image. The absolute value of the difference between the minimum gray value within the second preset range and the average gray value within the sub-detection area is determined as the soil characterization value of the suspected mouse hole. The second preset range is determined based on the suspected mouse hole.

[0025] Furthermore, the mouse hole identification module is used to filter characteristic mouse holes, wherein,

[0026] The mouse hole identification module filters the suspected mouse holes as characteristic mouse holes based on the judgment results that the characteristic activity trajectory and soil characterization of the suspected mouse holes meet the characteristic mouse hole conditions;

[0027] The characteristic mouse hole condition is that the number of characteristic activity trajectories exceeds a preset reference value, and the soil characterization quantity exceeds a preset soil characterization reference value.

[0028] Furthermore, the model determination output module is used to classify each characteristic mouse hole into different mouse hole systems, wherein,

[0029] The model determination output module calculates the distance between any feature mouse hole and other feature mouse holes, and classifies feature mouse holes whose distance does not exceed the preset third distance threshold into the same mouse hole system. Each feature mouse hole exists only in a unique mouse hole system.

[0030] Furthermore, the mouse-catching control module is used to determine the growth tendency coefficient, wherein,

[0031] The rodent control module selects the characteristic rodent burrows as newly formed characteristic rodent burrows based on the soil characterization of the characteristic rodent burrows in the rodent burrow system, which meets the conditions for newly formed characteristic rodent burrows.

[0032] The condition for the newly formed characteristic mouse burrows is that the soil characterization amount exceeds the preset soil characterization threshold.

[0033] The growth tendency coefficient is the ratio of the number of newly formed characteristic mouse holes to the number of characteristic mouse holes in the mouse hole system.

[0034] Furthermore, the mouse-catching control module is used to adjust the frequency of mouse catching in the mouse burrow system, and the mouse-catching frequency is positively correlated with the growth tendency coefficient.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention sets up a feature acquisition module, an imprint analysis module, a mouse hole identification module, a model judgment output module, and a mouse trapping control module. The feature acquisition module acquires surface images of each sub-detection area on the grassland, marks suspected mouse holes based on the surface images, and performs imprint identification within a first preset range determined by the suspected mouse holes. The imprint analysis module filters several related imprint groups to determine the activity trajectory of the suspected mouse holes. The mouse hole identification module determines the activity tendency characterization quantity and filters out characteristic imprints. Based on the activity tendency characterization quantity and characteristic imprints, characteristic activity trajectories are identified and characteristic mouse holes are filtered. The model judgment output module outputs the division result of the mouse hole system. The mouse trapping control module determines the growth tendency coefficient and adjusts the mouse trapping frequency for the mouse hole system. Thus, it realizes the judgment of mouse hole activity based on the intermittent imprints around the identified mouse holes, performs joint analysis on active mouse holes, and adaptively adjusts the mouse trapping frequency, thereby improving the reliability and flexibility of the mouse trapping system.

[0036] In particular, this invention uses an imprint analysis module to determine the activity trajectory of suspected mouse holes based on the location distribution of associated imprint groups. It is understood that numerous depressions resembling mouse holes exist in the natural environment, such as wormholes and plant root cavities. The YOLO model's identification of mouse holes is easily affected by lighting and shadows, resulting in a high false positive rate. By analyzing the activity trajectory of rodents, false positives can be distinguished from real mouse holes. Simultaneously, the activity trajectory can provide data support for analyzing the activity level of mouse holes. When rodents move, they leave imprints on soft soil surfaces such as grasslands and farmland. The distance between adjacent imprints can characterize whether adjacent imprints belong to the same activity trajectory. By constructing associated imprint vectors based on adjacent imprints, the activity trajectory of rodents can be determined through the correlation between these vectors, thereby obtaining the rodent activity trajectory of suspected mouse holes identified by the YOLO model. This invention uses an imprint analysis module to determine the activity trajectory of suspected mouse holes based on the location distribution of associated imprint groups, thus realizing the identification of the activity trajectory of suspected mouse holes and improving the reliability and flexibility of the rodent-catching system.

[0037] In particular, this invention uses a mouse hole identification module to screen characteristic mouse holes based on characteristic activity trajectories and soil characterization of suspected mouse holes. It is understood that by comprehensively considering characteristic activity trajectories and soil characterization, real and active mouse holes can be more accurately distinguished from other similar structures. Relying solely on a single soil characterization or activity trajectory feature can lead to misjudgment. For example, some soil imprints formed by natural factors such as rainwater erosion may resemble mouse holes in grayscale values, but combining activity trajectory features can eliminate them. Similarly, some accidentally formed imprints resembling rodent activity trajectories will not be identified as characteristic mouse holes without corresponding soil characterization support. This comprehensive screening method effectively reduces errors. The system improves the accuracy of rat hole identification by increasing the number of characteristic activity tracks. The more characteristic activity tracks a rat hole has, the more frequently rats enter and exit it, indicating that the rat hole is in active use. For rodent control, these active rat holes are key targets for management. Screening out such characteristic rat holes allows for targeted rodent control measures such as trapping, improving control efficiency, reducing interference with non-target holes, better controlling rodent populations, and reducing the damage of rodents to the grassland ecosystem. Furthermore, it enables the determination of rat hole activity based on intermittent imprints around the identified rat holes, improving the reliability and flexibility of the rodent trapping system.

[0038] In particular, this invention outputs the classification results of mouse burrow systems based on the location distribution of each characteristic mouse burrow through the model judgment output module. It can be understood that dividing characteristic mouse burrows into different mouse burrow systems can more accurately assess the size of the mouse population in each mouse burrow system. Mouse burrows in the same system are usually used by one mouse population. After clarifying the system to which each mouse burrow belongs, mouse-catching resources can be rationally allocated according to the activity level and mouse population size of different mouse burrow systems, improving mouse-catching efficiency, avoiding waste of resources, and achieving precise strikes against different mouse-infested areas. By analyzing mouse burrows in the same system, the approximate size of the mouse population in that system can be inferred, and the mouse-catching frequency of that system can be adjusted in a timely manner to ensure the effectiveness of mouse-catching, effectively control the development of mouse infestation, and protect the original ecology in the grassland ecosystem. It can also balance and avoid the negative impact of excessive mouse-catching on other organisms and the environment. In this way, it realizes the joint analysis of active mouse burrows and improves the reliability and flexibility of the mouse-catching system.

[0039] In particular, this invention adjusts the mouse-catching frequency based on a growth tendency coefficient using a mouse-catching control module. This means that adjusting the mouse-catching frequency based on the growth tendency coefficient allows for real-time optimization of the mouse-catching strategy according to the reproductive dynamics of the mouse burrow system. When the growth tendency coefficient is high, it indicates a large number of newborn mice and active mouse reproduction; in this case, increasing the mouse-catching frequency can effectively control the expansion of the mouse population. If the growth tendency coefficient is low, the mouse-catching frequency can be appropriately reduced to avoid wasting resources, achieving precise and dynamic mouse-catching management, and greatly improving mouse-catching efficiency and effectiveness. This method adjusts the mouse-catching frequency based on the reproductive signs of the mouse burrow system. This method can prevent excessive rodent trapping from damaging the grassland ecosystem, ensuring the balance of grassland biodiversity. At the same time, it can accurately control the rodent population, avoiding the damage of rodent infestation to grassland vegetation, soil, and other ecological environments. It achieves a balance between rodent control and ecological protection. Since rodent trapping resources are limited, the growth tendency coefficient can be used to accurately locate rodent burrows that are actively breeding and require key management. This allows for the concentration of resources for efficient rodent trapping, avoiding the waste of scattered resources and ensuring that limited human and material resources are invested in the areas that need them most, thereby improving the overall rodent trapping efficiency. Furthermore, it enables adaptive adjustment of the rodent trapping frequency, improving the reliability and flexibility of the rodent trapping system. Attached Figure Description

[0040] Figure 1 This is a functional block diagram of a mouse-catching system based on the YOLO mouse hole detection model according to an embodiment of the present invention;

[0041] Figure 2 This is a schematic diagram of the activity trajectory of a suspected mouse hole according to an embodiment of the present invention;

[0042] Figure 3 This is a flowchart illustrating the logic of the mouse hole recognition module in this embodiment of the invention for recognizing characteristic activity trajectories.

[0043] Figure 4This is a flowchart illustrating the logic of the mouse hole identification module in this embodiment of the invention for filtering characteristic mouse holes.

[0044] In the diagram: 1 - suspected mouse hole; 2 - imprint; 3 - associated imprint vector. Detailed Implementation

[0045] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0046] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0047] It should be noted that in the description of this invention, the terms "upper," "lower," "inner," "outer," etc., which indicate the direction or positional relationship, are based on the direction or positional relationship shown in the drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0048] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0049] Please see Figure 1 The diagram shown is a functional block diagram of a mouse-catching system based on the YOLO mouse hole detection model according to an embodiment of the present invention. The mouse-catching system based on the YOLO mouse hole detection model of the present invention includes:

[0050] The feature acquisition module is used to mark suspected mouse holes based on the surface images of each sub-detection area on the grassland, and to perform imprint recognition within a first preset range set based on the suspected mouse holes;

[0051] Specifically, the YOLO algorithm can be used to mark suspected mouse holes based on surface images of each sub-detection area on the grassland.

[0052] Specifically, the YOLO mouse hole detection model can be used to mark suspected mouse holes in the surface images of each sub-detection region. The YOLO mouse hole detection model is trained using architectures such as YOLOv5 / v8 after data augmentation such as rotation and brightness adjustment based on a large dataset of mouse hole images. The model performs image preprocessing on the surface images of the obtained sub-detection regions. The preprocessed sub-detection region images are then sequentially input into the trained YOLO mouse hole detection model to extract texture and shape features from the images and fuse multi-scale features to provide suspected mouse hole marking results for the feature acquisition module.

[0053] Specifically, imprint recognition within a first preset range of suspected mouse holes can be performed using an image recognition model based on a convolutional neural network, such as YOLO. This model is trained by collecting a large dataset of images with imprint annotations and performing imprint recognition within the first preset range. The output imprint coordinates are then correlated with the mouse hole location to provide the imprint analysis module with raw data for filtering related imprint groups.

[0054] Specifically, the area of ​​the sub-detection region is the product of the grassland area to be detected and the area division factor, dividing the grassland into several grid-like sub-detection regions. The area division factor can be set by those skilled in the art according to the accuracy requirements of mouse hole detection. The higher the accuracy requirement, the smaller the area division factor. The value range of the area division factor can be [0.01, 0.03]. Preferably, the area division factor can be 0.02.

[0055] Specifically, the first preset range can be set by those skilled in the art based on the activity range of rodents on the grassland to be detected near their burrows. The first preset range is a circular area with the center point of the burrow as the center and the activity distance from the center point of the burrow as the radius. The activity distance is the product of the activity distance reference value and the activity distance factor. The activity distance reference value is the average activity distance of the rodents to be detected on the grassland in historical data. The activity distance factor can be set by those skilled in the art based on the accuracy requirements of burrow detection. The higher the accuracy requirements, the larger the activity distance factor is set. The value range of the activity distance factor can be [1.3, 1.45]. Preferably, the activity distance factor can be 1.4.

[0056] Specifically, the embodiments of the present invention do not specifically limit the structure of the feature acquisition module. Preferably, it can be a processor used in a computer to mark suspected mouse holes and perform imprint recognition, which will not be elaborated further.

[0057] The imprint analysis module, which is connected to the feature acquisition module, is used to filter several associated imprint groups of the suspected mouse hole based on the identified imprints, and to determine the activity trajectory of the suspected mouse hole based on the location distribution of the associated imprint groups.

[0058] Specifically, the embodiments of the present invention do not specifically limit the structure of the imprint analysis module. Preferably, it can be a microprocessor used to filter associated imprint groups and determine the activity trajectory of suspected mouse holes, which will not be elaborated further.

[0059] The mouse hole identification module is connected to the feature acquisition module and the imprint analysis module respectively. It is used to determine the activity tendency characterization quantity and screen out the feature imprints based on the imprint distribution of the activity trajectory. It identifies the feature activity trajectory based on the activity tendency characterization quantity and the feature imprints, and screens the feature mouse holes based on the feature activity trajectory and the soil characterization quantity of the suspected mouse hole.

[0060] Specifically, the embodiments of the present invention do not impose specific limitations on the structure of the mouse hole recognition module. Preferably, it can be constructed using logic components, such as field-programmable logic components, microprocessors, processors used in computers, etc., to determine activity tendency representation quantities, filter feature imprints, identify feature activity trajectories, and filter feature mouse holes, which will not be elaborated further.

[0061] The model determination output module is connected to the mouse hole recognition module and is used to determine the division result of the mouse hole system based on the location distribution of each characteristic mouse hole;

[0062] Specifically, the embodiments of the present invention do not specifically limit the structure of the model determination output module. Preferably, it can be a processor used in a computer to determine the division result of the mouse hole system based on the location distribution of each feature mouse hole. This will not be elaborated further.

[0063] The rodent control module is connected to the feature acquisition module and the model determination output module, respectively. It is used to determine the growth tendency coefficient based on the soil characterization of characteristic rodent holes in each rodent hole system, and to adjust the rodent trapping frequency based on the growth tendency coefficient, and push the rodent trapping frequency to the output end.

[0064] Specifically, the rat-catching frequency is the number of times rat trapping operations are carried out on rat burrows within a unit of time period. Rat-catching operations can include placing rodenticides or setting up physical rat traps.

[0065] Specifically, the embodiments of the present invention do not impose specific limitations on the structure of the output end. Preferably, it can be a display, a mobile terminal device, etc., which will not be elaborated further.

[0066] Specifically, the embodiments of the present invention do not specifically limit the structure of the model determination output module. Preferably, it can be a microprocessor used to determine the growth tendency coefficient and adjust the frequency of mouse trapping in the mouse hole system. This will not be elaborated further.

[0067] Specifically, the imprint analysis module is used to filter several related imprint groups, wherein,

[0068] The imprint analysis module filters two adjacent imprints into an associated imprint group based on the determination result that the interval between two adjacent imprints meets the conditions for an associated imprint group.

[0069] If the distance between two adjacent imprints does not meet the conditions for associated imprint group, the imprint analysis module will not filter the two adjacent imprints.

[0070] The condition for the associated imprint group is that the distance between two adjacent imprints does not exceed a preset first distance threshold.

[0071] Specifically, two adjacent imprints are spatially adjacent imprints. By calculating the interval between any imprint and the other imprints, the two imprints corresponding to the minimum interval distance are determined as two adjacent imprints. The interval distance between two adjacent imprints is the interval distance between the center points of the imprints.

[0072] Specifically, the preset first interval distance threshold is the product of the first interval distance threshold reference value and the first spacing factor. The first interval distance threshold reference value is the average interval distance between adjacent imprints of rodents to be detected on the grassland in historical data. The first spacing factor can be set by those skilled in the art according to the accuracy requirements of rodent hole detection. The higher the accuracy requirement, the larger the first spacing factor is set. The value range of the first spacing factor can be [1.25, 1.35]. Preferably, the first spacing factor can be 1.3.

[0073] Specifically, the imprint analysis module is used to determine the activity trajectory of the suspected mouse burrow, wherein...

[0074] The activity trajectory is determined based on several interconnected imprint vectors. The interconnected imprint vectors are constructed with the imprint with the largest distance between two adjacent imprints and the suspected mouse hole as the vector starting point and the imprint with the smallest distance between two adjacent imprints and the suspected mouse hole as the vector ending point.

[0075] For example, please refer to Figure 2 As shown, it is a schematic diagram of the activity trajectory of a suspected mouse hole in an embodiment of the present invention. The activity trajectory is a trajectory composed of several interconnected imprint vectors.

[0076] The imprint analysis module determines the trajectory composed of several interconnected imprint vectors as the active trajectory. The interconnected imprint vectors are constructed with the imprint with the largest distance from the suspected mouse hole among the adjacent imprints as the vector starting point and the imprint with the smallest distance from the suspected mouse hole among the adjacent imprints as the vector ending point.

[0077] Specifically, the distance between the imprint and the suspected mouse hole is the distance between the center point of the imprint and the center point of the mouse hole.

[0078] Specifically, this embodiment of the invention uses an imprint analysis module to determine the activity trajectory of suspected mouse holes based on the location distribution of associated imprint groups. It is understood that numerous depressions resembling mouse holes exist in the natural environment, such as wormholes and plant root cavities. The YOLO model's identification of mouse holes is easily affected by lighting and shadows, resulting in a high false positive rate. By analyzing the activity trajectory of rodents, false positives can be distinguished from real mouse holes. Simultaneously, the activity trajectory can provide data support for analyzing the activity level of mouse holes. When rodents move, they leave imprints on soft soil surfaces such as grasslands and farmland. The distance between adjacent imprints can characterize whether adjacent imprints belong to the same activity trajectory. By constructing associated imprint vectors based on adjacent imprints, the activity trajectory of rodents can be determined through the correlation between these vectors, thereby obtaining the rodent activity trajectory of suspected mouse holes identified by the YOLO model. This embodiment of the invention uses an imprint analysis module to determine the activity trajectory of suspected mouse holes based on the location distribution of associated imprint groups, thus realizing the identification of the activity trajectory of suspected mouse holes and improving the reliability and flexibility of the rodent-catching system.

[0079] Specifically, the mouse hole recognition module is used to determine activity tendency representation quantities and feature imprints, wherein,

[0080] The mouse hole identification module is used to calculate the distance between each imprint in the activity trajectory and the suspected mouse hole, determine the maximum distance as the activity tendency characteristic of the activity trajectory, and determine the imprint with the minimum distance as the feature imprint.

[0081] Please see Figure 3 The diagram shown is a logical flowchart of the mouse hole recognition module for recognizing characteristic activity trajectories according to an embodiment of the present invention. The mouse hole recognition module is used to recognize characteristic activity trajectories.

[0082] The mouse hole identification module is used to identify the activity trajectory as a characteristic activity trajectory based on the activity tendency representation quantity of the activity trajectory and the determination result that the characteristic imprint meets the condition of the characteristic activity trajectory.

[0083] If the activity tendency representation quantity and feature imprint of the activity trajectory do not meet the feature activity trajectory conditions, the mouse hole recognition module will not filter the activity trajectory;

[0084] The condition for the characteristic activity trajectory is that the activity tendency representation exceeds a preset activity tendency representation threshold, and the distance between the characteristic imprint and the suspected mouse hole does not exceed a preset second distance threshold.

[0085] Specifically, the preset activity tendency representation threshold is the product of the activity tendency representation reference value and the activity tendency factor. The activity tendency representation reference value is the average value of the activity tendency representation of the rodents to be detected on the grassland in historical data. The activity tendency factor can be set by those skilled in the art according to the accuracy requirements of rodent hole detection. The higher the accuracy requirement, the smaller the activity tendency factor. The value range of the activity tendency factor can be [1.2, 1.35]. Preferably, the activity tendency factor can be 1.3.

[0086] Specifically, the preset second interval distance threshold is the product of the second interval distance threshold reference value and the second spacing factor. The second interval distance threshold reference value is the average distance between the characteristic imprints of the rodents to be detected on the grassland and the suspected rodent burrows in historical data. The second spacing factor can be set by those skilled in the art according to the accuracy requirements of rodent burrow detection. The higher the accuracy requirement, the larger the second spacing factor is set. The value range of the second spacing factor can be [1.25, 1.35]. Preferably, the second spacing factor can be 1.3.

[0087] Specifically, the mouse burrow identification module is used to determine soil characterization quantities of suspected mouse burrows, wherein,

[0088] The mouse hole identification module is used to determine the gray values ​​of each monitoring point in the sub-detection area where the suspected mouse hole is located and the gray values ​​within a second preset range based on the surface image. The absolute value of the difference between the minimum gray value within the second preset range and the average gray value within the sub-detection area is determined as the soil characterization value of the suspected mouse hole. The second preset range is determined based on the suspected mouse hole.

[0089] For example, a specific embodiment for determining the soil characterization of a suspected mouse burrow is given here. The minimum gray value of the suspected mouse burrow within a second preset range is 35, and the average gray value of the sub-detection area is 80. Then the soil characterization is |35-80|=45.

[0090] Specifically, the monitoring points in each sub-detection area can be evenly distributed within the sub-detection area. The density of monitoring points in each sub-detection area can be set by those skilled in the art based on the accuracy requirements of mouse hole detection. The higher the accuracy requirements, the higher the density should be set. The density is 10 to 15 monitoring points in each sub-detection area.

[0091] Specifically, the second preset range can be set by those skilled in the art based on the distance between the soil accumulation position around the mouse burrow on the grassland to be detected and the center point of the mouse burrow. The second preset range is a circular area with the center point of the mouse burrow as the center and the distance between the soil accumulation position and the center point of the mouse burrow as the radius. The distance between the soil accumulation position and the center point of the mouse burrow is the product of the distance reference value and the soil spacing factor. The distance reference value is the average distance between the soil accumulation position around the mouse burrow on the grassland to be detected and the center point of the mouse burrow in historical data. The soil spacing factor can be set by those skilled in the art based on the accuracy requirements of mouse burrow detection. The higher the accuracy requirements, the larger the set soil spacing factor. The value range of the soil spacing factor can be [1.3, 1.5]. Preferably, the soil spacing factor can be 1.4.

[0092] Please see Figure 4 The diagram shown is a flowchart illustrating the logic of the mouse hole identification module in an embodiment of the present invention for filtering characteristic mouse holes. The mouse hole identification module is used to filter characteristic mouse holes.

[0093] The mouse hole identification module filters the suspected mouse holes as characteristic mouse holes based on the judgment results that the characteristic activity trajectory and soil characterization of the suspected mouse holes meet the characteristic mouse hole conditions;

[0094] If the characteristic activity trajectory and soil characterization of a suspected mouse burrow do not meet the characteristics of a mouse burrow, the mouse burrow identification module will not screen the suspected mouse burrow.

[0095] The characteristic mouse hole condition is that the number of characteristic activity trajectories exceeds a preset reference value, and the soil characterization quantity exceeds a preset soil characterization reference value.

[0096] Specifically, the preset quantity reference value is the product of the quantity threshold and the quantity factor. The quantity threshold is the average number of characteristic activity trajectories of mouse holes on the grassland to be detected in historical data. The quantity factor can be set by those skilled in the art according to the accuracy requirements of mouse hole detection. The higher the accuracy requirement, the larger the quantity factor is set. The value range of the quantity factor can be [1.23, 1.4]. Preferably, the quantity factor can be 1.3.

[0097] Specifically, the preset soil characterization reference value is the product of the historical soil characterization average value and the soil factor. The historical soil characterization average value is the average value of the soil characterization of mouse burrows on the grassland to be detected in the historical data. The soil factor can be set by those skilled in the art according to the accuracy requirements of mouse burrow detection. The higher the accuracy requirement, the larger the soil factor is set. The value range of the soil factor can be [1.2, 1.35]. Preferably, the soil factor can be 1.25.

[0098] Specifically, in this embodiment of the invention, a mouse burrow identification module filters characteristic mouse burrows based on feature activity trajectories and soil characterization of suspected mouse burrows. It is understood that by comprehensively considering feature activity trajectories and soil characterization, real and active mouse burrows can be more accurately distinguished from other similar structures. Relying solely on a single soil characterization or activity trajectory feature can lead to misjudgment. For example, some soil imprints formed by natural factors such as rainwater erosion may resemble mouse burrows in grayscale values, but combining activity trajectory features can eliminate them. Similarly, some accidentally formed imprints resembling rodent activity trajectories will not be identified as characteristic mouse burrows without corresponding soil characterization support. This comprehensive screening method effectively reduces the risk of misjudgment. This reduces the false positive rate and improves the accuracy of rat hole identification. The number of characteristic activity tracks can also characterize the activity level of a rat hole. The more characteristic activity tracks a rat hole has, the more frequently rats enter and exit the hole, meaning that the rat hole is in active use. For rodent control, these active rat holes are key targets for management. Screening out such characteristic rat holes allows for targeted rodent trapping and other control measures, improving control efficiency, reducing interference with non-target holes, better controlling rodent populations, and reducing the damage of rodents to the grassland ecosystem. Furthermore, it enables the determination of rat hole activity based on intermittent imprints around the identified rat holes, improving the reliability and flexibility of the rodent trapping system.

[0099] Specifically, it can be understood that rodents leave activity tracks when entering and exiting their burrows. The distribution of these tracks reflects the behavioral patterns of rodents. During foraging and exploration, rodents will start from the burrow entrance and gradually move away from the burrow at a certain distance. Characteristic activity tracks include tracks near the burrow entrance and tracks further away from the burrow. The more characteristic activity tracks a burrow has, the more frequently rodents use the burrow for various activities, and the more active the burrow is. The color of the soil around the burrow is closely related to the formation of the burrow. When rodents dig burrows, they excavate the soil inside and pile it around the burrow entrance. The soil around these suspected mouse holes, which have not undergone long-term weathering and vegetation growth, differs in color from the naturally formed soil in the surrounding area. It is usually darker and therefore has a relatively low gray value. In contrast, the soil in normal areas that have not been disturbed by mouse holes is relatively uniform and lighter in color due to long-term natural environmental influences, i.e., it has a higher gray value. By calculating the difference in gray value between the suspected mouse hole area and its surroundings, the soil characterization can be obtained to help determine whether the suspected mouse hole is a real mouse hole. In this way, the activity of mouse holes can be judged based on the discontinuous imprints around the identified mouse holes, thereby improving the reliability and flexibility of the mouse trapping system.

[0100] Specifically, the model determination output module is used to classify each feature mouse hole into different mouse hole systems, wherein,

[0101] The model determination output module calculates the distance between any feature mouse hole and other feature mouse holes, and classifies feature mouse holes whose distance does not exceed the preset third distance threshold into the same mouse hole system. Each feature mouse hole exists only in a unique mouse hole system.

[0102] For example, a specific implementation of dividing a mouse hole system is given here. The interval distance between characteristic mouse hole A and characteristic mouse hole B is 2m, the interval distance between characteristic mouse hole A and characteristic mouse hole C is 1.8m, the interval distance between characteristic mouse hole A and characteristic mouse hole D is 3.8m, the interval distance between characteristic mouse hole A and characteristic mouse hole E is 2.5m, and the interval distance between characteristic mouse hole A and characteristic mouse hole F is 4m. The third interval distance threshold is set to 3.5m. Therefore, characteristic mouse hole A, characteristic mouse hole B, characteristic mouse hole C, and characteristic mouse hole E are divided into the same mouse hole system.

[0103] Specifically, the preset third interval distance threshold is the product of the third interval distance threshold reference value and the third spacing factor. The third interval distance threshold reference value is the average of the interval distances between rodent burrows in the same grassland burrow system in historical data. The third spacing factor can be set by those skilled in the art according to the accuracy requirements of rodent burrow detection. The higher the accuracy requirement, the smaller the third spacing factor is set. The value range of the third spacing factor can be [1.15, 1.25]. Preferably, the third spacing factor can be 1.2.

[0104] Specifically, in this embodiment of the invention, the model-based output module outputs the classification results of mouse burrow systems based on the location distribution of each characteristic mouse burrow. It can be understood that dividing characteristic mouse burrows into different mouse burrow systems allows for a more accurate assessment of the mouse population size of each system. Mouse burrows within the same system are typically used by one mouse population. Clearly identifying the system to which each mouse burrow belongs allows for the rational allocation of mouse-catching resources based on the activity level and population size of different systems, improving efficiency, avoiding resource waste, and enabling precise targeting of different mouse-infested areas. Analyzing mouse burrows within the same system allows for the estimation of the approximate size of the mouse population within that system, enabling timely adjustment of the mouse-catching frequency to ensure effectiveness and control the development of rodent infestations. This also protects the original ecology of the grassland ecosystem, balancing and avoiding the negative impacts of excessive mouse-catching on other organisms and the environment. Furthermore, it enables joint analysis of active mouse burrows, improving the reliability and flexibility of the mouse-catching system.

[0105] Specifically, it is understandable that rodents typically exhibit family-based or community-based living habits. A burrow system often corresponds to a relatively stable rodent population. They will dig multiple interconnected or adjacent burrows within a certain area, forming a burrow system as their living, breeding, and refuge from predators. At the same time, rodents have a limited range of activity, generally foraging and breeding around the area where their burrow system is located. They rarely leave their burrow system for long distances to open up new activity areas. Dividing rodents into different burrow systems based on the distance between characteristic burrows conforms to the actual living behavior patterns of rodents and can accurately reflect the distribution of rodent populations on grasslands. In this way, it enables joint analysis of active rodent burrows, improving the reliability and flexibility of the rodent-catching system.

[0106] Specifically, the mouse-catching control module is used to determine the growth tendency coefficient, wherein,

[0107] The rodent control module selects the characteristic rodent burrows as newly formed characteristic rodent burrows based on the soil characterization of the characteristic rodent burrows in the rodent burrow system, which meets the conditions for newly formed characteristic rodent burrows.

[0108] If the soil characterization of characteristic mouse burrows in the mouse burrow system does not meet the conditions for newly formed characteristic mouse burrows, then the mouse-catching control module will not screen the characteristic mouse burrows.

[0109] The condition for the newly formed characteristic mouse burrows is that the soil characterization amount exceeds the preset soil characterization threshold.

[0110] The growth tendency coefficient is the ratio of the number of newly formed characteristic mouse holes to the number of characteristic mouse holes in the mouse hole system.

[0111] Specifically, the preset soil characterization threshold is the product of the historical soil characterization average and the newly formed soil factor. The historical soil characterization average is the average soil characterization amount of the mouse burrows on the grassland to be detected in the historical data. The newly formed soil factor can be set by those skilled in the art according to the accuracy requirements of mouse burrow detection. The higher the accuracy requirement, the larger the newly formed soil factor is set. The value range of the newly formed soil factor can be [1.4, 1.5]. Preferably, the newly formed soil factor can be 1.45.

[0112] Specifically, the mouse-catching control module is used to adjust the frequency of mouse catching in the mouse burrow system, and the mouse-catching frequency is positively correlated with the growth tendency coefficient.

[0113] Specifically, the adjustment amount of the mouse-catching frequency is the initial mouse-catching frequency × (growth tendency coefficient × growth tendency factor). The initial mouse-catching frequency is the average value of the historical mouse-catching frequency. The growth tendency factor can be set by those skilled in the art based on the average value of historical data on the grassland to be tested. The value range of the growth tendency factor can be [0.1, 0.3] to avoid the mouse-catching frequency being adjusted too large or too small. Preferably, the growth tendency factor can be 0.2.

[0114] Specifically, this embodiment of the invention uses a rodent control module to adjust the rodent-catching frequency based on a growth tendency coefficient. This means that adjusting the rodent-catching frequency based on the growth tendency coefficient allows for real-time optimization of the rodent-catching strategy according to the reproductive dynamics of the rodent burrow system. When the growth tendency coefficient is high, it indicates a large number of newborn mice and active rodent reproduction; in this case, increasing the rodent-catching frequency can effectively control the expansion of the rodent population. If the growth tendency coefficient is low, the rodent-catching frequency can be appropriately reduced to avoid wasting resources, achieving precise and dynamic rodent-catching management, and greatly improving rodent-catching efficiency and effectiveness. The rodent-catching frequency is adjusted based on the reproductive signs of the rodent burrow system. This method can prevent excessive rodent trapping from damaging the grassland ecosystem, ensuring the balance of grassland biodiversity. At the same time, it can accurately control the rodent population, avoiding the damage to grassland vegetation, soil, and other ecological environments caused by rodent infestations. It achieves a balance between rodent control and ecological protection. Since rodent trapping resources are limited, the growth tendency coefficient can be used to accurately locate rodent burrows that are actively breeding and require key management. This allows for the concentration of resources for efficient rodent trapping, avoiding the waste of resources by spreading them out. It ensures that limited human and material resources are invested in the areas that need them most, improving the overall efficiency of rodent trapping. Furthermore, it enables adaptive adjustment of the rodent trapping frequency, improving the reliability and flexibility of the rodent trapping system.

[0115] Specifically, it is understandable that rodent reproduction leads to the expansion of rodent populations. As newborn rodents grow up, they require more space and resources, thus digging new burrows. When rodents dig new burrows, they excavate the soil inside and pile it around the burrow entrance, forming new mounds. New burrows can be identified through soil characterization. The higher the growth tendency coefficient, the better the rodent population reproduction within the burrow system, the faster the rodent population expands, and the higher the future risk of rodent infestation. The growth tendency coefficient is obtained by calculating the ratio of newly formed characteristic rodent burrows to total characteristic rodent burrows. This coefficient can characterize the reproductive dynamics of rodents within the burrow system, providing a scientific basis for adjusting rodent control strategies. By following the natural reproductive laws of rodent populations, precise control of rodent population numbers can be achieved, effectively curbing the development of rodent infestations. Furthermore, adaptive adjustments to the rodent trapping frequency can be made, improving the reliability and flexibility of the rodent trapping system.

[0116] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0117] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A mouse-catching system based on the YOLO mouse hole detection model, characterized in that, include: The feature acquisition module is used to mark suspected mouse holes based on the surface images of each sub-detection area on the grassland, and to perform imprint recognition within a first preset range set based on the suspected mouse holes; The imprint analysis module, which is connected to the feature acquisition module, is used to filter several associated imprint groups of the suspected mouse hole based on the identified imprints, and to determine the activity trajectory of the suspected mouse hole based on the location distribution of the associated imprint groups. The mouse hole identification module is connected to the feature acquisition module and the imprint analysis module respectively. It is used to determine the activity tendency characterization quantity and screen out the feature imprints based on the imprint distribution of the activity trajectory. It identifies the feature activity trajectory based on the activity tendency characterization quantity and the feature imprints, and screens the feature mouse holes based on the feature activity trajectory and the soil characterization quantity of the suspected mouse hole. The model determination output module is connected to the mouse hole recognition module and is used to determine the division result of the mouse hole system based on the location distribution of each characteristic mouse hole; The rodent control module is connected to the feature acquisition module and the model determination output module, respectively. It is used to determine the growth tendency coefficient based on the soil characterization of characteristic rodent holes in each rodent hole system, and to adjust the rodent trapping frequency based on the growth tendency coefficient, and push the rodent trapping frequency to the output end.

2. The mouse-catching system based on the YOLO mouse hole detection model according to claim 1, characterized in that, The imprint analysis module is used to filter several related imprint groups, wherein... The imprint analysis module filters two adjacent imprints into an associated imprint group based on the determination result that the interval between two adjacent imprints meets the conditions for an associated imprint group. The condition for the associated imprint group is that the distance between two adjacent imprints does not exceed a preset first distance threshold.

3. The mouse-catching system based on the YOLO mouse hole detection model according to claim 2, characterized in that, The imprint analysis module is used to determine the activity trajectory of the suspected mouse burrow, wherein... The activity trajectory is determined based on several interconnected imprint vectors. The interconnected imprint vectors are constructed with the imprint with the largest distance between two adjacent imprints and the suspected mouse hole as the vector starting point and the imprint with the smallest distance between two adjacent imprints and the suspected mouse hole as the vector ending point.

4. The mouse-catching system based on the YOLO mouse hole detection model according to claim 3, characterized in that, The mouse hole identification module is used to determine the activity tendency representation quantity and feature imprint, wherein... The mouse hole identification module is used to calculate the distance between each imprint in the activity trajectory and the suspected mouse hole, determine the maximum distance as the activity tendency characteristic of the activity trajectory, and determine the imprint with the minimum distance as the feature imprint.

5. The mouse-catching system based on the YOLO mouse hole detection model according to claim 4, characterized in that, The mouse hole recognition module is used to identify characteristic activity trajectories, wherein... The mouse hole identification module is used to identify the activity trajectory as a characteristic activity trajectory based on the activity tendency representation quantity of the activity trajectory and the determination result that the characteristic imprint meets the condition of the characteristic activity trajectory. The condition for the characteristic activity trajectory is that the activity tendency representation exceeds a preset activity tendency representation threshold, and the distance between the characteristic imprint and the suspected mouse hole does not exceed a preset second distance threshold.

6. The mouse-catching system based on the YOLO mouse hole detection model according to claim 5, characterized in that, The mouse burrow identification module is used to determine the soil characterization quantities of suspected mouse burrows, wherein, The mouse hole identification module is used to determine the gray values ​​of each monitoring point in the sub-detection area where the suspected mouse hole is located and the gray values ​​within a second preset range based on the surface image. The absolute value of the difference between the minimum gray value within the second preset range and the average gray value within the sub-detection area is determined as the soil characterization value of the suspected mouse hole. The second preset range is determined based on the suspected mouse hole.

7. The mouse-catching system based on the YOLO mouse hole detection model according to claim 6, characterized in that, The mouse hole identification module is used to filter characteristic mouse holes, wherein... The mouse hole identification module filters the suspected mouse holes as characteristic mouse holes based on the judgment results that the characteristic activity trajectory and soil characterization of the suspected mouse holes meet the characteristic mouse hole conditions; The characteristic mouse hole condition is that the number of characteristic activity trajectories exceeds a preset reference value, and the soil characterization quantity exceeds a preset soil characterization reference value.

8. The mouse-catching system based on the YOLO mouse hole detection model according to claim 7, characterized in that, The model determination output module is used to classify each characteristic mouse hole into different mouse hole systems, wherein... The model determination output module calculates the distance between any feature mouse hole and other feature mouse holes, and classifies feature mouse holes whose distance does not exceed the preset third distance threshold into the same mouse hole system. Each feature mouse hole exists only in a unique mouse hole system.

9. The mouse-catching system based on the YOLO mouse hole detection model according to claim 8, characterized in that, The rat-catching control module is used to determine the growth tendency coefficient, wherein... The rodent control module selects the characteristic rodent burrows as newly formed characteristic rodent burrows based on the soil characterization of the characteristic rodent burrows in the rodent burrow system, which meets the conditions for newly formed characteristic rodent burrows. The condition for the newly formed characteristic mouse hole is that the soil characterization amount exceeds the preset soil characterization threshold. The growth tendency coefficient is the ratio of the number of newly formed characteristic mouse holes to the number of characteristic mouse holes in the mouse hole system.

10. The mouse-catching system based on the YOLO mouse hole detection model according to claim 9, characterized in that, The rat-catching control module is used to adjust the rat-catching frequency of the rat hole system, and the rat-catching frequency is positively correlated with the growth tendency coefficient.

Images