Air-air cooperative grassland mouse hole identification method, device and equipment and medium

Abstract

The invention discloses an air-air cooperative grassland mouse hole recognition method, device and equipment and a medium, and the method comprises the steps: carrying out the mouse hole recognition based on a multispectral image, so as to determine the confidence coefficient of a to-be-recognized mouse hole; according to the confidence coefficients of the to-be-identified mouse holes, the mouse hole types of the to-be-identified mouse holes are determined, and the mouse hole types include suspected mouse holes and determined mouse holes; based on an inter-frame difference algorithm, determining geographic coordinates of the to-be-identified mouse hole; step S14, obtaining geographic coordinates of a to-be-identified mouse hole as a suspected mouse hole, and obtaining fluorescence data at the corresponding geographic coordinates; s15, according to the fluorescence data, determining whether the mouse hole type of the to-be-identified mouse hole as the suspected mouse hole is corrected into a determined mouse hole; and S16, taking the to-be-identified mouse holes of which the mouse hole types are the determined mouse holes as target mouse holes, and sending geographic coordinates of the target mouse holes to a target client. The invention belongs to the field of mouse hole identification, and can realize accurate identification of mouse holes.

Description

Technical Field

[0001] This invention relates to the field of mouse hole identification, and more particularly to a method, apparatus, equipment, and medium for identifying mouse holes in grasslands using a coordinated air-to-air approach. Background Technology

[0002] In grassland ecosystems, the frequent burrowing, root digging, and excessive foraging by rodents can trigger a series of ecological problems once their population exceeds the environmental carrying capacity. These problems manifest as a reduction in usable grassland area and a significant decrease in livestock carrying capacity, posing a serious threat to the health of grassland ecosystems and the sustainable development of animal husbandry. Traditional rodent monitoring methods primarily rely on manual surveys; however, in recent years, modern monitoring technologies such as satellite remote sensing and drone remote sensing have been gradually introduced and widely applied.

[0003] Manual surveys typically employ methods such as fixed-point observation and day / night trapping to estimate rodent populations. This involves setting up rodent traps within a specific area and counting the number of rodents captured in a single day. However, this method requires continuous, long-term monitoring to obtain reliable statistical results, and its drawbacks are obvious: it is time-consuming, labor-intensive, costly, has a limited survey scope, and the results lack representativeness. In contrast, satellite remote sensing technology has developed a relatively mature theoretical and technical system in the field of forestry pest and disease monitoring. However, when applied to rodent monitoring, the low resolution of satellite remote sensing images makes it difficult to accurately identify minute features such as rodent burrows. Furthermore, satellite remote sensing monitoring is costly and lacks real-time performance, all of which limit its effectiveness in rodent monitoring. Therefore, a combined air-to-air method for grassland rodent burrow identification is urgently needed to improve the accuracy and timeliness of rodent burrow identification. Summary of the Invention

[0004] This invention provides an air-to-air collaborative grassland rodent burrow identification method, device, equipment, and medium, which solves the technical problems of low accuracy and insufficient timeliness of rodent burrow identification in the prior art, and achieves the technical effect of improving the accuracy of rodent burrow identification and the real-time nature of monitoring.

[0005] In a first aspect, the present invention provides a method for identifying grassland rodent burrows through aerial-to-aerial cooperation, comprising: Based on the navigator aircraft, execute steps S11-S13, in which the navigator aircraft only cruises along a preset route: Step S11: Acquire a multispectral image of the target area and perform mouse hole identification based on the multispectral image to determine the confidence level of the mouse hole to be identified; Step S12: Determine the type of each mouse hole to be identified based on the confidence level of the mouse hole to be identified. The mouse hole types include suspected mouse holes and confirmed mouse holes. Step S13: Determine the geographic coordinates of the mouse hole to be identified based on the inter-frame difference algorithm; Based on the auxiliary patrol aircraft, execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse hole to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution or fluorescence area. Step S15: Based on the fluorescence data, determine whether the mouse hole type of the suspected mouse hole to be identified should be corrected to a confirmed mouse hole. Step S16: All mouse holes to be identified that are classified as "determined mouse holes" are designated as target mouse holes, and the geographical coordinates of the target mouse holes are sent to the target client.

[0006] Furthermore, based on fluorescence data, it is determined whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole, including: If the spectral characteristics of the image show bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then the suspected mouse hole will be corrected to a confirmed mouse hole.

[0007] Furthermore, based on fluorescence data, it is determined whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole, including: When the spectral characteristics of an image show bright blue or blue-white fluorescence under ultraviolet light at a preset wavelength, if the fluorescence intensity of the image is higher than the intensity threshold, the proportion of the fluorescence display area is higher than the area threshold, and the fluorescence is distributed in a ring or path shape, then the suspected mouse hole will be corrected to a confirmed mouse hole.

[0008] Furthermore, determining whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole based on fluorescence data also includes: If the spectral characteristics of the image show no bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then the altitude of the auxiliary patrol vehicle is determined based on its shooting height, shooting radius, and the preset burrow opening range for grassland rodents, including: ,in, To ascend to altitude, For shooting height, For the shooting radius, Pre-defined burrow entrance areas for grassland rodents; After the auxiliary patrol vehicle ascends to the desired height, it re-acquires the image and determines whether the number of mouse holes to be identified in the image is within the preset range. If the suspected mouse hole is within the preset range, it will be corrected to a confirmed mouse hole; otherwise, it will not be corrected.

[0009] Furthermore, based on the inter-frame difference algorithm, the geographical coordinates of the mouse hole to be identified are determined, including: determining the distance between the mouse hole to be identified and the navigator aircraft, including: ,in, The lead aircraft was positioned at the first shooting location and the mouse hole to be identified. distance, The lead aircraft was positioned at the second shooting location and the mouse hole to be identified. distance, Let be the radius of the mouse hole to be identified. The imaging angle of the camera on the navigator aircraft. For the pixels of the image, The size of the image taken at the first shooting location of the mouse hole to be identified. The size of the image taken at the second shooting location of the mouse hole to be identified; , ,in, for and The distance; Based on the distance between the mouse hole to be identified and the navigator aircraft, the pixel difference between the first and second shooting positions of the navigator aircraft, and the skew angle information, the geographical coordinates of the mouse hole to be identified are determined. Further, based on the confidence level of the mouse hole to be identified, the type of mouse hole is determined, including: If the confidence level of the mouse hole to be identified is between 0.3 and 0.6, then the mouse hole to be identified is a suspected mouse hole; if the confidence level of the mouse hole to be identified is greater than 0.6, then the mouse hole to be identified is a confirmed mouse hole.

[0010] Furthermore, determining whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole based on fluorescence data also includes: ,in, Rate the entrance to the cave. All are weights. The normalized fluorescence intensity is... The similarity between the fluorescence spectrum and the normalized standard spectrum of mouse urine is used. This represents the percentage of the normalized fluorescent area. This is a preset value. If the fluorescence is distributed in a ring or path pattern, then... Pick Otherwise take ,and Greater than If the score of the hole is greater than the preset score, the corresponding suspected mouse hole will be corrected to a confirmed mouse hole.

[0011] Secondly, the present invention provides an air-to-air collaborative grassland rodent burrow identification device, comprising: The navigator module includes using a navigator to execute steps S11-S13, wherein the navigator only flies along a preset route: Step S11, acquiring multispectral images of the target area and identifying mouse holes based on the multispectral images to determine the confidence level of the mouse holes to be identified; Step S12, determining the mouse hole type of each mouse hole to be identified based on the confidence level of the mouse holes to be identified, wherein the mouse hole type includes suspected mouse holes and confirmed mouse holes; Step S13, determining the geographical coordinates of the mouse holes to be identified based on the inter-frame difference algorithm; The auxiliary patrol module includes using the auxiliary patrol device to execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse holes to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution, or fluorescence area; Step S15: Determine whether the mouse hole type of the suspected mouse holes to be identified has been corrected to a confirmed mouse hole based on the fluorescence data; Step S16: Treat all mouse holes to be identified with the mouse hole type of confirmed mouse hole as target mouse holes, and send the geographic coordinates of the target mouse holes to the target client.

[0012] Thirdly, the present invention provides an electronic device, comprising: A processor; a memory for storing processor-executable instructions; wherein the processor is configured to perform an air-to-air cooperative grassland mouse hole recognition as provided in the first aspect.

[0013] Fourthly, the present invention provides a non-transitory computer-readable storage medium, wherein when the instructions in the non-transitory computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to perform an air-to-air cooperative grassland rat hole identification as provided in the first aspect.

[0014] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention achieves efficient and accurate rodent burrow identification through the collaborative operation of a lead aircraft and a support patrol aircraft. The lead aircraft rapidly surveys along a pre-set route, using multispectral imaging to initially screen suspected areas, completing a large-scale screening and geolocation. The support patrol aircraft then conducts close-range, detailed verification of suspected targets, using methods such as ultraviolet fluorescence detection to verify rodent burrow activity, fully leveraging the complementary advantages of "wide-area coverage" and "precise identification." This air-to-air collaborative mode improves detection efficiency and reduces the false positive rate of a single device, achieving intelligent, hierarchical, and precise identification of rodent burrows in grasslands.

[0015] This invention utilizes fluorescence data to determine whether a suspected mouse burrow has been corrected to a confirmed mouse burrow. Based on the unique bright blue fluorescence characteristic of mouse urine under ultraviolet light, it achieves non-contact, high-sensitivity detection of rodent activity traces. By analyzing fluorescence intensity, spectral characteristics, and distribution patterns, this invention can effectively distinguish between real mouse burrows and non-biological cavities, identifying active burrows that, while lacking obvious structural features, exhibit frequent excretion markings. This significantly improves the accuracy and reliability of identification, especially in detecting low-activity or backup burrows that are easily missed by traditional visual methods.

[0016] This invention utilizes the spatial clustering characteristics of rodent burrow systems for indirect verification. When a suspected burrow does not show fluorescence, the imaging radius is expanded by increasing the altitude, allowing for the counting of similar burrows over a larger area. Since grassland rodent burrows typically cluster in groups of 3-4 within an area no larger than 5 meters in diameter, if multiple morphologically similar burrows are found in a cluster within this spatial range, even if individual burrows show no fluorescence, it can be inferred that they are part of the same rodent burrow system. Method 4 compensates for the missed detection problem caused by insufficient discharge and weak fluorescence from backup burrows, improves the ability to identify concealed or low-activity burrows, and enhances the robustness and overall detection coverage of the system. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating an air-to-air collaborative grassland mouse burrow identification method provided by the present invention; Figure 2 A schematic diagram illustrating the configuration of the detectors of the navigator provided by this invention; Figure 3 This is a schematic diagram of the architecture of the lightweight detection model for multispectral UAV images provided by the present invention; Figure 4 A schematic diagram of the structure of the high-efficiency cross-attention fusion module provided by the present invention; Figure 5 This is a schematic diagram of the structure of the small-scale information enhancement module provided by the present invention; Figure 6 This is a schematic diagram of the auxiliary patrol machine provided by the present invention. Detailed Implementation

[0019] This invention provides an air-to-air collaborative grassland mouse hole identification method, which solves the technical problem of low accuracy in mouse hole identification in the prior art.

[0020] The technical solution of this invention is to solve the above-mentioned technical problems, and the overall idea is as follows: A method for identifying mouse holes in grasslands using air-to-air cooperative methods includes: steps S11-S13 performed by a navigator aircraft, wherein the navigator aircraft flies along a preset route: Step S11, acquiring multispectral images of the target area and identifying mouse holes based on the multispectral images to determine the confidence level of the mouse holes to be identified; Step S12, determining the mouse hole type of each mouse hole to be identified based on the confidence level, wherein the mouse hole type includes suspected mouse holes and confirmed mouse holes; Step S13, determining the geographical coordinates of the mouse holes to be identified based on an inter-frame difference algorithm. Based on the auxiliary patrol device, execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse holes to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution, or fluorescence area; Step S15: Determine whether the mouse hole type of the suspected mouse holes to be identified has been corrected to a confirmed mouse hole based on the fluorescence data; Step S16: Take all the mouse holes to be identified with the mouse hole type of confirmed mouse hole as target mouse holes, and send the geographic coordinates of the target mouse holes to the target client.

[0021] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0022] First, it should be clarified that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0023] This invention provides, for example Figure 1 The method for identifying grassland mouse holes through air-to-air cooperation, as shown, includes steps S11-S16: Based on the navigator aircraft, execute steps S11-S13, in which the navigator aircraft only cruises along a preset route: Step S11: Acquire a multispectral image of the target area and identify mouse holes based on the multispectral image to determine the confidence level of the mouse holes to be identified.

[0024] The lead drone, typically a DJI Matrice 600 Pro hexacopter, is responsible for the initial screening of mouse holes. It can fly along a pre-set route. The lead drone can be equipped with a multispectral camera and a detector. The detector has a built-in image recognition module that can analyze multispectral images in real time, identify mouse holes, and calculate confidence levels.

[0025] For determining the confidence level of the rodent holes to be identified, please refer to the paper "Study on rodent damage and degradation in desert grassland based on UAV hyperspectral remote sensing", which will not be elaborated in this invention.

[0026] Step S12: Determine the mouse hole type of each mouse hole to be identified based on the confidence level of the mouse hole to be identified. The mouse hole types include suspected mouse holes and confirmed mouse holes.

[0027] Specifically, if the confidence level of the mouse hole to be identified is between 0.3 and 0.6, then the mouse hole to be identified is considered a suspected mouse hole; if the confidence level of the mouse hole to be identified is greater than 0.6, then the mouse hole to be identified is considered a confirmed mouse hole.

[0028] The meaning of "suspected mouse hole" is that the mouse hole to be identified is likely a mouse hole, and the meaning of "identified mouse hole" is that the mouse hole to be identified is a mouse hole.

[0029] If the confidence level is below 0.3, then the location of the mouse hole to be identified is a mouse hole.

[0030] Step S13: Determine the geographical coordinates of the mouse hole to be identified based on the inter-frame difference algorithm.

[0031] The lead aircraft's detectors mark suspected mouse holes with a confidence level between 0.3 and 0.6 and calculate their specific locations (geographic coordinates). Subsequently, the geographic coordinates of the suspected mouse holes are transmitted to the auxiliary patrol aircraft via the LoRa communication network. Upon receiving the location information, the auxiliary patrol aircraft proceeds to the corresponding location to further confirm the suspected mouse holes.

[0032] Determine the geographic coordinates of the mouse burrow to be identified, including: Determine the distance between the mouse hole to be identified and the navigator aircraft, including:

[0033] in, The lead aircraft was positioned at the first shooting location and the mouse hole to be identified. distance, The lead aircraft was positioned at the second shooting location and the mouse hole to be identified. distance, Let be the radius of the mouse hole to be identified. The imaging angle of the camera on the navigator aircraft. For the pixels of the image, The size of the image taken at the first shooting location of the mouse hole to be identified. The size of the image taken at the second shooting location of the mouse hole to be identified;

[0034]

[0035] in, for and The distance; Based on the geographic coordinates of the navigator (i.e., the second shooting location), the geographic coordinates of the mouse hole to be identified are determined according to the distance between the mouse hole to be identified and the navigator, and the pixel difference between the first and second shooting locations of the navigator.

[0036] Understandably, whether it's necessary to determine D1 or D2, as long as the distance between the two and the shooting position at another location are determined, the other location can be determined.

[0037] It is understandable that the method of determining the geographical coordinates of suspected mouse holes is not limited in this invention, as long as the geographical coordinates of suspected mouse holes can be reproduced to the auxiliary patrol machine.

[0038] Based on the auxiliary patrol aircraft, execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse hole to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution or fluorescence area.

[0039] The lead aircraft can transmit its geographic coordinates to the auxiliary patrol aircraft via a radio ad hoc network.

[0040] After receiving the geographic coordinates, the auxiliary patrol aircraft proceeds to the corresponding location to further confirm the suspected mouse burrow. The auxiliary patrol aircraft flies above the suspected mouse burrow to acquire data. The auxiliary patrol aircraft can be equipped with an ultraviolet-visible fluorescence imaging camera to acquire fluorescence intensity, spectral characteristics, spatial distribution, or fluorescence area. The flight altitude of the auxiliary patrol aircraft is usually determined by being able to take a comprehensive and clear picture of the suspected mouse burrow.

[0041] Because rats often urinate indiscriminately while active, and urine contains fluorescent proteins and phosphorus compounds, these substances emit bright blue or bluish-white fluorescence when irradiated with ultraviolet light at a predetermined wavelength (365nm for a certain prairie rat species). It is understandable that the urine fluorescence color varies slightly among different rodent species. This invention is designed for a rodent species found in a grassland, and the preset wavelength can be set to 365nm.

[0042] Step S15: Based on the fluorescence data, determine whether the mouse hole type of the suspected mouse hole to be identified should be corrected to a confirmed mouse hole.

[0043] Method 1 If the spectral characteristics of the image show bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then the suspected mouse hole will be corrected to a confirmed mouse hole.

[0044] This invention improves the accuracy of rodent burrow identification by detecting the presence of bright blue or bluish-white fluorescence at the burrow entrance to determine whether a suspected burrow is indeed a rodent burrow.

[0045] To further improve the accuracy of mouse hole identification, the present invention also provides the following three methods.

[0046] Method 2 Determining whether to correct the type of a suspected mouse hole to a confirmed mouse hole based on fluorescence data also includes: When the spectral characteristics of an image show bright blue or blue-white fluorescence under ultraviolet light at a preset wavelength, if the fluorescence intensity of the image is higher than the intensity threshold, the proportion of the fluorescence display area is higher than the area threshold, and the fluorescence is distributed in a ring or path shape, then the suspected mouse hole will be corrected to a confirmed mouse hole.

[0047] Method 1 relies solely on spectral characteristics to determine the presence of fluorescence, but it is susceptible to environmental interference (such as plant sap, mineral reflection, or artificial contaminants), leading to false alarms. Method 2, however, employs a multi-parameter fusion approach, comprehensively considering fluorescence intensity, area proportion, and spatial distribution patterns, significantly improving the reliability of the determination. High-intensity and large-area fluorescence indicates frequent excretion, consistent with the biological characteristics of active mouse burrows; ring-shaped or path-like distributions correspond to the behavior of mice marking their territory around burrow entrances, demonstrating high specificity. This method effectively eliminates accidental or non-biological fluorescence interference, reducing the false alarm rate.

[0048] Method 3 Determining whether to correct the type of a suspected mouse hole to a confirmed mouse hole based on fluorescence data also includes:

[0049] in, Rate the entrance to the cave. All are weights. The normalized fluorescence intensity is... The similarity between the fluorescence spectrum and the normalized standard spectrum of mouse urine is used. This represents the percentage of the normalized fluorescent area. This is a preset value. If the fluorescence is distributed in a ring or path pattern, then... Pick Otherwise take ,and Greater than ; If the score of the hole is greater than the preset score, the corresponding suspected mouse hole will be corrected to a confirmed mouse hole.

[0050]

Method 3

[0051] Method 4 Determining whether to correct the type of a suspected mouse hole to a confirmed mouse hole based on fluorescence data also includes: If the spectral characteristics of the image show no bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then the altitude of the auxiliary patrol vehicle is determined based on its shooting height, shooting radius, and the preset burrow opening range for grassland rodents, including:

[0052] in, To ascend to altitude, For shooting height, For the shooting radius, Pre-defined burrow entrance areas for grassland rodents; After the auxiliary patrol vehicle ascends to the desired height, it re-acquires images and determines whether the number of mouse holes to be identified in the image is within a preset range (including the original mouse holes to be identified). The preset range can be 3-4 or determined according to the actual situation of the grassland or mouse species. If the suspected mouse hole is within the preset range, it will be corrected to a confirmed mouse hole; otherwise, it will not be corrected.

[0053] It is understandable that controlling the auxiliary patrol vehicle to rise is to obtain information about the surrounding area of ​​the suspected mouse hole while keeping the viewing angle unchanged. If the viewing angle is changed to a wide-angle lens, the surrounding information can also be obtained. This invention does not impose any limitations on this.

[0054] In grasslands, rat burrow systems typically include 3-4 exits, forming a hierarchical security network. Although there are 3-4 entrances, rats will prioritize using 1-2 of the safest and most convenient entrances as primary passages, while using the others as backup or emergency exits. In other words, the remaining 1-2 exits will contain urine or have a lower urine volume compared to the other frequently used entrances. Within a circular or elliptical area of ​​approximately 2-5 meters around the 3-4 entrances, the maximum distance between the entrances generally does not exceed 5 meters.

[0055] Therefore, to further improve the accuracy of burrow identification, this invention utilizes the spatial clustering characteristics of rodent burrow systems for indirect verification. When no fluorescence is detected in a suspected burrow, the imaging radius is expanded by increasing the shooting height, allowing for the counting of similar burrows over a larger area. Since grassland rodent burrows are typically grouped in groups of 3-4 and distributed within an area with a diameter of no more than 5 meters, if multiple burrows with similar morphologies are found clustered within this spatial range, even if individual burrows show no fluorescence, it can be inferred that they are part of the same rodent burrow system. Method 4 compensates for the missed detection problem caused by insufficient discharge and weak fluorescence from backup burrows, improves the ability to identify concealed or low-activity burrows, and enhances the robustness and overall detection coverage of the system.

[0056] Step S16: All mouse holes to be identified that are classified as "determined mouse holes" are designated as target mouse holes, and the geographical coordinates of the target mouse holes are sent to the target client.

[0057] The target client can be configured with multiple ports, which can be set on mobile phones, computers, iPads, or navigation devices, and the geographical coordinates of the target mouse hole can be sent to the target client for relevant personnel to refer to.

[0058] In addition, after the auxiliary patrol aircraft has completed its inspection of all suspected rat holes, it can mark all suspected rat holes that have not been corrected or identified as rat holes as non-rat holes.

[0059] Once the auxiliary patrol aircraft has identified all target mouse holes, it can send the geographic coordinates of the target mouse holes to the navigator aircraft via wireless network, and the navigator aircraft will then relay the data to the target client.

[0060] In summary, this invention achieves efficient and accurate rodent burrow identification through the collaborative operation of a lead aircraft and a patrol aircraft. The lead aircraft rapidly surveys along a pre-set route, using multispectral imaging to initially screen suspected areas, completing a large-scale screening and geolocation. The patrol aircraft then conducts close-range, detailed verification of suspected targets, using methods such as ultraviolet fluorescence detection to verify rodent burrow activity, fully leveraging the complementary advantages of "wide coverage" and "precise identification." This air-to-air collaborative mode improves detection efficiency and reduces the false positive rate of a single device, achieving intelligent, hierarchical, and precise identification of rodent burrows in grasslands.

[0061] This invention utilizes fluorescence data to determine whether a suspected mouse burrow has been corrected to a confirmed mouse burrow. Based on the unique bright blue fluorescence characteristic of mouse urine under ultraviolet light, it achieves non-contact, high-sensitivity detection of rodent activity traces. By analyzing fluorescence intensity, spectral characteristics, and distribution patterns, this invention can effectively distinguish between real mouse burrows and non-biological cavities, identifying active burrows that, while lacking obvious structural features, exhibit frequent excretion markings. This significantly improves the accuracy and reliability of identification, especially in detecting low-activity or backup burrows that are easily missed by traditional visual methods.

[0062] In addition, the inventors disclosed the configuration of the navigator's detectors, such as... Figure 2 ,include: The detector comprises an STM32 controller, an NVIDIA Orin Nano module, a radar altimeter, a BeiDou module, an attitude sensor module, and a LoRa module. The NVIDIA Orin Nano module connects to the GaiaSky-mini multispectral camera via USB and runs a lightweight multispectral UAV image detection model based on a hybrid architecture of YOLO v8 and Transformer. This model stores calculated mousehole data and outputs pixel change and angular information for suspected mousehole locations, subsequently transmitting the information to the STM32 controller.

[0063] The radar altimeter is model NRA24, with a detection range of 200 meters, a ranging accuracy of 0.02 meters, and an operating frequency of 1Hz. It connects to the STM32 controller via a UART interface to output the UAV's relative altitude data. The BeiDou module is model M8N from ublox, operating at 1Hz. It outputs BDGGA protocol data to the STM32 controller via a UART interface, providing accurate latitude and longitude information for the UAV. The attitude sensor module is model BNO085 from Bosch. This module integrates a high-performance accelerometer, magnetometer, and gyroscope, mounted on the multispectral analyzer's camera, and connects to the STM32 controller via a UART interface to output attitude data such as pitch angle.

[0064] The LORA module uses the E32-900T30S model from EBITA, with a communication range of up to 8 kilometers. It employs a point-to-point transparent transmission mode and communicates with the STM32 controller via a UART interface. The STM32 controller uses the STMicroelectronics STM32H743VIT6 chip, responsible for reading pixel change and deflection information output from the Orin Nano module. It combines this information with data such as the relative altitude to the ground measured by the radar altimeter and the pitch angle measured by the attitude sensor. Using a frame difference algorithm, it calculates the relative distance to suspected mouse holes and, based on the deflection information, calculates the latitude and longitude of the suspected mouse holes. Finally, the information is transmitted to the auxiliary patrol aircraft via the LORA module.

[0065] like Figure 3 As shown, the inventors disclosed the architecture of a lightweight detection model for multispectral UAV images, including: YOLO-FT uses the YOLOv8 model as its baseline framework, comprising a backbone network, a neck network, and a detection head. The backbone network of YOLO-FT uses the C2F module from YOLOv8 as its basic unit. This module not only possesses excellent feature representation capabilities but also facilitates the transfer of gradient information during backpropagation. By combining the C2F module with downsampling convolutional layers, multiple parallel feature extraction networks are constructed, each responsible for extracting multi-scale features from multispectral images.

[0066] like Figure 4 As shown, the inventors disclose the structure of the Efficient Cross Attention Fusion Module (ECAFM) in the architecture, including: By modeling the long-range dependencies of different spectral feature maps on the fused feature map, we can improve the extraction and fusion of information between different spectral features. First, we adaptively integrate the information carried by different spectral feature maps through channel-dimensional concatenation and a 3×3 convolutional module (CBS) to obtain a simple fused feature map. Secondly, downsampling is performed on these three feature maps to obtain... , , To improve computational efficiency, downsampling operation is employed. It consists of max pooling and average pooling:

[0067] The cross-attention mechanism is then calculated. Three 1×1 convolutions are used to linearly project the three downsampled feature maps to obtain the query matrix. and Then, the key-value matrices K and V are merged. Subsequently, , K and V are converted into sequence data, and calculations are performed respectively. and The attention matrix is ​​obtained by scaling the dot product attention relative to KV. and Using 3×3 convolution from and Extract local information and compare it with the information obtained from the previous step. and Perform element-wise addition. Integrate through channel-level splicing operations. and And perform linear projection using 1×1 convolution. Then... Adding the projected feature map to the feature map yields the deeply fused feature map. The feedforward network is designed using an MLP module consisting of two 1×1 convolutional modules (CBS) and residual connections to achieve channel-level information exchange. The channel dilation rate between the two convolutional layers is set to 2. Finally, the two feature maps are added together to obtain the final fused feature map for this stage. .

[0068] The inventors also disclosed the structure of the small-scale information enhancement module, such as Figure 5 As shown, it includes: The small-scale information enhancement module selects feature maps of three scales generated by the path aggregation network. , , As input, and Downsampling operations with feature scaling factors of 4 and 2 are applied respectively, and the data is processed by a 1×1 convolutional module CBS. Small-scale, medium-scale, and large-scale feature maps are obtained respectively. Subsequently, these three feature maps are integrated through channel-level concatenation and a 1×1 convolutional module (CBS) to obtain a simple fused feature map (DP). The next step is to calculate the self-attention mechanism of DP, computing the query matrix Q, key matrix K, value matrix V, and attention matrix Map to effectively fuse detailed and semantic information. Map is pre-sampled, and 3×3 convolution is used to extract... The local information is combined with the data through element-wise addition to achieve multi-scale information fusion. Subsequently, a 1×1 convolutional layer is used to linearly project the map, and residual connections are introduced. The same MLP structure as in ECAFM is used to realize the interaction of information between channels, obtaining the output feature map of SSIEM. to replace the original As input to the detection head, it enhances the ability to represent small-scale branch features.

[0069] like Figure 6 As shown, the inventors also disclosed the structure of the auxiliary patrol machine, including: The navigator and auxiliary patrol aircraft consist of a Pixhawk 6X controller, an NVIDIA Orin Nano module, a BeiDou module, and a LoRa module. The Pixhawk 6X controller is equipped with an STM32H753 main processor and an STM32F103 coprocessor, and has three built-in industrial-grade IMUs. It can also connect to an external IMU via an I2C interface to enhance altitude measurement accuracy and monitor sensor malfunctions. Furthermore, the controller supports multiple interfaces, including UART, CAN FD, I2C, SPI, and PWM, and can run the PX4 open-source flight control system, efficiently handling complex navigation, obstacle avoidance, and control algorithms. The ultraviolet camera is a MER3-810-36G3M-P-UV F02 model, with a wide spectral acquisition capability from 10nm to 400nm. The NVIDIA Orin Nano module connects to the ultraviolet camera via Ethernet. The BeiDou module is a ublox M8N model, operating at 1Hz, and connects to the GPS1 interface of the Pixhawk 6X controller to determine the UAV's precise latitude and longitude. The attitude sensor module utilizes the ICM-45686X3 accelerometer / gyroscope and BMM150 magnetometer built into the Pixhawk 6X controller to adjust the drone's attitude in real time, ensuring stability during filming. The LORA module is the E32-900T30S model from Ebix, with a communication range of up to 8 kilometers, employing a point-to-point pass-through mode and connecting to the UART4 interface of the Pixhawk 6X controller. The Pixhawk 6X controller reads the mousehole information output by the Orin Nano module and sends the result of whether a suspected mousehole is a confirmed mousehole to the client (which can be the navigator drone) via the LORA module. Furthermore, the Pixhawk 6X controller embeds the PX4 flight control system to control the drone's flight path, ensuring it flies according to the predetermined route and mission requirements.

[0070] Based on the same inventive concept, this invention provides an air-to-air collaborative grassland rodent burrow identification device, comprising: The navigator module includes using a navigator to execute steps S11-S13, wherein the navigator only flies along a preset route: Step S11, acquiring multispectral images of the target area and identifying mouse holes based on the multispectral images to determine the confidence level of the mouse holes to be identified; Step S12, determining the mouse hole type of each mouse hole to be identified based on the confidence level of the mouse holes to be identified, wherein the mouse hole type includes suspected mouse holes and confirmed mouse holes; Step S13, determining the geographical coordinates of the mouse holes to be identified based on the inter-frame difference algorithm; The auxiliary patrol module includes using the auxiliary patrol device to execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse holes to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution, or fluorescence area; Step S15: Determine whether the mouse hole type of the suspected mouse holes to be identified has been corrected to a confirmed mouse hole based on the fluorescence data; Step S16: Treat all mouse holes to be identified with the mouse hole type of confirmed mouse hole as target mouse holes, and send the geographic coordinates of the target mouse holes to the target client.

[0071] Based on the same inventive concept, the present invention also provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to execute a method for identifying grassland rat holes in a coordinated air-to-air manner, as described above.

[0072] Based on the same inventive concept, the present invention also provides a non-transitory computer-readable storage medium, which, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to execute the aforementioned air-to-air cooperative grassland mouse hole identification method.

[0073] Since the electronic device described in this embodiment is an electronic device used to implement the information processing method in the embodiments of the present invention, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the information processing method described in the embodiments of the present invention. Therefore, how the electronic device implements the method in the embodiments of the present invention will not be described in detail here. Any electronic device used by those skilled in the art to implement the information processing method in the embodiments of the present invention falls within the scope of protection of the present invention.

[0074] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0075] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0076] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0077] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0078] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0079] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for identifying grassland rodent burrows through aerial-to-aerial collaboration, characterized in that, include: Based on the navigator aircraft, steps S11-S13 are executed, wherein the navigator aircraft only patrols along a preset route: Step S11, acquires multispectral images of the target area, and identifies mouse holes based on the multispectral images to determine the confidence level of the mouse holes to be identified; Step S12, determines the mouse hole type of each mouse hole to be identified based on the confidence level of the mouse holes to be identified, wherein the mouse hole type includes suspected mouse holes and confirmed mouse holes; Step S13, determines the geographic coordinates of the mouse holes to be identified based on the inter-frame difference algorithm; Based on the auxiliary patrol aircraft, steps S14-S16 are executed: Step S14, acquires the geographic coordinates of the mouse holes to be identified as suspected mouse holes, and acquires fluorescence data at the corresponding geographic coordinates, the fluorescence data including fluorescence intensity, spectral characteristics, spatial distribution or fluorescence area; Step S15, determines whether the mouse hole type of the suspected mouse holes to be identified has been corrected to confirmed mouse holes based on the fluorescence data; Step S16, all mouse holes to be identified with the mouse hole type of confirmed mouse holes are designated as target mouse holes, and the geographic coordinates of the target mouse holes are sent to the target client.

2. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Based on fluorescence data, it is determined whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole, including: If the spectral characteristics of the image show bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then the suspected mouse hole will be corrected to a confirmed mouse hole.

3. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Based on fluorescence data, it is determined whether the type of a suspected mouse hole to be identified should be corrected to a confirmed mouse hole, including: When the spectral characteristics of an image show bright blue or blue-white fluorescence under ultraviolet light at a preset wavelength, if the fluorescence intensity of the image is higher than the intensity threshold, the proportion of the fluorescence display area is higher than the area threshold, and the fluorescence is distributed in a ring or path shape, then the suspected mouse hole will be corrected to a confirmed mouse hole.

4. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Determining whether to correct a suspected rodent burrow to a confirmed rodent burrow based on fluorescence data also includes: if the spectral characteristics of the image show no bright blue or bluish-white fluorescence under ultraviolet light at a preset wavelength, then determining the altitude of the auxiliary patrol vehicle based on its shooting height, shooting radius, and the preset burrow entrance range for grassland rodents, including: in, To ascend to altitude, For shooting height, For the shooting radius, The range of burrow entrances for grassland rodents is preset; after the auxiliary patrol vehicle ascends to the specified height, it re-acquires images and determines whether the number of burrows to be identified in the image is within the preset range; if it is within the preset range, the suspected burrows are corrected to confirmed burrows; if it is not within the preset range, no correction is made.

5. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Based on the inter-frame difference algorithm, the geographical coordinates of the mouse hole to be identified are determined, including: determining the distance between the mouse hole to be identified and the navigator aircraft, including: ,in, The lead aircraft was positioned at the first shooting location and the mouse hole to be identified. distance, The lead aircraft was positioned at the second shooting location and the mouse hole to be identified. distance, Let the radius be the mouse hole to be identified. The imaging angle of the camera on the navigator aircraft. For the pixels of the image, The size of the image taken at the first shooting location of the mouse hole to be identified. The size of the image taken at the second shooting location of the mouse hole to be identified; in, for and The distance; The geographical coordinates of the mouse hole to be identified are determined based on the distance between the mouse hole to be identified and the navigator, the pixel difference between the first and second shooting positions of the navigator, and the skew angle information.

6. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Based on the confidence level of the mouse hole to be identified, the mouse hole type of each mouse hole to be identified is determined, including: if the confidence level of the mouse hole to be identified is between 0.3 and 0.6, then the mouse hole to be identified is a suspected mouse hole; if the confidence level of the mouse hole to be identified is greater than 0.6, then the mouse hole to be identified is a confirmed mouse hole.

7. The method for identifying grassland rodent burrows through aerial collaboration as described in claim 1, characterized in that, Determining whether to correct the type of a suspected mouse hole to a confirmed mouse hole based on fluorescence data also includes: in, Rate the entrance to the cave. All are weights. The normalized fluorescence intensity is... The similarity between the fluorescence spectrum and the normalized standard spectrum of mouse urine is used. This represents the percentage of the normalized fluorescent area. This is a preset value. If the fluorescence is distributed in a ring or path pattern, then... Pick Otherwise take ,and Greater than If the score of the hole is greater than the preset score, the corresponding suspected mouse hole will be corrected to a confirmed mouse hole.

8. An air-to-air collaborative grassland rodent burrow identification device, characterized in that, include: The navigator module includes steps S11-S13, whereby the navigator flies along a preset route: Step S11, acquiring multispectral images of the target area and identifying mouse holes based on the multispectral images to determine the confidence level of the mouse holes to be identified; Step S12, determining the mouse hole type of each mouse hole to be identified based on the confidence level, where the mouse hole type includes suspected mouse holes and confirmed mouse holes; Step S13, determining the geographical coordinates of the mouse holes to be identified based on an inter-frame difference algorithm; the auxiliary patrol module includes... Using the auxiliary patrol device, execute steps S14-S16: Step S14: Obtain the geographic coordinates of the suspected mouse holes to be identified, and obtain fluorescence data at the corresponding geographic coordinates. The fluorescence data includes fluorescence intensity, spectral characteristics, spatial distribution, or fluorescence area; Step S15: Determine whether the mouse hole type of the suspected mouse holes to be identified has been corrected to a confirmed mouse hole based on the fluorescence data; Step S16: Take all the mouse holes to be identified with the mouse hole type of confirmed mouse hole as target mouse holes, and send the geographic coordinates of the target mouse holes to the target client.

9. An electronic device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor is configured to perform an air-to-air cooperative grassland mouse burrow identification as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, When the instructions in the non-transitory computer-readable storage medium are executed by the processor of the electronic device, the electronic device is enabled to perform an air-to-air cooperative grassland rat hole identification as described in any one of claims 1 to 7.

Images