Rat bird recognition algorithm and multifunctional Rat bird rack

Abstract

The invention provides a raptor recognition algorithm. The algorithm comprises the following steps: S001, obtaining a video and weight data of a raptor; s002, scaling, normalization and data enhancement are carried out on the obtained image of the rapling bird, normalization processing is carried out on the obtained weight data of the rapling bird, the weight data of the rapling bird are paired with image data of the rapling bird, and a multi-modal data set is formed; s003, performing model selection and training on the image data and the weight data of the raptor; s004, performing feature fusion classification on the image data and the weight data of the raptor; and S005, outputting a result. And S006, carrying out actual deployment on the trained model. The invention further provides a multifunctional raptor frame which is used for implementing the raptor recognition algorithm.

Description

Technical Field

[0001] The invention relates to the technical field of grassland rodent control, in particular to a raptor recognition algorithm and a multifunctional raptor stand. Background Art

[0002] Rodent damage is one of the major biological disasters on grasslands, which can lead to forage loss, grassland degradation, and spread plague and other zoonotic diseases. In order to maintain the ecological balance of grasslands, ensure food security, ecological security and people's health and safety, it is very necessary to control grassland rodent damage. Raptor racks are a green prevention and control measure that uses the natural enemies of rodents to prevent and control rodent damage. The principle is to take advantage of the lack of trees on grasslands and the fact that raptors have nowhere to live. Raptor racks of a certain density are set up in rodent-infested areas to provide more habitats for raptors and increase their range of activities on the grassland, thereby increasing the predation rate of raptors on grassland pests, thereby achieving long-term prevention and control of pests. It has been promoted in Xinjiang, Qinghai, Sichuan, Inner Mongolia and other places in my country. However, since raptor racks are generally set up in remote grassland areas, it is not convenient for long-term monitoring. In addition, during the dry season on grasslands, raptors may not go deep into water-scarce areas on grasslands, which is not conducive to grassland rodent control. Therefore, the actual utilization effect of raptor racks, as well as the types, numbers, weights and other related conditions of raptors lack sufficient evaluation, and their actual rodent control capabilities are unclear. Traditional bird identification methods rely on manual classification, which is time-consuming and error-prone. With the development of deep learning technology, vision-based bird identification methods have made significant progress. However, relying solely on visual recognition still has certain limitations in complex environments. This method can provide additional identification features based on weight monitoring information obtained by the raptor rack device to improve identification accuracy. Summary of the invention

[0003] Therefore, it is necessary to provide a raptor identification algorithm and a multifunctional raptor stand to overcome the above problems or at least partially solve or alleviate the above problems.

[0004] To achieve the above object, the present invention proposes a raptor identification algorithm, comprising the following steps:

[0005] S001: Obtain video and weight data of raptors;

[0006] S002: scaling, normalizing and data enhancing the acquired images of the raptors, normalizing the acquired weight data of the raptors, and pairing them with the image data of the raptors to form a multimodal data set;

[0007] S003: Model selection and training for raptor image data and weight data;

[0008] S004: Perform feature fusion classification on the image data and weight data of birds of prey;

[0009] S005: Result output;

[0010] S006: Actual deployment of the trained model.

[0011] In step S003, the total loss function of the model training is obtained by weighted summation of classification loss, positioning loss and confidence loss, where the positioning loss function adopts the CIoU method, and the calculation formula is as follows:

[0012] Positioning loss function:

[0013] The intersection of the predicted box and the true box:

[0014] Trade-off parameters:

[0015] Measuring consistency:

[0016] Among them, A is the predicted bounding box, B is the true bounding box, d is the distance between the center points of the two bounding boxes, c is the maximum diagonal length of the minimum enclosing rectangle of the two bounding boxes, w, h, w gt and h gt are the width and height of the ground-truth bounding box of the predicted bounding box, respectively.

[0017] In step S004, the raptor identification model and the raptor classification model are scored, and the scores of different categories obtained by the two models are scored and summed to obtain the calibrated raptor species probability. The seat with the highest probability is the final raptor species. The scoring formula is as follows:

[0018]

[0019] Among them, i is different types of raptors, T i is the final calibration score predicted by the ensemble method for a certain raptor species, R i To predict the raptor species scores for the recognition model, C i Predict raptor species scores for a classification model.

[0020] The present invention also provides a multifunctional bird of prey stand, comprising:

[0021] A vertical pole, a fixed pulley is arranged on the top of the vertical pole;

[0022] A support, a ferrule is arranged on one side of the support, the ferrule is sleeved on the vertical pole, a hanging ring is arranged on the top of the support, a connecting rope is connected to the hanging ring, the connecting rope passes around the fixed pulley and leads to the ground, and a support platform is arranged at the bottom of the support;

[0023] A horizontal rod, wherein the horizontal rod is arranged on the support platform, both ends of the horizontal rod extend out of the support platform, and a weight sensor is arranged between the horizontal rod and the support platform;

[0024] A water tank, wherein a water trough and an electromagnetic valve seat are arranged at the lower part of the water tank, and the electromagnetic valve seat is connected with the water tank and the water trough;

[0025] A camera, wherein the camera is arranged on the support and faces two ends of the horizontal rod;

[0026] A controller is provided with a processor, a battery module, a memory and a communication module. A signal receiving end of the controller is electrically connected to the weight sensor, and two signal output ends of the controller are electrically connected to the solenoid valve seat and the camera.

[0027] The multifunctional bird of prey stand of the present invention also has the following features.

[0028] It also includes a solar panel, which is arranged on the top of the support and is electrically connected to the battery module through a charging circuit.

[0029] It also includes bird-proof spikes, which are arranged on the edge of the solar cell panel or on both sides and the middle connecting seam of the vertical pole.

[0030] It also includes a GPS module, which is connected to the communication module in the controller through a circuit.

[0031] It also includes a proximity switch or an ultrasonic distance measuring device, and the proximity switch or the ultrasonic distance measuring device is arranged on the support or the horizontal rod.

[0032] A chip sensor is also included, and the chip sensor is arranged at the lower part of the support.

[0033] The raptor recognition algorithm of the present invention realizes efficient and accurate bird recognition by combining the improved lightweight YOLOv5s model and the random forest classification model, which not only improves the accuracy and real-time performance of grassland eagle recognition, but also provides strong technical support for bird protection and ecological research.

[0034] The multifunctional raptor stand of the present invention mainly has a liftable support arranged on the vertical pole, which is convenient for staff to maintain the raptor stand on the ground. At the same time, it does not affect the height of the vertical pole, which easily attracts raptors to land. A camera and a weight sensor are arranged on the support to take pictures and weigh the raptors landing on the horizontal pole, monitor the type, number and individual size of the raptors, and can also attract raptors through water supply through a water tank, so that the raptors can enter the water-deficient area of ​​the grassland during the dry season, which is beneficial to the prevention and control of grassland pests. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of a raptor identification algorithm of the present invention;

[0036] Figure 2 This is the YOLOv5s-MobileNetV3 network structure diagram;

[0037] Figure 3 It is a schematic diagram of the algorithm structure of the raptor identification algorithm of the present invention;

[0038] Figure 4 This is a recognition effect diagram of the raptor recognition algorithm of the present invention;

[0039] Figure 5 A structural diagram of the multifunctional bird of prey stand of the present invention from one viewing angle;

[0040] Figure 6 This is a structural diagram of the multifunctional raptor stand of the present invention from another perspective.

[0041] In the above figures:

[0042] 1 vertical pole; 2 fixed pulley; 3 support; 301 support platform; 302 connecting seat; 303 top frame; 4 rings; 5 lifting rings; 6 horizontal rod; 7 weight sensor; 8 water tank; 9 water tank; 10 solenoid valve seat; 11 camera; 12 controller; 13 solar panel; 14 GPS module; 15 ultrasonic ranging device; 16 chip sensor; 17 anti-bird thorn. DETAILED DESCRIPTION

[0043] In order to explain the technical content, structural features, achieved objectives and effects of the technical solution in detail, the following is a detailed description in conjunction with specific embodiments and accompanying drawings.

[0044] like Figure 1 and Figure 2 As shown, an embodiment of the present invention proposes a raptor identification algorithm, comprising the following steps:

[0045] S001: Use a multifunctional raptor stand to obtain video and weight data of raptors; collect image data and weight data of prairie eagles under different lighting environments at different times through the camera 11 on the vertical pole 1, and the weight data of the raptor is obtained through the weight sensor 7 installed at the bottom of the horizontal pole 6.

[0046] S002: scaling, normalizing and data enhancement of the acquired raptor images, normalizing the acquired raptor weight data, and pairing them with the image data to form a multimodal data set;

[0047] S003: Model selection and training for raptor image data and weight data;

[0048] like Figure 2 and Figure 3 As shown in the figure, in order to further reduce the amount of calculation and save energy consumption, the visual recognition model selects the lightweight model YOLOv5s, and uses MobileNetV3 to replace the original CSPDarkNet backbone network in YOLOv5s. Through transfer learning technology, the MobileNetV3 model pre-trained on ImageNet is used and fine-tuned on the bird dataset. After training, the algorithm is deployed on the installed edge computing device to achieve the purpose of fast and accurate monitoring of wild birds of prey.

[0049] In step S003, the total loss function of the model training is obtained by weighted summation of classification loss, positioning loss and confidence loss, where the positioning loss function adopts the CIoU method, and the calculation formula is as follows:

[0050] Positioning loss function: The quality of model training is solved and evaluated by minimizing the positioning loss function. Generally, the smaller the positioning loss function, the more consistent the predicted box and the real box used to locate the target are;

[0051] The intersection of the predicted box and the true box: Calculate the intersection-union ratio of the predicted box and the true box, that is, the ratio of the intersection area to the union area of ​​the two target boxes;

[0052] Trade-off parameters: Used to define the proportion of consistency;

[0053] Measuring consistency: Measure the consistency of the aspect ratio of two target boxes;

[0054] Among them, A is the predicted bounding box, B is the true bounding box, d is the distance between the center points of the two bounding boxes, c is the maximum diagonal length of the minimum enclosing rectangle of the two bounding boxes, w, h, w gt and h gt are the width and height of the ground-truth bounding box of the predicted bounding box, respectively.

[0055] S004: Perform feature fusion classification on the image data and weight data of birds of prey;

[0056] In step S004, the scale data and weight features of the raptors identified by the image are fused, and the random forest algorithm (RF) is used for joint feature extraction and classification to improve the robustness and accuracy of the model. The raptor recognition model and the raptor classification model are scored, and the scoring ratio is 1:1. The scores of different categories obtained by the two models are scored and summed to obtain the calibrated raptor species probability. The raptor species with the highest probability is the final raptor species. The scoring formula is as follows:

[0057]

[0058] Among them, i is different types of raptors, T i is the final calibration score predicted by the ensemble method for a certain raptor species, R i To predict the raptor species scores for the recognition model, C i Predict raptor species scores for a classification model.

[0059] S005: Model result output; the results obtained after multimodal feature fusion detection are raptor species, raptor location, raptor weight and raptor size, among which, raptor species is the calibrated category, raptor location is the location information obtained by the raptor recognition model, raptor weight is obtained by an electronic scale, and raptor size is the diagonal length of the raptor location rectangular box.

[0060] S006: Deploy the trained model. Deploy the trained model to actual applications, and finally apply it on the hawk scaffold to achieve real-time identification of grassland eagles. Display the identification results through the front-end interface and provide relevant bird information to assist researchers and bird watchers in bird identification. Figure 4 shown.

[0061] refer to Figure 5 and Figure 6 The embodiment of the present invention further proposes a multifunctional raptor stand for implementing the raptor identification algorithm described above, wherein the multifunctional raptor stand comprises: a vertical pole 1, a support 3, a horizontal pole 6, a water tank 8, a camera 11 and a controller 12; a fixed pulley 2 is arranged on the top of the vertical pole 1; a ferrule 4 is arranged on one side of the support 3, the ferrule 4 is sleeved on the vertical pole 1, a hanging ring 5 is arranged on the top of the support 3, a connecting rope is connected to the hanging ring 5, the connecting rope passes around the fixed pulley 2 and leads to the ground, and a supporting platform 301 is arranged at the lower part of the support 3; the horizontal pole 6 is arranged on the supporting platform 301, and the horizontal pole 6 is arranged on the supporting platform 301. The two ends extend out of the support platform 301, and a weight sensor 7 is arranged between the horizontal rod 6 and the support platform 301; a water tank 9 and an electromagnetic valve seat 10 are arranged at the lower part of the water tank 8, and the electromagnetic valve seat 10 is connected with the water tank 8 and the water tank 9; two cameras 11 are arranged on the support 3, and the two cameras 11 are respectively facing the two ends of the horizontal rod 6; a processor, a battery module, a memory and a communication module are arranged in the controller 12, a signal receiving end of the controller 12 is electrically connected to the weight sensor 7, and two signal output ends of the controller 12 are electrically connected to the electromagnetic valve seat 10 and the camera 11.

[0062] The pole 1 can be made of wood or metal. The pole 1 is erected on the ground. A fixed pulley 2 is arranged on the top of the pole 1. The support 3 is a vertical metal plate or wooden board. A ring 4 is arranged on the back side of the support 3, which is slidably mounted on the pole 1 through the ring 4. A lifting ring 5 is arranged on the top of the support 3. The lifting ring 5 is connected to a cable that passes around the fixed pulley 2 and is wound and anchored on the ground. By untying the wound cable, the support 3 can be lowered from the top of the pole 1 to the ground through the fixed pulley 2, which is convenient for the staff to maintain the raptor rack.

[0063] A horizontal support platform 301 is provided at the lower end of the support 3, a connecting seat 302 is horizontally provided on the support platform 301, the middle part of the horizontal rod 6 is fixed in the connecting seat 302, and a weight sensor 7 is provided between the lower part of the connecting seat 302 and the support platform 301. When the bird of prey falls on the horizontal rod 6, the weight sensor 7 will automatically weigh the weight of the bird of prey.

[0064] One side of the solenoid valve seat 10 is fixedly connected to the support 3, the upper port of the solenoid valve seat 10 is connected to the water tank 8, and a water tank 9 is provided on one side of the solenoid valve seat 10. The solenoid valve seat 10 can release the water in the water tank 8 into the water tank 9 when powered on for the birds of prey to drink.

[0065] A top frame 303 is also provided on the upper part of the support 3, and a camera 11 is provided on the lower side of the top frame 303, and can be used to take pictures of the raptors landing on the horizontal rod 6;

[0066] The controller 12 is fixed on the inner side of the support 3. A processor, a battery module, a memory and a communication module are arranged in the controller 12. When the raptor falls on the horizontal rod 6, the controller 12 receives the weight measurement signal from the weight sensor 7, and controls the camera 11 to shoot the raptor. The controller 12 also turns on the electromagnetic switch in the electromagnetic valve seat 10, and fills the water tank 9 with water for the raptor to drink. The processor in the controller 12 processes the received weight measurement signal and video signal, and stores the video information in the memory. Finally, the weight information and video information of the raptor are sent to the network through the communication module, and the staff can receive it through the communication equipment.

[0067] like Figure 5 and Figure 6 According to an embodiment of the present invention, it also includes a solar panel 13, which is arranged on the top of the support 3 and is electrically connected to the battery module through a charging circuit.

[0068] The solar panel 13 is fixed above the top frame 303. The solar panel 13 is electrically connected to the battery module through a charging circuit, and can charge the battery to improve the endurance of the raptor frame.

[0069] like Figure 5 and Figure 6, and also includes bird-proof spikes 17, which are arranged on the edges of the solar cell panel 13 or on both sides and the middle connecting seam of the vertical pole 1.

[0070] The bird-proof spikes can prevent the raptors from landing directly on the solar panel 13, which would make it impossible to collect the raptor data, and only allow the raptors to land on the horizontal rod 6.

[0071] refer to Figure 5 and Figure 6 According to an embodiment of the present invention, it also includes a GPS module 14, and the GPS module 14 is connected to the communication module in the controller 12 through a circuit.

[0072] The GPS module 14 is electrically connected to the communication module and can transmit signals wirelessly, making it convenient for staff to remotely receive image and video data.

[0073] refer to Figure 5 and Figure 6 According to an embodiment of the present invention, it also includes a proximity switch or an ultrasonic distance measuring device 15, and the proximity switch 15 is arranged on the support 3 or the horizontal rod 6.

[0074] In order to save electricity, a proximity switch or an ultrasonic ranging device 15 can be set, and the detection end of the proximity switch or ultrasonic ranging device 15 is facing the two ends of the horizontal rod 6. When the raptor falls on the horizontal rod 6, the proximity switch or ultrasonic ranging device 15 is triggered. At this time, the controller 12 starts to start the camera 11 to shoot the raptor, and at the same time starts the weight sensor 7 to measure the weight of the raptor.

[0075] like Figure 5 and Figure 6 According to an embodiment of the present invention, it also includes a chip sensor 16, which is arranged at the lower part of the support 3.

[0076] After the raptor is released after being equipped with a sensing chip, when the raptor lands on the horizontal rod 6, the chip sensor 16 can sense the chip fixed on the raptor, thereby immediately identifying the information of the raptor.

[0077] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "include..." or "comprise..." do not exclude the existence of other elements in the process, method, article or terminal device including the elements. In addition, in this article, "greater than", "less than", "exceed" and the like are understood to exclude the number itself; "above", "below", "within" and the like are understood to include the number itself.

[0078] Although the above embodiments have been described, once those skilled in the art know the basic creative concepts, they can make additional changes and modifications to these embodiments. Therefore, the above description is only an embodiment of the present invention and does not limit the patent protection scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the specification and drawings of the present invention, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A raptor recognition algorithm, characterized in that: The following steps are involved: S001: Obtain video and weight data of raptors; S002: scaling, normalizing and data enhancement of the acquired raptor images, normalizing the acquired raptor weight data, and pairing them with the raptor image data to form a multimodal data set; S003: Model selection and training for raptor image data and weight data; S004: Perform feature fusion classification on the image data and weight data of birds of prey; S005: Model result output; S006: Deploy the trained model.

2. The raptor identification algorithm according to claim 1, characterized in that: In step S003, the total loss function of the model training is obtained by weighted summation of classification loss, positioning loss and confidence loss, where the positioning loss function adopts the CIoU method, and the calculation formula is as follows: Positioning loss function: The intersection of the predicted box and the true box: Trade-off parameters: Measuring consistency: Among them, A is the predicted bounding box, B is the true bounding box, d is the distance between the center points of the two bounding boxes, c is the maximum diagonal length of the minimum enclosing rectangle of the two bounding boxes, w, h, w gt and h gt are the width and height of the ground-truth bounding box of the predicted bounding box, respectively.

3. The raptor identification algorithm according to claim 2, characterized in that: In step S004, the raptor identification model and the raptor classification model are scored, and the scores of different categories obtained by the two models are scored and summed to obtain the calibrated raptor species probability. The seat with the highest probability is the final raptor species. The scoring formula is as follows: Among them, i is different types of raptors, T i is the final calibration score predicted by the ensemble method for a certain raptor species, R i To predict the raptor species scores for the recognition model, C i Predict raptor species scores for a classification model.

4. A multifunctional raptor stand, characterized in that: include: A vertical pole (1), wherein a fixed pulley (2) is arranged at the top of the vertical pole (1); A support (3); a ring (4) is provided on one side of the support (3), the ring (4) is sleeved on the vertical pole (1), a hanging ring (5) is provided on the top of the support (3), a connecting rope is connected to the hanging ring (5), the connecting rope passes around the fixed pulley (2) and leads to the ground, and a support platform (301) is provided at the bottom of the support (3) a horizontal rod (6), wherein the horizontal rod (6) is arranged on the support platform (301), both ends of the horizontal rod (6) extend out of the support platform (301), and a weight sensor (7) is arranged between the horizontal rod (6) and the support platform (301); A water tank (8), wherein a water tank (9) and an electromagnetic valve seat (10) are provided at the lower part of the water tank (8), and the electromagnetic valve seat (10) is connected to the water tank (8) and the water tank (9); A camera (11), wherein the camera (11) is arranged on the support (3), and the camera (11) faces two ends of the horizontal rod (6); A controller (12), wherein the controller (12) is provided with a processor, a battery module, a memory and a communication module, a signal receiving end of the controller (12) is electrically connected to the weight sensor (7), and two signal output ends of the controller (12) are electrically connected to the electromagnetic valve seat (10) and the camera (11).

5. The multifunctional raptor stand according to claim 1, characterized in that: It also comprises a solar panel (13), wherein the solar panel (13) is arranged on the top of the support (3), and the solar panel (13) is connected to the battery module circuit via a charging circuit.

6. The multifunctional raptor stand according to claim 1, characterized in that: It also includes bird-proof spikes (17), which are arranged at the edge of the solar cell panel (13) or at the two sides and the middle connection seam of the vertical pole (1).

7. The multifunctional raptor stand according to claim 1, characterized in that: It also includes a GPS module (14), and the GPS module (14) is connected to the communication module in the controller (12) via a circuit.

8. The multifunctional raptor stand according to claim 1, characterized in that: It also includes a proximity switch, wherein the proximity switch or the ultrasonic distance measuring device (15) is arranged on the support (3) or the horizontal rod (6).

9. The multifunctional raptor stand according to claim 1, characterized in that: It also includes a chip sensor (16), which is arranged at the lower part of the support (3).

Images