Method and device for trapping wild animals

Abstract

The application discloses a wild animal trapping method and device, and relates to the field of ecological research devices. The wild animal trapping method comprises a triggering mechanism, a multi-modal bait system, a safe detention cabin and a data integration module. The triggering mechanism is integrated with an AI image recognition module and an edge computing chip, and is used for real-time analysis of the morphological characteristics and behavior patterns of target species. The wild animal trapping method and device realize accurate trapping through an AI-driven multi-modal collaborative mechanism, integrate species identification, adopt confidence dynamic threshold judgment and environment response bait release, adopt segmented function control and stress grading release, adopt a three-level strategy, utilize the above three core technologies, reduce animal injury rate while improving capture success rate, and construct a full-link data closed loop from field sensing to cloud analysis, thereby providing a high-reliability tool for ecological research.

Description

Technical Field

[0001] This invention relates to the field of ecological research equipment technology, specifically to a method and apparatus for trapping wild animals. Background Technology

[0002] In the process of rescue, conservation, and scientific research, wildlife centers rely on trapping or capturing wild animals to obtain research samples and implement conservation actions. Whether ecologists are studying the population dynamics and behavioral habits of wild animals, or veterinarians are treating injured wild animals, the safe capture of wild animals is a necessary first step.

[0003] However, current traditional wildlife trapping or capture devices have the following shortcomings:

[0004] First, traditional mechanical traps (such as spring clips or cages) rely on physical triggering mechanisms, which can easily lead to animal injury or death. They also lack intelligent recognition capabilities, resulting in a high probability of accidentally trapping non-target species.

[0005] Secondly, rigid confinement structures can lead to collision damage, and traditional confinement chambers lack stress monitoring, which may cause captured animals to be injured or die due to panic.

[0006] Thirdly, regarding data recording, existing technologies largely rely on manual inspections, which cannot transmit captured information to remote monitoring centers in real time, thus limiting the timeliness and data integrity of ecological research.

[0007] To address the shortcomings of existing technologies, this invention provides a method and apparatus for trapping wild animals, thereby solving the aforementioned problems. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method and apparatus for trapping wild animals. It achieves precise trapping through an AI-driven multimodal collaborative mechanism, integrating three core technologies: species identification (dynamic confidence threshold), environmentally responsive bait delivery (piecewise function control), and stress-level release (three-level strategy). This improves the capture success rate while reducing animal injury rate, and constructs a closed-loop data system from on-site perception to cloud analysis, providing a highly reliable tool for ecological research.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a method for trapping wild animals, comprising:

[0010] S1, set a trigger mechanism, which integrates an AI image recognition module and an edge computing chip. The AI ​​image recognition module and the edge computing chip are used to analyze the morphological characteristics and behavioral patterns of the target species in real time. When the recognition accuracy is ≥95%, a trigger signal is output.

[0011] S2, a multimodal bait system is set up, including an odor release unit, a sound wave generation unit, and a food delivery unit that are dynamically linked through environmental sensors, wherein the bait type and delivery dosage satisfy a functional relationship:

[0012]

[0013] in, The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH.

[0014] S3, set up a safe detention chamber. The safe detention chamber is equipped with a pressure sensor and a stress monitoring module. When the sound pressure level SPL in the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is implemented.

[0015] S4, set up a data integration module. Each module interacts with the other via a CAN bus. The trigger mechanism controls the activation sequence of the multimodal decoy system, and the release signal from the detention chamber is fed back to the data integration module to generate an operation log.

[0016] Preferably, the AI ​​image recognition module integrates a transfer learning framework, loading a species-specific feature extraction layer on top of a lightweight CNN model, and dynamically adjusting the recognition confidence threshold α:

[0017]

[0018] The threshold is β, which is the humidity influence coefficient. The triggering mechanism is activated when the recognition results are consistent for N ≥ 5 consecutive frames.

[0019] Preferably, the edge computing chip has a built-in FPGA accelerator to perform hardware-level optimization of the convolutional layers of the CNN model, so that the computation latency is ≤0.3 seconds, and is equipped with a temperature-adaptive clock circuit that automatically reduces the frequency by 10% when T≥40℃.

[0020] Preferably, the odor-releasing unit comprises a microfluidic chip-driven array of replaceable capsules, each capsule storing volatile organic compounds (VOCs), and the mixing ratio R is controlled by environmental parameters.

[0021] ;

[0022] Organic VOCs are mixed in proportion using a piezoelectric micropump.

[0023] Preferably, the secure detention chamber adopts a gradient buffer structure: the inner layer is memory foam with a porosity of ≥90%, the outer layer is a honeycomb aluminum plate, and the opening angle θ of the chamber door is precisely controlled by a stepper motor to satisfy θ=arcsin(v / v_max), where v is the animal's entry speed and v_max is the design maximum speed;

[0024] The three-level release strategy includes:

[0025] Level 1 release: Electromagnetic unlocking of the hatch;

[0026] Secondary release: Activate the sedative spray device, spray volume Q=0.01M, where M is the animal's weight estimated based on infrared thermal imaging;

[0027] Level 3 Release: Sends an encrypted alert to the remote monitoring center.

[0028] Preferably, the nozzle surface of the sedative spray device is provided with a nano-needle structure, the spray particle size D_p≤20μm, the hydrophobic coating contact angle≥150°, and the spray direction can be dynamically deflected ±30° according to the position of the animal in the cabin.

[0029] Preferably, the data integration module automatically associates individual IDs with RFID tags and biometric identification devices, and synchronizes the capture time t, GPS coordinates, environmental parameters, and behavioral video V to the cloud at intervals of ≤5 minutes; the data integration module has a built-in communication decision engine, which switches to LoRa mode when the 5G signal strength RSSI ≤ -120dBm and starts the data compression protocol. The compression ratio is dynamically adjusted according to the transmission delay to ensure data transmission reliability ≥99.9%.

[0030] Preferably, the food dispensing unit is a centrifugal rotating disc with an electromagnetically controlled valve on the disc surface. The valve opening time Δt is positively correlated with the size of the target species.

[0031] Preferably, the stress monitoring module integrates voiceprint feature matching and body temperature change analysis. When the voiceprint matching degree is ≥85% and the body temperature change rate ΔT / Δt is ≥2℃ / min, it is determined to be a state of high stress and the secondary release is initiated first.

[0032] A second aspect of this invention discloses a wildlife trapping device, applied to the aforementioned wildlife trapping method, the device comprising:

[0033] The triggering mechanism integrates an AI image recognition module and an edge computing chip. The AI ​​image recognition module and the edge computing chip are used to analyze the morphological characteristics and behavioral patterns of the target species in real time. When the recognition accuracy is ≥95%, a trigger signal is output.

[0034] A multimodal bait system comprises an odor release unit, a sound wave generation unit, and a food delivery unit, dynamically linked by environmental sensors. The bait type and dosage satisfy a functional relationship:

[0035]

[0036] in, The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH.

[0037] The safe confinement chamber is equipped with a built-in pressure sensor and stress monitoring module. When the sound pressure level SPL inside the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is implemented.

[0038] The data integration module enables data interaction between various modules via a CAN bus. The trigger mechanism controls the activation sequence of the multimodal decoy system, and the release signal from the safe detention chamber is fed back to the data integration module to generate an operation log.

[0039] The device also includes a self-cleaning system that performs after each release:

[0040] Irradiate with an ultraviolet lamp at a wavelength of 265nm for ≥30 seconds;

[0041] High-pressure airflow is ejected tangentially along the inner wall of the chamber, with an airflow velocity of v = 0.5 m / s;

[0042] The residue is recovered to a sealed container via a negative pressure pipeline.

[0043] Its beneficial effects are as follows:

[0044] 1. Wildlife trapping methods and devices with high trapping success rates. By loading species-specific feature layers through a transfer learning framework and combining them with temperature and humidity adaptive confidence thresholds, the target recognition accuracy is improved. The multimodal bait system dynamically adjusts the VOCs mixing ratio, sound wave frequency, and food delivery amount based on environmental parameters, thereby improving attraction efficiency.

[0045] 2. Wildlife trapping methods and devices ensure animal safety. The secure confinement chamber utilizes a memory foam and honeycomb aluminum panel structure with a gradient cushioning design. The memory foam provides a soft and comfortable contact surface for the wild animals, reducing physical injuries caused by struggling and collisions during confinement. The honeycomb aluminum panels, while ensuring structural strength, also offer excellent cushioning performance, further reducing the risk of animal injury. Simultaneously, a stress monitoring module monitors the animal's stress state in real time, and a three-level release strategy significantly reduces animal stress responses. The three-level release strategy determines the stress level through voiceprint matching and body temperature change rate, while the sedative spray particle size is ≤20μm, precisely controlling the dosage and reducing animal injury rates.

[0046] 3. The wildlife trapping method and apparatus are highly intelligent. The data integration and transmission system enables the aggregation of image data collected by the triggering mechanism, environmental data and bait release status data from the multimodal bait system, and pressure and stress monitoring data from the secure confinement chamber via wired or wireless transmission. The data integration module achieves real-time transmission and remote monitoring of the captured data, providing strong support for ecological research and intelligently controlling each device according to preset logic and algorithms. Furthermore, data can be transmitted to a remote monitoring terminal, allowing staff to monitor the device's operation and the status of captured wildlife in real time, significantly improving work efficiency and reducing labor costs. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0049] Figure 2 This is a schematic diagram of the overall device structure of the present invention;

[0050] Figure 3 This is a schematic diagram of the structure of the memory foam and honeycomb aluminum plate of the present invention;

[0051] Figure 4 This is the overall control flowchart of the device of the present invention;

[0052] Figure 5 This is the AI ​​recognition trigger logic diagram for the present invention;

[0053] Figure 6 This is the control diagram for the multimodal decoy of the present invention;

[0054] Figure 7This is a diagram illustrating the three-level release strategy of the present invention;

[0055] Figure 8 This is a data integration and transmission diagram of the present invention;

[0056] Figure 9 This is a flowchart of the self-cleaning system of the present invention.

[0057] In the diagram: 100, triggering mechanism; 101, AI image recognition module; 102, edge computing chip; 200, multimodal decoy system; 201, odor release unit; 202, sound wave generation unit; 203, food delivery unit; 204, environmental sensor; 300, secure detention chamber; 301, pressure sensor; 302, stress monitoring module; 303, sedative spray device; 304, memory foam; 305, honeycomb aluminum panel. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] 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.

[0060] This invention discloses a method for trapping wild animals, according to the appendix. Figure 1 To be continued Figure 9 As shown, it includes:

[0061] S1, set a trigger mechanism 100, the trigger mechanism 100 integrates an AI image recognition module 101 and an edge computing chip 102, the AI ​​image recognition module 101 and the edge computing chip 102 are used to analyze the morphological characteristics and behavioral patterns of the target species in real time, and output a trigger signal when the recognition accuracy is ≥95%;

[0062] S2, a multimodal bait system 200 is set up, including an odor release unit 201, a sound wave generation unit 202, and a food delivery unit 203, which are dynamically linked through an environmental sensor 204. The bait type and the delivery dosage D satisfy a functional relationship:

[0063]

[0064] in, The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH.

[0065] S3, set up a safe detention chamber 300. The safe detention chamber 300 has a built-in pressure sensor 301 and a stress monitoring module 302. When the sound pressure level SPL in the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is executed.

[0066] S4, set up a data integration module. Each module realizes data interaction through the CAN bus. The trigger mechanism 100 controls the activation sequence of the multimodal decoy system 200, and the release signal of the safe detention chamber 300 is fed back to the data integration module to generate an operation log.

[0067] According to the appendix Figure 2 As shown, the triggering mechanism 100 integrates an AI image recognition module 101 and an edge computing chip 102. The AI ​​image recognition module 101 adopts a transfer learning framework, loading a species-specific feature extraction layer on top of a lightweight CNN model. It achieves high-precision target species recognition by dynamically adjusting the recognition confidence threshold α, which is a base threshold plus an adjustment value for the humidity influence coefficient β. When the recognition results are consistent for N≥5 consecutive frames and the recognition accuracy is ≥95%, the edge computing chip 102 outputs a trigger signal. The edge computing chip 102 has a built-in FPGA accelerator, performs hardware-level optimization of the convolutional layers of the CNN model to ensure a computational latency of ≤0.3 seconds, and is equipped with a temperature-adaptive clock circuit that automatically reduces the frequency by 10% in high-temperature environments (T≥40℃) to prevent overheating.

[0068] According to the appendix Figure 2 , 6 As shown, the multimodal bait system 200 includes an odor release unit 201, a sound wave generation unit 202, and a food delivery unit 203, which are dynamically linked through an environmental sensor 204. The bait type and the delivery dosage D satisfy a functional relationship:

[0069]

[0070] The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH.

[0071] The multimodal bait system 200 comprises an odor release unit 201, a sound wave generating unit 202, and a food delivery unit 203, all dynamically linked via an environmental sensor 204. The odor release unit 201 employs a microfluidic chip-driven array of replaceable capsules, each storing different volatile organic compounds (VOCs). The mixing ratio R is controlled by environmental parameters such as temperature and relative humidity, and the odor is released proportionally via a piezoelectric micropump to attract target species. The sound wave generating unit 202 emits sound waves of specific frequencies based on the auditory characteristics of the target species, enhancing the trapping effect. The food delivery unit 203 uses a centrifugal rotating disc design with an electromagnetically controlled valve. The valve opening time Δt is positively correlated with the target species' size, ensuring an appropriate amount of food is delivered.

[0072] According to the appendix Figure 2 , 7 As shown, the safe detention chamber 300 has a built-in pressure sensor 301 and a stress monitoring module 302. When the sound pressure level SPL inside the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is implemented:

[0073] Level 1 release: Electromagnetic unlocking of the hatch;

[0074] Secondary release: Activate sedative spray device 303, spray volume Q=0.01M, where M is the animal's weight estimated based on infrared thermal imaging;

[0075] Level 3 Release: Sends an encrypted alert to the remote monitoring center;

[0076] According to the appendix Figure 2 , 3 As shown in Figure 7, the secure detention chamber 300 incorporates a pressure sensor 301 and a stress monitoring module 302. It employs a gradient buffer structure, with an inner layer of memory foam 304 with a porosity ≥90% and an outer layer of honeycomb aluminum plate 305, effectively absorbing impact and protecting the captured animal. The chamber door opening angle θ is precisely controlled by a stepper motor, satisfying θ=arcsin(v / v_max), where v is the animal's entry speed and v_max is the design maximum speed. When the sound pressure level SPL inside the chamber ≥80dB or the animal's movement frequency F ≥5Hz, a three-level release strategy is implemented: Level 1 release is electromagnetic unlocking of the chamber door; Level 2 release is activation of the sedative spray device 303, with a spray volume Q=0.01M, where M is the animal's weight estimated based on infrared thermal imaging; Level 3 release is sending an encrypted alarm to a remote monitoring center.

[0077] According to the appendix Figure 2 , 8 As shown, the data integration module automatically associates individual IDs with RFID tag 401 and biometric identification device 402, and synchronizes the capture time t, GPS coordinates, environmental parameters and behavioral video V to the cloud at intervals of ≤5 minutes.

[0078] The data integration module automatically associates individual IDs with RFID tags 401 and biometric identification devices 402, and synchronizes the capture time t, GPS coordinates, environmental parameters, and behavioral video V to the cloud at intervals of ≤5 minutes. This module has a built-in communication decision engine that automatically switches to LoRa mode when the 5G signal strength RSSI is ≤-120dBm and initiates a data compression protocol. The compression ratio is dynamically adjusted based on transmission latency to ensure data transmission reliability ≥99.9%.

[0079] Each module interacts with the other via a CAN bus. The trigger mechanism 100 controls the activation sequence of the multimodal decoy system 200, and the release signal of the safe detention chamber 300 is fed back to the data integration module to generate an operation log.

[0080] The AI ​​image recognition module 101 integrates a transfer learning framework, loading a species-specific feature extraction layer on top of a lightweight CNN model, and dynamically adjusting the recognition confidence threshold α.

[0081]

[0082] The baseline threshold is β, which is the humidity influence coefficient. The trigger mechanism 100 is activated when the recognition results are consistent for N ≥ 5 consecutive frames.

[0083] The edge computing chip 102 has a built-in FPGA accelerator that performs hardware-level optimization on the convolutional layers of the CNN model, making the computation latency ≤0.3 seconds. It is also equipped with a temperature-adaptive clock circuit that automatically reduces the clock frequency by 10% when T≥40℃.

[0084] According to the appendix Figure 4 As shown, the overall control process of this wildlife trapping method is as follows: First, the AI ​​image recognition module 101 of the triggering mechanism 100 identifies the target species. After successful identification, the multimodal bait system 200 is triggered to release the bait. After the animal enters the safe confinement chamber 300, the pressure sensor 301 and the stress monitoring module 302 monitor the animal's status. The corresponding release strategy is executed according to the monitoring results. Finally, the data integration module synchronizes the captured data to the cloud.

[0085] According to the appendix Figure 5 As shown, the AI ​​recognition trigger logic is as follows: the AI ​​image recognition module 101 continuously collects image data and performs real-time analysis through a lightweight CNN model; when the recognition results are consistent for N≥5 consecutive frames and the recognition accuracy is ≥95%, the edge computing chip 102 confirms the target species and outputs a trigger signal; the trigger signal activates the multimodal decoy system 200.

[0086] Odor release unit 201 includes a microfluidic chip-driven array of replaceable capsules, each storing volatile organic compounds (VOCs), with the mixing ratio R controlled by environmental parameters.

[0087] ;

[0088] Organic VOCs are mixed in proportion using a piezoelectric micropump.

[0089] The safe detention chamber 300 adopts a gradient buffer structure: the inner layer is memory foam 304 with a porosity of ≥90%, and the outer layer is honeycomb aluminum plate 305. The opening angle θ of the chamber door is precisely controlled by a stepper motor to satisfy θ=arcsin(v / v_max), where v is the animal's entry speed and v_max is the design maximum speed.

[0090] The nozzle surface of the sedative spray device 303 is provided with a nano needle-like structure, the spray particle size D_p≤20μm, the hydrophobic coating contact angle≥150°, and the spray direction can be dynamically deflected ±30° according to the position of the animal in the cabin.

[0091] The data integration module has a built-in communication decision engine. When the 5G signal strength RSSI is ≤-120dBm, it switches to LoRa mode and starts the data compression protocol. The compression ratio is dynamically adjusted according to the transmission delay to ensure data transmission reliability ≥99.9%.

[0092] The food dispensing unit 203 uses a centrifugal rotating disc with an electromagnetically controlled valve on the disc surface. The valve opening time Δt is positively correlated with the body size of the target species.

[0093] The stress monitoring module 302 integrates voiceprint feature matching and body temperature change analysis. When the voiceprint matching degree is ≥85% and the body temperature change rate ΔT / Δt is ≥2℃ / min, it is determined to be a state of high stress and the secondary release is initiated first.

[0094] The second aspect of this invention discloses a wildlife trapping device, applied to a wildlife trapping method, the device comprising:

[0095] The triggering mechanism 100 integrates an AI image recognition module 101 and an edge computing chip 102. The AI ​​image recognition module 101 and the edge computing chip 102 are used to analyze the morphological characteristics and behavioral patterns of the target species in real time, and output a trigger signal when the recognition accuracy is ≥95%.

[0096] A multimodal bait system 200 includes an odor release unit 201, a sound wave generation unit 202, and a food delivery unit 203, which are dynamically linked via an environmental sensor 204. The bait type and the delivery dosage D satisfy a functional relationship:

[0097]

[0098] Where is the optimal temperature for the species, is the temperature tolerance threshold, is the weighting coefficient, and RH is the relative humidity;

[0099] The safe confinement chamber 300 has a built-in pressure sensor 301 and a stress monitoring module 302. When the sound pressure level SPL in the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is implemented.

[0100] The data integration module enables data interaction between various modules via a CAN bus. The trigger mechanism 100 controls the activation timing of the multimodal decoy system 200, and the release signal of the safe detention chamber 300 is fed back to the data integration module to generate an operation log.

[0101] The device also includes a self-cleaning system that performs after each release:

[0102] Irradiate with an ultraviolet lamp at a wavelength of 265nm for ≥30 seconds;

[0103] High-pressure airflow is ejected tangentially along the inner wall of the chamber, with an airflow velocity of v = 0.5 m / s;

[0104] The residue is recovered to a sealed container via a negative pressure pipeline.

[0105] According to the appendix Figure 9 As shown, the self-cleaning system performs the following operations after each release: an ultraviolet lamp irradiates the inner wall of the chamber at a wavelength of 265nm for ≥30 seconds to kill residual microorganisms; a high-pressure airflow is sprayed tangentially along the inner wall of the chamber at a speed of v=0.5m / s to remove residues; and the residues are recovered to a sealed container through a negative pressure pipeline to prevent cross-contamination.

[0106] The working principle of this wildlife trapping method is as follows: Triggering mechanism 100 detects the target wildlife using AI image recognition and edge computing technology. Multimodal bait system 200 dynamically adjusts the combination of scent, sound, and food to attract the target into the safe confinement chamber 300. The safe confinement chamber 300 confines the wildlife after entry and monitors its status through pressure sensor 301 and stress monitoring module 302. Stress monitoring module 302 triggers a tiered release to avoid animal injury, and if necessary, activates a sedative spray device 303. The memory foam 304 and honeycomb aluminum panel 305 structure provide a comfortable and safe confinement environment for the wildlife, reducing its risk of injury. Capture data is compressed and transmitted to the cloud to guide subsequent trapping strategy optimization.

[0107] First, based on the habits and activity areas of the target wild animal, a suitable placement location is selected and the device is installed, with GPS automatically recording the coordinates. Then, the environmental sensor 204 activates the multimodal bait system 200, and the AI ​​image recognition module 101 begins monitoring; after the animal enters the safe confinement chamber 300, the system automatically performs locking and stress assessment; based on the stress level, it selects natural release or sedation-assisted release; the self-cleaning system disinfects the chamber, preparing it for the next use.

[0108] The beneficial effects of this wildlife trapping method are as follows: the transfer learning framework improves species identification accuracy to ≥95%, and the humidity-adaptive threshold reduces false triggering rates; environmental parameters are linked to bait delivery, such as temperature-segmented control of VOCs ratios to enhance attraction efficiency, and the gradient buffer structure (304 memory foam + 305 honeycomb aluminum plate) and three-stage release strategy minimize stress damage. 5G / LoRa dual-mode transmission ensures data reliability to ≥99.9% in harsh environments; the data integration module enables real-time transmission and remote monitoring of captured data, providing strong support for ecological research.

[0109] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0110] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for trapping wild animals, characterized in that, include: S1, set a trigger mechanism (100), the trigger mechanism (100) integrates an AI image recognition module (101) and an edge computing chip (102), the AI ​​image recognition module (101) and the edge computing chip (102) are used to analyze the morphological characteristics and behavioral patterns of the target species in real time, and output a trigger signal when the recognition accuracy is ≥95%; S2, a multimodal bait system (200) is set up, including an odor release unit (201), a sound wave generation unit (202), and a food delivery unit (203) that are dynamically linked through an environmental sensor (204), wherein the bait type and the delivery dosage (D) satisfy a functional relationship: in, The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH. S3, set up a safe detention chamber (300), the safe detention chamber (300) is equipped with a pressure sensor (301) and a stress monitoring module (302), when the sound pressure level SPL in the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is executed; S4, set up a data integration module. Each module realizes data interaction through the CAN bus. The trigger mechanism (100) controls the activation sequence of the multimodal decoy system (200). The release signal of the detention chamber (300) is fed back to the data integration module to generate an operation log.

2. The method for trapping wild animals according to claim 1, characterized in that, The AI ​​image recognition module (101) integrates a transfer learning framework, loads a species-specific feature extraction layer on top of a lightweight CNN model, and dynamically adjusts the recognition confidence threshold α: The baseline threshold is β, which is the humidity influence coefficient. The trigger mechanism is activated when the recognition results are consistent for N ≥ 5 consecutive frames (100).

3. The method for trapping wild animals according to claim 2, characterized in that, The edge computing chip (102) has a built-in FPGA accelerator to perform hardware-level optimization of the convolutional layers of the CNN model, so that the computation delay is ≤0.3 seconds, and is equipped with a temperature adaptive clock circuit that automatically reduces the frequency by 10% when T≥40℃.

4. The method for trapping wild animals according to claim 1, characterized in that, The odor release unit (201) comprises a microfluidic chip-driven array of replaceable capsules, each storing volatile organic compounds (VOCs), with the mixing ratio R controlled by environmental parameters. ； Organic VOCs are mixed in proportion using a piezoelectric micropump.

5. A method for trapping wild animals according to claim 1, characterized in that, The safe detention chamber (300) adopts a gradient buffer structure: the inner layer is memory foam (304) with a porosity of ≥90%, and the outer layer is a honeycomb aluminum plate (305). The opening angle θ of the chamber door is precisely controlled by a stepper motor to satisfy θ=arcsin(v / v_max), where v is the animal's entry speed and v_max is the design maximum speed. The three-level release strategy includes: Level 1 release: Electromagnetic unlocking of the hatch; Secondary release: Activate the sedative spray device (303), spray volume Q=0.01M, where M is the animal's weight estimated based on infrared thermal imaging; Level 3 Release: Sends an encrypted alert to the remote monitoring center.

6. The method for trapping wild animals according to claim 1, characterized in that, The sedative spray device (303) has a nano needle-like structure on the nozzle surface, a spray particle size D_p≤20μm, a hydrophobic coating contact angle≥150°, and the spray direction can be dynamically deflected ±30° according to the position of the animal in the cabin.

7. The method for trapping wild animals according to claim 1, characterized in that, The data integration module automatically associates individual IDs with RFID tags (401) and biometric identification devices (402), and synchronizes the capture time t, GPS coordinates, environmental parameters, and behavioral video V to the cloud at intervals of ≤5 minutes. The data integration module has a built-in communication decision engine. When the 5G signal strength RSSI is ≤-120dBm, it switches to LoRa mode and starts the data compression protocol. The compression ratio is dynamically adjusted according to the transmission delay to ensure data transmission reliability ≥99.9%.

8. A method for trapping wild animals according to claim 4, characterized in that, The food delivery unit (203) adopts a centrifugal rotating disc with an electromagnetically controlled valve on the disc surface. The valve opening time Δt is positively correlated with the body size of the target species.

9. A method for trapping wild animals according to claim 1, characterized in that, The stress monitoring module (302) integrates voiceprint feature matching and body temperature change analysis. When the voiceprint matching degree is ≥85% and the body temperature change rate ΔT / Δt is ≥2℃ / min, it is determined to be a state of high stress and the secondary release is initiated first.

10. A wildlife trapping device, applied to a wildlife trapping method according to any one of claims 1-9, characterized in that, The device includes: The triggering mechanism (100) integrates an AI image recognition module (101) and an edge computing chip (102). The AI ​​image recognition module (101) and the edge computing chip (102) are used to analyze the morphological characteristics and behavioral patterns of the target species in real time, and output a trigger signal when the recognition accuracy is ≥95%. A multimodal bait system (200) includes an odor release unit (201), a sound wave generation unit (202), and a food delivery unit (203) dynamically linked via an environmental sensor (204), wherein the bait type and the delivery dosage (D) satisfy a functional relationship: in, The optimal temperature for the species This is the temperature tolerance threshold. , Here, RH is the relative humidity, and the weighting factor is RH. The safe confinement chamber (300) has a built-in pressure sensor (301) and stress monitoring module (302). When the sound pressure level SPL in the chamber is ≥80dB or the animal movement frequency F is ≥5Hz, a three-level release strategy is implemented. The data integration module enables data interaction between modules via a CAN bus. The trigger mechanism (100) controls the activation sequence of the multimodal decoy system (200), and the release signal of the safe detention chamber (300) is fed back to the data integration module to generate an operation log. The device also includes a self-cleaning system that performs after each release: Irradiate with an ultraviolet lamp at a wavelength of 265nm for ≥30 seconds; High-pressure airflow is ejected tangentially along the inner wall of the chamber, with an airflow velocity of v = 0.5 m / s; The residue is recovered to a sealed container via a negative pressure pipeline.

Images