Animal target detection method, device, equipment, medium and computer program product

Abstract

The invention provides an animal target detection method, device and equipment, a medium and a computer program product, and relates to the technical field of ecological detection, and the method comprises the steps: obtaining a preprocessed radar echo signal; multiple features are extracted from the preprocessed radar echo signals; the various characteristics comprise micro-Doppler characteristics, geometric characteristics, polarization characteristics and Doppler spectral moments; and inputting the multiple features into a pre-constructed random forest model to obtain an animal target category output after the pre-constructed random forest model classifies the multiple features. According to the invention, animal target detection is carried out by using a radar detection technology, traditional satellite, aircraft and infrared camera detection is replaced, the cost is reduced, and compared with an infrared camera, the system is easier to maintain and has no damage to animals, the economic cost and time cost of ecological detection are reduced, and the efficiency of animal target detection is improved.

Description

Technical Field

[0001] The present invention relates to the field of ecological detection technology, and in particular to an animal target detection method, device, equipment, medium and computer program product. Background Art

[0002] In wildlife conservation and ecological research, accurate animal target detection is crucial for monitoring biodiversity and ecosystem health.

[0003] Existing methods for detecting animal targets include traditional remote sensing technology, radio telemetry technology, and infrared camera technology. Among them, traditional remote sensing technology refers to the use of satellites or aircraft to obtain ground images and identify animals through image analysis. This method has low resolution and is difficult to accurately identify small animals in complex terrain (such as dense vegetation). In addition, the use and maintenance costs of satellites and aircraft are expensive. Radio telemetry technology is to install radio transmitters on animals and track the location of animals by receiving signals. This method requires capturing animals and installing equipment, and regularly replacing batteries or maintaining equipment. It is impossible to cover all animals in a short period of time, and the timeliness and comprehensiveness of detection are limited. Infrared camera technology refers to setting up infrared trigger cameras in the wild to automatically capture images when animals pass by. This method can only monitor animals passing in front of the camera, and the coverage range is not wide enough. In addition, a large amount of image data needs to be manually analyzed, which is time-consuming and labor-intensive.

[0004] It can be seen that the existing animal target detection methods either do not have a wide enough coverage, or the resolution is not high enough, or require high time and economic costs. Therefore, it can be seen that the detection efficiency of the existing animal target detection methods is not high enough. Summary of the invention

[0005] The present invention provides an animal target detection method, device, equipment, medium and computer program product, which are used to solve the defects of high cost and low resolution of detection using satellites or aircraft or detection blind spots using infrared cameras in the prior art, and realize wide coverage, high precision and low cost animal target detection.

[0006] The present invention provides an animal target detection method, comprising the following steps.

[0007] Obtaining preprocessed radar echo signals; Extracting multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; The multiple features are input into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

[0008] According to an animal target detection method provided by the present invention, before obtaining the pre-processed radar echo signal, the method includes: Acquire radar echo signal; The radar echo signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

[0009] According to an animal target detection method provided by the present invention, the multiple features are extracted from the pre-processed radar echo signal, including: Inputting the preprocessed radar echo signal into an OS-CFAR detector to obtain a time delay; calculating a distance to a target based on the time delay; According to the distance and Doppler map of the target, the speed of the target is obtained; The geometric characteristics of the target are obtained according to the speed of the target and the duration of the target signal.

[0010] According to an animal target detection method provided by the present invention, before filtering the radar echo signal using a preset range gate to obtain the preprocessed radar echo signal, the method further includes: Filtering the radar echo signal using an elliptical bandpass filter to obtain a first filtered signal; Decomposing the first filtered signal using a three-level wavelet packet to obtain a decomposed signal; The using a preset range gate to filter the radar echo signal to obtain the preprocessed radar echo signal includes: The decomposed signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

[0011] According to an animal target detection method provided by the present invention, after decomposing the first filtered signal using a three-level wavelet packet to obtain a decomposed signal, the method further includes: Compensating the decomposed signal for distance attenuation by using time-varying gain control to obtain a compensated signal; The using a preset range gate to filter the radar echo signal to obtain the preprocessed radar echo signal includes: The compensated signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

[0012] According to an animal target detection method provided by the present invention, the bands of the radar echo signal include P band, X band and L band.

[0013] The present invention also provides an animal target detection device, comprising the following modules: A signal acquisition module, used to acquire the pre-processed radar echo signal; A feature extraction module, used to extract multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; The target classification module is used to input the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the animal target detection methods described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the animal target detection methods described above.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the animal target detection method as described above is implemented.

[0017] The animal target detection method, device, equipment, medium and computer program product provided by the present invention obtain pre-processed radar echo signals; extract multiple features from the pre-processed radar echo signals; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; input the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classification for the multiple features. The present application uses radar detection technology to detect animal targets, replacing traditional satellite, aircraft and infrared camera detection, reducing costs, and is easier to maintain than infrared cameras, without any damage to animals, reducing the economic cost and time cost of ecological detection, and improving the efficiency of animal target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is one of the flow charts of the animal target detection method provided by the present invention.

[0020] Figure 2 This is the second flow chart of the animal target detection method provided by the present invention.

[0021] Figure 3 It is a structural schematic diagram of the animal target detection device provided by the present invention.

[0022] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Combine the following Figure 1-Figure 4 Specific embodiments of the present invention are described.

[0025] Figure 1 FIG. 1 is one of the flow charts of the animal target detection method provided by the present invention, such as Figure 1 As shown, the method comprises the following steps: Step 101, obtaining a preprocessed radar echo signal; Among them, the radar echo signal refers to the signal reflected back by the electromagnetic wave emitted by the radar after encountering the target. This signal carries information such as the position, speed, shape, etc. of the target, and is the basis for the radar system to detect targets and estimate parameters.

[0026] Specifically, in this application, a mobile radar is used to detect while moving in the target area. A mobile radar refers to a radar system with mobility or portability, such as a vehicle, unmanned vehicle, or drone equipped with a radar system to move in the target area, transmit radar signals, use a phased array antenna or a flat antenna and a digital signal processor, adjust the pulse signal beam direction in real time through a phase control matrix, use a GPS positioning system and an inertial navigation module to provide real-time coordinate information, and drive a microwave switch array to update the radar signal coverage area. When the radar reflection signal encounters an obstacle (such as trees, animals, etc.), it will be reflected back and received by the radar receiver.

[0027] The radar system performs signal preprocessing on the radar echo signal. The preprocessing steps include signal amplification, filtering, analog-to-digital conversion, denoising and interference suppression to improve the signal-to-noise ratio and signal clarity, and finally obtain the preprocessed radar echo signal.

[0028] It is worth mentioning that the mobile radar in this application uses the transmission signals of three radar bands: P-band, X-band and L-band, which are suitable for the detection of biological characteristics of animal targets. Among them, the P-band is a frequency band in the radio spectrum, with a frequency range of 230MHz~1000MHz and a wavelength of 30cm~130cm. Due to its long wavelength, it has a strong penetration ability and can penetrate vegetation and a certain depth of the surface layer, which is suitable for underground detection and forest monitoring; the resolution of the P-band is low, but its coverage range is large. The frequency range of the X-band is 8GHz~12GHz, and the wavelength is between 2.5cm~3.75cm. It has high resolution and good directivity, is less affected by atmospheric absorption and rain attenuation, and is suitable for high-resolution imaging and target detection. The frequency range of L band is 1GHz~2GHz, and the wavelength is between 15cm~30cm. It has good penetration ability and moderate resolution. It can penetrate clouds, vegetation and a certain depth of the ground. It is suitable for underground detection and forest monitoring. It is less affected by atmospheric absorption and rain attenuation. It is suitable for high-precision measurement, especially in applications that need to penetrate vegetation and the ground. In practical applications, different bands can be selected according to factors such as terrain, different animal populations to be detected, and weather changes.

[0029] Step 102, extracting multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization features and Doppler spectrum moments; Among them, micro-Doppler characteristics refer to the subtle changes in Doppler frequency shift caused by small movements of the target or on the target. The principle is: when the target moves radially relative to the radar, the frequency of the radar echo signal will shift, which is called Doppler frequency shift; small movements on the target (such as the gait of the moving target, vibrating blades, etc.) will produce additional frequency modulation on the basis of the main Doppler frequency shift, forming micro-Doppler characteristics. Geometric characteristics refer to the aspect ratio of the target; polarization characteristics refer to the vibration direction and the change characteristics of the vibration direction presented by the radar transmission wave after encountering different targets under the action of different radar electromagnetic wave polarization modes. The state of the target can be identified through polarization characteristics, such as distinguishing between stationary targets and moving targets, and different types of targets can also be classified, such as distinguishing different types of animals; Doppler spectrum moments refer to a series of statistics obtained by analyzing and calculating the Doppler spectrum, such as the zero-order moment (total power), the first-order moment (average speed), the second-order moment (spectral width), the third-order moment (skewness), the fourth-order moment (kurtosis), etc. These statistics are used to describe the motion characteristics of the target. These moments are widely used in the radar field to extract information such as target speed and spectrum width.

[0030] Specifically, for micro-Doppler features, we can extract micro-Doppler features by performing time-frequency analysis (such as short-time Fourier transform, wavelet transform, etc.) on the preprocessed radar echo signal to observe the change of frequency over time. We can also extract the frequency components of micro-Doppler features by performing spectral analysis on the preprocessed radar echo signal to identify the motion pattern of the target. We can also use convolutional neural networks (CNN) to extract key parameters about micro-Doppler features, such as modulation frequency, modulation depth, period, etc., for target recognition and classification.

[0031] For geometric features, a pre-trained neural network model can be used to extract the geometric features in the signal, that is, the radar image (the image generated by the preprocessed radar echo signal) is input into the pre-trained neural network model to obtain the geometric features of the animal target in the image (such as the aspect ratio).

[0032] By using radar signals with different polarization modes, the polarization characteristics of the target can be detected.

[0033] For the preprocessed radar echo signal, the corresponding range gate signal is selected for processing according to the distance of the target, and the selected range gate signal is subjected to fast Fourier transform to convert the time domain signal into a frequency domain signal to obtain the Doppler spectrum. The frequency components therein are analyzed to obtain the micro-Doppler characteristics and Doppler spectrum moment.

[0034] Step 103: input the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

[0035] Among them, the random forest model constructs multiple decision trees to predict different features, and finally combines the prediction results of these decision trees to improve the accuracy of the model.

[0036] Specifically, the micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments of the above radar images are input into a pre-built random forest model to obtain the animal target category output by the model.

[0037] Optionally, the distance and speed of the animal target may be calculated based on the above features, and finally the category, distance, speed and movement trajectory of the animal target may be output to a display screen.

[0038] The above embodiment obtains the preprocessed radar echo signal; extracts multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; the multiple features are input into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classification of the multiple features. The present application uses radar detection technology to detect animal targets, replacing traditional satellite, aircraft, and infrared camera detection, reducing costs, and is easier to maintain than infrared cameras, without any damage to animals, reducing the economic cost and time cost of ecological detection, and improving the efficiency of animal target detection.

[0039] In one embodiment, the above step 101 includes: acquiring a radar echo signal; and filtering the radar echo signal using a preset range gate to obtain the preprocessed radar echo signal.

[0040] Among them, the range gate is a virtual "time window" or "distance window" that only allows signals within a specific time period (corresponding to a specific distance) to pass through, and other signals are blocked. For example, in radar, after the electromagnetic wave is emitted, the time delay of the echo is proportional to the target distance. The range gate selects target signals within a specific distance by setting the starting point and width of the time window.

[0041] Specifically, the radar locks the distance information of specific animal targets through the range gate, filters out clutter (such as clouds, flying birds) and other interference, and improves tracking accuracy. Use range gate control to adjust the signal range of the radar echo signal to ensure that the echo only includes signals within the operating frequency of the animal detection radar.

[0042] Optionally, before using the range gate control, it also includes: using an elliptical bandpass filter to filter the radar echo signal to obtain a first filtered signal; using a three-level wavelet packet to decompose the first filtered signal to obtain a decomposed signal; using time-varying gain control to compensate the decomposed signal for distance attenuation to obtain a compensated signal.

[0043] To elaborate, an elliptical bandpass filter (200MHz-4GHz) is used for bandpass filtering, a three-level wavelet packet decomposition is used to achieve multi-resolution noise reduction, and time-varying gain control (0-80dB) is used to compensate for distance attenuation. Then, a range gate control is used to adjust the signal range of the radar echo signal to ensure that the echo only includes signals within the operating frequency of the animal detection radar.

[0044] The above embodiment can remove interference and implement multi-resolution noise reduction through range gate control, which is beneficial to subsequent signal analysis.

[0045] In one embodiment, the above step 102 includes: inputting the preprocessed radar echo signal into an OS-CFAR detector to obtain a time delay; calculating the distance of the target based on the time delay; obtaining the speed of the target according to the distance and Doppler map of the target; and obtaining the geometric characteristics of the target according to the speed of the target and the duration of the target signal.

[0046] Among them, the ordered statistics constant false alarm rate (OS-CFAR) detector is mainly used for target detection. Its core function is to maintain a constant false alarm rate in a clutter and noise environment, that is, by sorting reference units and selecting thresholds, a constant false alarm probability is maintained, thereby effectively detecting targets in clutter and noise backgrounds. In this application, the OS-CFAR detector can also be used to achieve feature extraction.

[0047] Specifically include: Figure 2 As shown, Figure 2 The second flowchart of the animal target detection method is shown. The preprocessed radar echo signal is a time domain signal and is segmented according to the radar pulse repetition period. Each segment corresponds to an echo within a detection period. The preprocessed radar echo signal is input into the OS-CFAR detector, the power value of each sampling point in the reference unit is calculated, the power values ​​of the N reference units are arranged in ascending order, and the kth value is selected as the noise estimate. If the power detected by the detection unit in the OS-CFAR detector is greater than the preset power threshold, it is determined that there is a target at the sampling point and its time delay is recorded. Traverse all sampling points to generate a target time delay set {t1, t2, ..., tM}, where M is the number of detected targets.

[0048] By detecting the time delay of the target, the distance of the target can be calculated, and the straight-line distance R between the target and the radar can be calculated according to the time delay t: R=2c t, where c is the speed of light and the factor 2 accounts for the round-trip path of the signal.

[0049] By combining the distance to the target with the Doppler effect (i.e., the Doppler map), the target's speed can be estimated, and by analyzing the duration and amplitude changes of the target signal, the size of the target (i.e., geometric characteristics) can be estimated.

[0050] In the above embodiment, the geometric features of the target are extracted by the OS-CFAR detector, thereby providing effective data basis for subsequent target category recognition.

[0051] In another embodiment, after initial target detection based on the CA-CFAR algorithm, a multi-pass tracking gate technology is used in combination with an extended Kalman filter to predict target tracking parameters; and time domain correlation matching filtering is used, pulse cancellation MTI (two-point cancellation, weight coefficient [-1,1]) is used to suppress stationary clutter, and adaptive clutter map processing is combined to improve blind speed, and a Chebyshev high-pass filter (cut-off frequency 1Hz) is used to filter out remaining clutter.

[0052] Specifically, the multi-pass tracking gate technology is combined with the extended Kalman filter (EKF) to predict the target trajectory with a nonlinear motion model. The multi-frame detection results are associated through the tracking gate, and the state estimation is iteratively corrected using the observed data to achieve continuous tracking of the target track and false alarm suppression.

[0053] In the above embodiment, when multiple targets exist at the same time, it is necessary to define a tracking area (wave gate) for each target independently to avoid signal confusion. The multi-path tracking gate technology dynamically sets multiple tracking gates (Tracking Gate) and combines signal characteristics to separate multipath signals or multi-target echoes, thereby improving the robustness of detection and tracking.

[0054] The animal target detection device provided by the present invention is described below. The animal target detection device described below and the animal target detection method described above can be referenced to each other.

[0055] like Figure 3 As shown, Figure 3 The schematic diagram of the structure of the animal target detection device is shown, which includes the following modules: The signal acquisition module 301 is used to acquire the pre-processed radar echo signal; A feature extraction module 302 is used to extract multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; The target classification module 303 is used to input the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

[0056] In one embodiment, the animal target detection device further includes a signal preprocessing unit, which is used to: Acquire a radar echo signal; and filter the radar echo signal using a preset range gate to obtain the preprocessed radar echo signal.

[0057] In one embodiment, the feature extraction module 302 is further used to: The preprocessed radar echo signal is input into the OS-CFAR detector to obtain a time delay; the distance of the target is calculated based on the time delay; the speed of the target is obtained according to the distance and Doppler map of the target; and the geometric characteristics of the target are obtained according to the speed of the target and the duration of the target signal.

[0058] In one embodiment, the signal preprocessing unit is further used for: The radar echo signal is filtered using an elliptical bandpass filter to obtain a first filtered signal; the first filtered signal is decomposed using a three-level wavelet packet to obtain a decomposed signal; the decomposed signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

[0059] In one embodiment, the signal preprocessing unit is further used for: The decomposed signal is compensated for distance attenuation by using time-varying gain control to obtain a compensated signal; the compensated signal is filtered by using a preset range gate to obtain the preprocessed radar echo signal.

[0060] In one embodiment, the bands of the radar echo signal include a P band, an X band, and an L band.

[0061] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the animal target detection method, which includes: obtaining a pre-processed radar echo signal; extracting a plurality of features from the pre-processed radar echo signal; the plurality of features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; inputting the plurality of features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classification of the plurality of features.

[0062] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0063] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the animal target detection method provided by the above-mentioned methods, which includes: obtaining a preprocessed radar echo signal; extracting multiple features from the preprocessed radar echo signal; the multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; inputting the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

[0064] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the animal target detection method provided by the above-mentioned methods, the method comprising: obtaining a preprocessed radar echo signal; extracting a plurality of features from the preprocessed radar echo signal; the plurality of features comprising: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; inputting the plurality of features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the plurality of features.

[0065] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0066] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting an animal target, characterized in that: include: Obtaining preprocessed radar echo signals; extracting a plurality of features from the preprocessed radar echo signal; The multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; The multiple features are input into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

2. The animal target detection method according to claim 1, characterized in that: Before obtaining the preprocessed radar echo signal, the method includes: Acquire radar echo signal; The radar echo signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

3. The animal target detection method according to claim 1, characterized in that: The extracting multiple features from the preprocessed radar echo signal includes: Inputting the preprocessed radar echo signal into an OS-CFAR detector to obtain a time delay; calculating a distance to a target based on the time delay; According to the distance and Doppler map of the target, the speed of the target is obtained; The geometric characteristics of the target are obtained according to the speed of the target and the duration of the target signal.

4. The animal target detection method according to claim 2, characterized in that: Before filtering the radar echo signal by using a preset range gate to obtain the preprocessed radar echo signal, the method further includes: Filtering the radar echo signal using an elliptical bandpass filter to obtain a first filtered signal; Decomposing the first filtered signal using a three-level wavelet packet to obtain a decomposed signal; The using a preset range gate to filter the radar echo signal to obtain the preprocessed radar echo signal includes: The decomposed signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

5. The animal target detection method according to claim 4, characterized in that: After decomposing the first filtered signal using a three-level wavelet packet to obtain a decomposed signal, the method further includes: Compensating the decomposed signal for distance attenuation by using time-varying gain control to obtain a compensated signal; The using a preset range gate to filter the radar echo signal to obtain the preprocessed radar echo signal includes: The compensated signal is filtered using a preset range gate to obtain the preprocessed radar echo signal.

6. The animal target detection method according to any one of claims 1 to 5, characterized in that: The bands of the radar echo signal include P band, X band and L band.

7. An animal target detection device, characterized in that: include: A signal acquisition module, used to acquire the pre-processed radar echo signal; A feature extraction module, used to extract multiple features from the preprocessed radar echo signal; The multiple features include: micro-Doppler features, geometric features, polarization characteristics and Doppler spectrum moments; The target classification module is used to input the multiple features into a pre-constructed random forest model to obtain the animal target category output by the pre-constructed random forest model after classifying the multiple features.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the animal target detection method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the animal target detection method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the animal target detection method according to any one of claims 1 to 6 is implemented.

Images