Animal voiceprint monitoring method and device, medium and product

Abstract

The invention relates to the technical field of animal identity recognition, in particular to an animal voiceprint monitoring method and device, a medium and a product, and the method comprises the steps: carrying out the sound source orientation of an activity sound when the existence of the activity sound is detected, controlling a sound collection device to face the direction where the activity sound is emitted, and carrying out the audio collection. Then, the audio data is preprocessed, voiceprint feature extraction is carried out based on the target audio data, a multi-dimensional voiceprint feature vector is determined, and the multi-dimensional voiceprint feature vector can comprehensively and meticulously display multi-aspect sound characteristics of the target audio data. And finally, matching with each voiceprint model in an animal voiceprint model library is carried out based on the multi-dimensional voiceprint feature vector, and animal information is determined. Through sound source positioning, the audio data is less interfered by background noise and irrelevant sound, the quality of the audio data is remarkably improved by executing the preprocessing operation, animal voiceprint monitoring is executed by using the high-quality target audio data, and the accuracy of animal voiceprint monitoring is improved.

Description

Technical Field

[0001] The present application relates to the technical field of animal identification, and in particular to an animal voiceprint monitoring method, device, medium and product. Background Art

[0002] In the field of ecology and wildlife conservation, monitoring animal behavior and population dynamics is crucial. Traditional monitoring methods, such as direct observation, trapping, and radio tracking, although providing valuable data to a certain extent, often have limitations such as high cost, low efficiency, and great interference with animals.

[0003] With the advancement of science and technology, especially the rapid development of audio processing technology and machine learning algorithms, animal voiceprint monitoring, as a non-invasive, efficient and relatively low-cost monitoring method, has gradually attracted the attention of researchers and conservation agencies. However, when collecting audio data, the sound collection equipment cannot accurately align with the direction of the sound source, resulting in the collected audio data containing a large amount of background noise and irrelevant sounds. The complex and changeable background noise and irrelevant sounds will seriously interfere with the collection and identification of animal voiceprints, reducing the accuracy of animal voiceprint monitoring.

[0004] Therefore, how to improve the accuracy of animal voiceprint monitoring is a technical problem that needs to be urgently solved by those skilled in the art. Summary of the invention

[0005] The purpose of this application is to provide an animal voiceprint monitoring method, device, medium and product to solve at least one of the above technical problems.

[0006] The above invention objectives of the present application are achieved through the following technical solutions: In the first aspect, the present application provides an animal voiceprint monitoring method, which adopts the following technical solution: An animal voiceprint monitoring method, comprising: When active sound is detected in the monitoring area, the sound source of the active sound is oriented, the sound collection device is controlled to collect the sound in the direction of the active sound and the audio is collected; Acquire the audio data in the monitoring area collected by the sound collection device, and pre-process the audio data using an audio processing algorithm to obtain target audio data, wherein the pre-processing includes: denoising, filtering, audio segmentation, and mono separation; Extract voiceprint features based on the target audio data to determine a multidimensional voiceprint feature vector, wherein the multidimensional voiceprint feature vector is a multidimensional vector composed of frequency features, amplitude features, time domain features, and frequency domain features; An animal voiceprint model library is obtained, and animal information is determined by matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library, wherein the animal information includes: animal species and animal status.

[0007] By adopting the above technical solution, in order to collect purer audio data that is less disturbed by background noise and irrelevant sounds, when active sounds are detected in the monitoring area, the active sounds are oriented to the sound source, the sound collection direction of the sound collection device is controlled to be the direction of the active sound and audio collection is performed, and the sound source direction is accurately aligned, so as to effectively capture the sound of animals in a wide monitoring area. Then, the audio data in the monitoring area collected by the sound collection device is obtained, and the audio processing algorithm is used to pre-process the audio data to obtain the target audio data. The pre-processing is used to significantly improve the quality of the audio data and enhance the accuracy and processing efficiency of voiceprint recognition. Then, voiceprint feature extraction is performed based on the target audio data to determine a multi-dimensional voiceprint feature vector. The multi-dimensional voiceprint feature vector can comprehensively and meticulously display the multi-faceted sound characteristics of the target audio data, which helps to improve the accuracy and robustness of subsequent voiceprint model matching. Finally, the multi-dimensional voiceprint feature vector is matched with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information. Audio collection is performed through sound source positioning, making the audio data less affected by background noise and irrelevant sounds, and preprocessing operations are performed to significantly improve the quality of the audio data. Animal voiceprint monitoring is performed using high-quality target audio data, thereby improving the accuracy of animal voiceprint monitoring.

[0008] In a preferred example, the present application may be further configured as follows: the audio data is preprocessed using an audio processing algorithm to obtain target audio data, including: Performing denoising and filtering processing on the audio data to obtain first audio data, wherein the denoising and filtering processing is used to remove complex and changeable background noise in the audio data; Performing mono separation based on the first audio data to determine a target source audio signal; The continuous target source audio signal is segmented using an audio segmentation technique to obtain target audio data.

[0009] In a preferred example, the present application may be further configured as follows: after matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information, the configuration further includes: Controlling the shooting direction of the image acquisition device to the direction where the activity sound is emitted and taking images to obtain animal images; Verify the monitoring result based on the animal image and the animal information, and determine the monitoring verification result; When the monitoring verification result is verification passed, data association is performed based on the animal image, the target audio data and the animal information, and stored in an animal information database.

[0010] In a preferred example, the present application may be further configured as follows: after verifying the monitoring result based on the animal image and the animal information and determining the monitoring verification result, the following may also be included: When the monitoring verification result is a verification failure, data association is performed based on the animal image and the target audio data, and the data is stored in the animal information database; In the animal information database, the data to be confirmed are data-marked, wherein the data marking facilitates manual animal voiceprint recognition of the data to be confirmed, and the data to be confirmed are the animal image that failed verification, the audio data, and the target audio data; An animal voiceprint confirmation instruction carrying target animal information is obtained, and based on the animal voiceprint confirmation instruction, data to be confirmed in the animal information database is filled with data, wherein the data filling is used to enrich the animal information database information.

[0011] In a preferred example, the present application can be further configured as follows: The animal information corresponding to different areas is obtained, and multi-dimensional analysis is performed based on the animal information corresponding to each area to obtain animal activity analysis results, wherein the animal activity analysis results include: population distribution information, living habit information and activity range information.

[0012] In a preferred example, the present application can be further configured as follows: When no active sound is detected in the monitoring area, the sound collecting device is controlled to enter a dormant state, so that the sound collecting device can save energy and extend its service life.

[0013] In a second aspect, the present application provides an electronic device, which adopts the following technical solution: at least one processor; Memory; At least one application, wherein the at least one application is stored in the memory and is configured to be executed by at least one processor, and the at least one application is configured to: execute the above-mentioned animal voiceprint monitoring method.

[0014] In a third aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a computer program, which, when executed in a computer, causes the computer to execute the above-mentioned animal voiceprint monitoring method.

[0015] In a fourth aspect, the present application provides a computer program product, which adopts the following technical solution: A computer program product includes a computer program, and when the computer program is executed by a processor, the animal voiceprint monitoring method is implemented.

[0016] In summary, the present application includes at least one of the following beneficial technical effects: In order to collect purer audio data that is less affected by background noise and irrelevant sounds, when active sounds are detected in the monitoring area, the active sounds are oriented to the sound source, the sound collection device is controlled to collect audio in the direction of the active sounds, and the sound source is accurately aligned to the direction of the sound source, so as to effectively capture the sounds of animals in a wide monitoring area. Then, the audio data in the monitoring area collected by the sound collection device is obtained, and the audio processing algorithm is used to pre-process the audio data to obtain the target audio data. The pre-processing is used to significantly improve the quality of the audio data and enhance the accuracy and processing efficiency of voiceprint recognition. Then, voiceprint feature extraction is performed based on the target audio data to determine the multi-dimensional voiceprint feature vector, which can comprehensively and meticulously display the multi-faceted sound characteristics of the target audio data, and help improve the accuracy and robustness of subsequent voiceprint model matching. Finally, the multi-dimensional voiceprint feature vector is matched with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information. Audio collection is performed through sound source positioning, making the audio data less affected by background noise and irrelevant sounds, and preprocessing operations are performed to significantly improve the quality of the audio data. Animal voiceprint monitoring is performed using high-quality target audio data, thereby improving the accuracy of animal voiceprint monitoring.

[0017] The image acquisition device is controlled to shoot in the direction of the active sound and take images to obtain animal images. Then, the monitoring result is verified based on the animal image and animal information to determine the monitoring verification result. When the monitoring verification result is verified, data association is performed based on the animal image, target audio data and animal information, and stored in the animal information database. By combining the two technologies of voiceprint monitoring and image recognition to achieve dual verification of animal identity, the combination of the two technologies can significantly improve the accuracy of animal monitoring and reduce misjudgments and missed judgments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flowchart of an animal voiceprint monitoring method according to one embodiment of the present application; Figure 2 It is a structural schematic diagram of an animal voiceprint monitoring device according to one embodiment of the present application; Figure 3It is a structural schematic diagram of an electronic device according to one embodiment of the present application. DETAILED DESCRIPTION

[0019] The following combination Figures 1 to 3 This application is described in further detail.

[0020] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, a person skilled in the art may make non-creative modifications to the present embodiment as needed, but such modifications are protected by the patent law as long as they are within the scope of the present application.

[0021] In order to make the purpose, technical scheme and advantages of the embodiment of the present application clearer, the technical scheme in the embodiment of the present application will be clearly and completely described in conjunction with the drawings in the embodiment of the present application. Obviously, the described embodiment is a part of the embodiment of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application. It should be noted that in the optional embodiments of the present application, the object information and other related data involved, when the embodiments in the present application are applied to specific products or technologies, need to obtain the permission or consent of the object, and the collection, use and processing of the relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions. In other words, if the data related to the object is involved in the embodiment of the present application, it needs to be obtained through the authorization and consent of the object, the authorization and consent of the relevant departments, and in accordance with the relevant laws, regulations and standards of the country and region. If personal information is involved in the embodiment, the acquisition of all personal information needs to obtain the consent of the individual. If sensitive information is involved, the separate consent of the information subject needs to be obtained, and the embodiment also needs to be implemented with the authorization and consent of the object.

[0022] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0023] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0024] The embodiment of the present application provides an animal voiceprint monitoring method, which is executed by an electronic device, which can be a server or a terminal device, wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The terminal device can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited to this. The terminal device and the server can be directly or indirectly connected via wired or wireless communication, and the embodiment of the present application does not limit this. Figure 1 As shown, the method includes step S101, step S102, step S103 and step S104, wherein: Step S101: when an active sound is detected in the monitoring area, the active sound is oriented to the sound source, the sound collection device is controlled to collect the direction of the active sound and the audio is collected.

[0025] For the embodiments of the present application, in the fields of wildlife protection and ecological monitoring, the demand for monitoring animal voiceprints is getting higher and higher. However, the audio data collected by animal voiceprint monitoring in the related art often contains a large amount of background noise and irrelevant sounds. The complex and changeable background noise and irrelevant sounds will seriously interfere with the collection and identification of animal voiceprints. Therefore, in order to collect purer audio data that is less interfered by background noise and irrelevant sounds, the sound source localization operation is performed to enable the sound collection device to be more accurately aligned with the direction of the sound source, so as to effectively capture the sound of animals in a wide monitoring area. Therefore, through sound signal processing algorithms, such as energy detection, short-time Fourier transform, etc., it is detected whether there is active sound in the monitoring area; when the presence of active sound is detected, the sound source can be preliminarily located using algorithms such as arrival time difference or arrival phase difference. Then, based on the preliminary positioning, a more advanced sound source localization algorithm is used, such as a multiple signal classification algorithm, a minimum variance distortionless response beamformer, etc., to accurately estimate the direction of the sound source and determine the precise sound source directional information. Then, based on the precise sound source directional information, the sound collection device is controlled to collect the direction of the sound, so that the sound collection device is aimed at the direction of the active sound. Furthermore, according to factors such as the distance of the sound source and the strength of the sound signal, the gain, sampling rate and other parameters of the sound collection device are automatically adjusted to ensure the best quality of the collected audio data.

[0026] Step S102: Acquire the audio data in the monitoring area collected by the sound collection device, and pre-process the audio data using an audio processing algorithm to obtain target audio data, wherein the pre-processing includes: denoising, filtering, audio segmentation, and mono separation.

[0027] For the embodiment of the present application, the audio data collected by the sound collection device in the monitoring area is obtained, and the audio data includes some noise, and there are long audio data with animal sounds and no animal sounds alternating. The unprocessed audio data will affect the accuracy and processing complexity of the subsequent animal voiceprint recognition. In order to significantly improve the quality of audio data, enhance the accuracy and processing efficiency of voiceprint recognition, the audio data is preprocessed by an audio processing algorithm to obtain target audio data, wherein the preprocessing includes: denoising, filtering, audio segmentation, and mono separation. For the denoising and filtering operations in the preprocessing, the background noise and irrelevant sounds in the audio data can be reduced and eliminated, and the audio signal can be further smoothed by removing high-frequency or low-frequency noise to improve the accuracy of subsequent voiceprint recognition. For audio segmentation in the preprocessing, the continuous long audio data is cut to obtain a short audio clip with rich voiceprint features, so as to facilitate the subsequent voiceprint feature extraction and voiceprint model matching. For mono separation in the preprocessing, it is used to separate at least one source signal from the audio data to avoid mutual interference between animal sounds of different species.

[0028] There are many ways to pre-process audio data, which are not limited in the embodiments of the present application. In one feasible method, denoising and filtering are performed based on the audio data to obtain first audio data, wherein the denoising and filtering are used to remove complex and changeable background noise in the audio data; monaural separation is performed based on the first audio data to determine at least one source audio signal, and audio signal screening is performed based on each source audio signal to determine a target source audio signal; and audio segmentation is performed on the continuous target source audio signal using audio segmentation technology to obtain target audio data.

[0029] Step S103: extract voiceprint features based on the target audio data to determine a multi-dimensional voiceprint feature vector, wherein the multi-dimensional voiceprint feature vector is a multi-dimensional vector composed of frequency features, amplitude features, time domain features and frequency domain features.

[0030] For the embodiment of the present application, voiceprint feature extraction is performed based on the target audio data to determine a multi-dimensional voiceprint feature vector, which can comprehensively and meticulously display the various aspects of the sound characteristics of the target audio data, and help improve the accuracy and robustness of the subsequent voiceprint model matching. The specific implementation process for voiceprint feature extraction is as follows: frequency feature extraction is performed based on the target audio data to determine the frequency feature, that is, the fundamental frequency information is extracted by algorithms such as the autocorrelation method and the cepstrum method, and the sound signal of the target audio data is Fourier transformed to obtain its representation in the frequency domain, recorded as the spectrum, and the fundamental frequency information and the spectrum are integrated to obtain the frequency feature corresponding to the target audio data. Amplitude feature extraction is performed based on the target audio data to determine the amplitude feature, that is, the energy feature is extracted by calculating the energy or root mean square value of the sound signal, and the envelope feature is extracted by methods such as the Hilbert transform, and the energy feature and the envelope feature are integrated to obtain the amplitude feature corresponding to the target audio data, wherein the energy feature is the square of the amplitude; the envelope feature of the sound signal reflects the change of the amplitude of the sound signal over time. Based on the target audio data, time domain features are extracted to determine the time domain features, that is, the pitch period information is extracted by the autocorrelation method or the cross-correlation method, and the intonation features are extracted by calculating the rate of change of the fundamental frequency or using methods such as dynamic time warping, and the pitch period information and intonation features are integrated to obtain the time domain features corresponding to the target audio data, wherein the pitch period refers to the length of the pitch vibration period in the sound waveform, which has strong individual differences; intonation refers to the pitch change of the sound signal, and the intonation of different animals is different. Based on the target audio data, frequency domain features are extracted to determine the frequency domain features, that is, by mapping the spectrum to the Mel frequency scale and then performing a logarithmic transformation to extract the Mel spectrum features; by performing Fourier transform, logarithmic transform and inverse Fourier transform on the sound signal, the cepstrum features are extracted, and the Mel spectrum features and cepstrum features are integrated to obtain the frequency domain features, wherein the cepstrum is a method of converting the frequency domain representation of the sound signal into the cepstrum representation, which can further reveal the spectrum structure of the sound signal. Furthermore, the frequency features, amplitude features, time domain features and frequency domain features extracted above are combined to form a multi-dimensional voiceprint feature vector. The multi-dimensional voiceprint feature vector contains the information of the sound signal in multiple dimensions, which can comprehensively reflect the characteristics and individual differences of the sound signal, and help improve the accuracy and robustness of subsequent voiceprint model matching.

[0031] Step S104: obtaining an animal voiceprint model library, matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library, and determining the animal information, wherein the animal information includes: animal species and animal status.

[0032] For the embodiment of the present application, an animal voiceprint model library is pre-stored in the electronic device, and the animal voiceprint model library stores voiceprint models corresponding to each species, and the voiceprint model includes the statistical characteristics and classification boundaries of the voiceprint features of the species. Therefore, based on the multi-dimensional voiceprint feature vector and the voiceprint model corresponding to each species in the animal voiceprint model library, the species and activity state corresponding to the successfully matched voiceprint model are recorded as animal information, and the animal information includes: animal species and animal state, wherein the animal state includes but is not limited to: health state, behavior pattern (for example, courtship, territory competition, warning, foraging), emotional state, etc. The training method for the voiceprint model includes: collecting animal sound samples from multiple channels to ensure that the animal sound samples can cover animals of different species, ages, genders and health conditions, and finely annotating the collected animal sound samples to clarify the animal species, health state, behavior pattern and emotional state corresponding to each animal sound sample. The fine annotation process usually requires professionals (for example, animal behaviorists, veterinarians) to participate in the annotation. Then, multi-dimensional voiceprint feature extraction is performed based on the animal sound sample to determine the sample voiceprint feature vector corresponding to the animal sound sample. Then, a suitable machine learning or deep learning model is selected, and the finely labeled animal sound samples and the extracted sample voiceprint feature vectors are used as training set data. The selected target model is trained using the training set data until the generalization ability of the target model meets the requirements to obtain a voiceprint model.

[0033] It can be seen that in the embodiment of the present application, in order to collect purer audio data that is less disturbed by background noise and irrelevant sounds, when active sounds are detected in the monitoring area, the active sounds are oriented to the sound source, the sound collection direction of the sound collection device is controlled to be the direction of the active sound and audio collection is performed, and the sound source direction is accurately aligned, so as to effectively capture the sound of animals in a wide monitoring area. Then, the audio data in the monitoring area collected by the sound collection device is obtained, and the audio processing algorithm is used to pre-process the audio data to obtain the target audio data. The pre-processing is used to significantly improve the quality of the audio data and enhance the accuracy and processing efficiency of voiceprint recognition. Furthermore, voiceprint feature extraction is performed based on the target audio data to determine a multi-dimensional voiceprint feature vector. The multi-dimensional voiceprint feature vector can comprehensively and meticulously display the multi-faceted sound characteristics of the target audio data, which helps to improve the accuracy and robustness of subsequent voiceprint model matching. Finally, the multi-dimensional voiceprint feature vector is matched with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information. Audio collection is performed through sound source positioning, making the audio data less affected by background noise and irrelevant sounds, and preprocessing operations are performed to significantly improve the quality of the audio data. Animal voiceprint monitoring is performed using high-quality target audio data, thereby improving the accuracy of animal voiceprint monitoring.

[0034] Furthermore, in order to significantly improve the quality of audio data and enhance the accuracy and processing efficiency of voiceprint recognition, in an embodiment of the present application, an audio processing algorithm is used to pre-process the audio data to obtain target audio data, including: Performing denoising and filtering processing on the audio data to obtain first audio data, wherein the denoising and filtering processing is used to remove complex and changeable background noise in the audio data; Performing mono separation based on the first audio data to determine a target source audio signal; The audio segmentation technology is used to segment the continuous target source audio signal to obtain the target audio data.

[0035] For the embodiment of the present application, in order to significantly improve the quality of audio data, enhance the accuracy and processing efficiency of voiceprint recognition, the audio data is preprocessed using an audio processing algorithm to obtain target audio data. Specifically, denoising and filtering are performed based on the audio data to obtain the first audio data, wherein the denoising and filtering are used to remove the complex and changeable background noise in the audio data, and further smooth the audio signal by removing high-frequency or low-frequency noise to improve the accuracy of subsequent voiceprint recognition. The audio data may include the calls of multiple animals, and different sound signals will be superimposed on each other to form a complex mixed signal. However, if the complex mixed signal is directly used for voiceprint recognition, there will be mutual interference between the sounds of different animals, resulting in a decrease in the accuracy of voiceprint recognition. Therefore, based on the first audio data, mono separation is performed to determine the target source audio signal, wherein the target source audio signal is composed of at least one single audio signal, and performing mono separation avoids mutual interference between the sounds of animals of different species. There are many ways to implement monaural separation, which are not limited in the embodiments of the present application. In one feasible way, blind source separation is used to perform monaural separation on the first audio data. Blind source separation is a technique for recovering the source signal from the mixed signal without knowing the source signal and the mixing process. Common blind source separation algorithms include independent component analysis and non-negative matrix decomposition. Since the target audio signal is a continuous and long audio data, there may be silent audio segments or audio segments with fewer voiceprint features in the middle, and these audio segments contribute less to voiceprint recognition. If the continuous target source audio signal is directly used for animal voiceprint recognition, there will be a problem of wasting computing resources and increasing the overall computing cost and time. Therefore, the continuous target source audio signal is segmented using audio segmentation technology to obtain target audio data, wherein the target audio data is an audio segment with rich voiceprint features from a single sound source.

[0036] It can be seen that in the embodiment of the present application, in order to significantly improve the quality of audio data and enhance the accuracy and processing efficiency of voiceprint recognition, denoising and filtering are performed based on the audio data to obtain first audio data, and monaural separation is performed based on the first audio data to determine the target source audio signal, and then, audio segmentation technology is used to perform audio segmentation on the continuous target source audio signal to obtain target audio data.

[0037] Furthermore, in order to significantly improve the accuracy of animal monitoring and reduce misjudgments and missed judgments, in the embodiment of the present application, after determining the animal information, the multi-dimensional voiceprint feature vector is matched with the voiceprint model corresponding to each species in the animal voiceprint model library, and the following is also included: Controlling the shooting direction of the image acquisition device to the direction where the activity sound is emitted and taking images to obtain animal images; Verify monitoring results based on animal images and animal information, and determine monitoring verification results; When the monitoring verification result is verification passed, data association is performed based on the animal image, target audio data and animal information, and stored in the animal information database.

[0038] For the embodiment of the present application, after detecting the presence of active sounds in the monitoring area and performing voiceprint extraction and recognition, the image acquisition device is controlled to turn to the sound source location and take images. The captured animal images can intuitively verify whether the animal information corresponding to the voiceprint monitoring is accurate. By combining the two technologies of voiceprint monitoring and image recognition, double verification of the animal's identity can be achieved. The combination of the two technologies can significantly improve the accuracy of animal monitoring and reduce misjudgments and missed judgments. At the same time, combining audio data and image data and storing them in the animal information database provides an important basis for scientific researchers to more comprehensively analyze animal and population dynamic monitoring.

[0039] Specifically, according to the result of sound positioning, the image acquisition device is controlled to turn to the direction where the active sound is emitted. When the shooting direction of the image acquisition device is adjusted in place, the shooting function is triggered, and then the image acquisition device sends the captured animal image to the electronic device through wireless transmission. Then, feature extraction is performed based on the animal image to determine the animal image features, wherein the animal image features include: color features, texture features, shape features, etc.; the extracted animal image features are matched with the known animal template, and the animal species is determined by calculating the similarity; then, the animal state is determined by analyzing the behavioral features (for example, posture, action, etc.) in the animal image in combination with animal behavior knowledge. Then, the animal species and animal state determined based on the animal image are compared with the animal information determined by the animal voiceprint monitoring. When both information items are matched successfully, the monitoring verification result is determined to be verification passed; otherwise, the monitoring verification result is determined to be verification failed. When the monitoring verification result is verified, data association is performed based on animal images, target audio data and animal information, and stored in the animal information database. Data association not only helps to deeply understand the ecological habits and behavioral patterns of animals, but also provides comprehensive and rich data support for subsequent scientific research analysis, ecological protection and management decisions.

[0040] It can be seen that in the embodiment of the present application, the shooting direction of the image acquisition device is controlled to be the direction where the active sound is emitted and the image is captured to obtain an animal image. Then, the monitoring result is verified based on the animal image and animal information, and the monitoring verification result is determined. When the monitoring verification result is verified, data association is performed based on the animal image, target audio data and animal information, and stored in the animal information database. By combining the two technologies of voiceprint monitoring and image recognition to achieve dual verification of the animal's identity, the combination of the two technologies can significantly improve the accuracy of animal monitoring and reduce misjudgments and missed judgments.

[0041] Furthermore, in order to increase the dimension and depth of the database and make the information in the database more comprehensive and detailed, in the embodiment of the present application, after the monitoring result verification is performed based on the animal image and animal information and the monitoring verification result is determined, it also includes: When the monitoring verification result is a verification failure, data association is performed based on the animal image and the target audio data, and stored in the animal information database; In the animal information database, the data to be confirmed are marked, wherein the data marking facilitates manual animal voiceprint recognition of the data to be confirmed, and the data to be confirmed are the animal images, audio data and target audio data that failed verification; An animal voiceprint confirmation instruction carrying target animal information is obtained, and based on the animal voiceprint confirmation instruction, data to be confirmed is filled in an animal information database, wherein the data filling is used to enrich the animal information database information.

[0042] For the embodiment of the present application, when the monitoring verification result is a verification failure, it indicates that the animal voiceprint model library and the animal information database do not store the relevant information of the animal species in advance, so that the animal information obtained by the animal voiceprint monitoring has a certain deviation. In order to enrich the animal information database and the animal voiceprint model library, data filling is performed by manual confirmation, which not only increases the dimension and depth of the database, but also makes the information in the database more comprehensive and detailed, providing a solid foundation for subsequent animal voiceprint monitoring.

[0043] Specifically, when the monitoring verification result is a verification failure, data association is performed based on the animal image and the target audio data, and stored in the animal information database, wherein the animal information database is designed with a reasonable table structure and fields to facilitate the storage and query of related data. Then, in the animal information database, the data to be confirmed is data-marked, and data marking can be achieved by setting specific fields or tags in the database, so as to highlight the data to be confirmed on the display interface. Furthermore, professionals manually identify the animal image and target audio data displayed in the display interface to determine the animal species and animal condition, and record the manually determined animal species and animal condition as the target animal information, and click or enter the target animal information on the display interface. Obtain an animal voiceprint confirmation instruction carrying the target animal information, and fill the data to be confirmed in the animal information database based on the animal voiceprint confirmation instruction, so as to accurately mark the animal information corresponding to the data to be confirmed.

[0044] It can be seen that in the embodiment of the present application, when the monitoring verification result is a verification failure, data association is performed based on the animal image and the target audio data, and stored in the animal information database. Furthermore, in the animal information database, the data to be confirmed is data-marked, and an animal voiceprint confirmation instruction carrying the target animal information is obtained, and based on the animal voiceprint confirmation instruction, the data to be confirmed in the animal information database is filled with data. In order to enrich the animal information database and the animal voiceprint model library, data filling is performed by manual confirmation, which not only increases the dimension and depth of the database, but also makes the information in the database more comprehensive and detailed, providing a solid foundation for subsequent animal voiceprint monitoring.

[0045] Furthermore, in order to improve the scientific nature of ecological protection decisions and promote biodiversity protection, in the embodiments of the present application, it also includes: The animal information corresponding to different areas is obtained, and multi-dimensional analysis is performed based on the animal information corresponding to each area to obtain the animal activity analysis results, wherein the animal activity analysis results include: population distribution information, living habit information and activity range information.

[0046] For the embodiments of the present application, in order to improve the scientific nature of ecological protection decisions and promote biodiversity conservation, animal information corresponding to different regions is obtained, and multidimensional analysis is performed based on the animal information corresponding to each region to obtain animal activity analysis results. Performing multidimensional analysis can understand the distribution of animal populations in various regions, accurately grasp the distribution patterns and change trends of different species in different geographical environments, and reveal the living habits of different species of organisms, which is helpful for predicting the response of animals to environmental changes. At the same time, accurately assessing the range of activities and space requirements of animals is of great significance for planning nature reserves and designing ecological corridors, etc., which can ensure that animals have enough space for activities and reduce the impact of human interference and habitat fragmentation.

[0047] Specifically, based on the animal information corresponding to each area, population distribution analysis is performed to determine population distribution information, that is, based on the animal information corresponding to each area and the geographical location information, an animal population distribution map is drawn. Through spatial analysis, the distribution density and distribution of animals in different areas can be intuitively displayed. Based on the animal information corresponding to each area, life habit analysis is performed to determine life habit information, that is, based on the animal status in the animal information corresponding to each area, integration and classification are performed to determine the activity patterns (e.g., diurnal, nocturnal, etc.), foraging behaviors (e.g., foraging locations, foraging methods, etc.) and social behaviors (e.g., domain division, group interaction, etc.) of animals in different time periods, which are recorded as life habit information. Based on the animal information corresponding to each area, activity range analysis is performed to determine activity range information, that is, based on the animal information corresponding to each area and the geographical location information, association marking is performed, and spatial analysis is performed using geographic information system software to draw an animal activity range map. The activity range map can intuitively display the activity trajectory and distribution area of ​​animals in different time periods.

[0048] It can be seen that in the embodiments of the present application, in order to improve the scientific nature of ecological protection decisions and promote biodiversity protection, animal information corresponding to different regions is obtained, and multidimensional analysis is performed based on the animal information corresponding to each region to obtain animal activity analysis results. Performing multidimensional analysis can understand the distribution of animal populations in each region, and can also reveal the living habits of different species of organisms, which is helpful for predicting the response of animals to environmental changes. At the same time, accurately assessing the activity range and space requirements of animals is of great significance for planning nature reserves and designing ecological corridors.

[0049] Furthermore, in order to reduce the energy consumption of the sound collection device, reduce unnecessary power waste, reduce the working time and load of internal components, and extend the service life of the device, in the embodiment of the present application, it also includes: When no active sound is detected in the monitoring area, the sound collection device is controlled to enter a dormant state, so that the sound collection device can save energy and extend its service life.

[0050] For the embodiments of the present application, the sound collection device is usually powered by a battery, and long battery life is a prerequisite for ensuring the efficient operation of the sound collection device. However, the sound collection device consumes a large amount of electricity during continuous operation, and also accelerates the wear and aging of its internal components. Therefore, when there is no active sound in the monitoring area, controlling the sound collection device to enter a dormant state can significantly reduce the energy consumption of the sound collection device, reduce unnecessary waste of electricity, and at the same time, reduce the working time and load of internal components, thereby extending the service life of the device. When an active sound is detected, the sound collection device quickly wakes up and starts collecting data; when there is no active sound, it automatically enters a dormant state to save resources. This on-demand work mode helps to reduce data redundancy and storage pressure, and improve the efficiency of data processing and analysis.

[0051] It can be seen that in the embodiment of the present application, when no active sound is detected in the monitoring area, the sound collection device is controlled to enter a sleep state, so that the sound collection device can save energy and extend its service life. In addition, the on-demand working mode helps to reduce data redundancy and storage pressure, and improve the efficiency of data processing and analysis.

[0052] The above embodiment introduces an animal voiceprint monitoring method from the perspective of method flow, and the following embodiment introduces an animal voiceprint monitoring device from the perspective of a virtual module or a virtual unit. For details, please refer to the following embodiment.

[0053] The present application embodiment provides an animal voiceprint monitoring device, such as Figure 2 As shown, the animal voiceprint monitoring device may specifically include: The sound source localization module 210 is used to orient the sound source of the active sound when detecting the presence of active sound in the monitoring area, control the sound collection device to collect the direction of the active sound and collect audio; The data preprocessing module 220 is used to obtain the audio data in the monitoring area collected by the sound collection device, and preprocess the audio data using an audio processing algorithm to obtain target audio data, wherein the preprocessing includes: denoising, filtering, audio segmentation, and mono separation; A voiceprint feature extraction module 230 is used to extract voiceprint features based on target audio data and determine a multi-dimensional voiceprint feature vector, wherein the multi-dimensional voiceprint feature vector is a multi-dimensional vector composed of frequency features, amplitude features, time domain features and frequency domain features; The voiceprint matching module 240 is used to obtain an animal voiceprint model library, match the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library, and determine the animal information, wherein the animal information includes: animal species and animal status.

[0054] For the embodiment of the present application, in order to collect purer audio data that is less affected by background noise and irrelevant sounds, when active sounds are detected in the monitoring area, the active sounds are oriented to the sound source, the sound collection direction of the sound collection device is controlled to be the direction of the active sounds and audio collection is performed, and the sound source direction is accurately aligned, so as to effectively capture the sounds of animals in a wide monitoring area. Then, the audio data in the monitoring area collected by the sound collection device is obtained, and the audio processing algorithm is used to pre-process the audio data to obtain the target audio data. The pre-processing is used to significantly improve the quality of the audio data and enhance the accuracy and processing efficiency of voiceprint recognition. Furthermore, voiceprint feature extraction is performed based on the target audio data to determine a multi-dimensional voiceprint feature vector. The multi-dimensional voiceprint feature vector can comprehensively and meticulously display the multi-faceted sound characteristics of the target audio data, which helps to improve the accuracy and robustness of subsequent voiceprint model matching. Finally, the multi-dimensional voiceprint feature vector is matched with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information. Audio collection is performed through sound source positioning, making the audio data less affected by background noise and irrelevant sounds, and preprocessing operations are performed to significantly improve the quality of the audio data. Animal voiceprint monitoring is performed using high-quality target audio data, thereby improving the accuracy of animal voiceprint monitoring.

[0055] In a possible implementation of the embodiment of the present application, when the data preprocessing module 220 performs preprocessing on the audio data using the audio processing algorithm to obtain the target audio data, it is used to: Performing denoising and filtering processing on the audio data to obtain first audio data, wherein the denoising and filtering processing is used to remove complex and changeable background noise in the audio data; Performing mono separation based on the first audio data to determine a target source audio signal; The audio segmentation technology is used to segment the continuous target source audio signal to obtain the target audio data.

[0056] In a possible implementation of the embodiment of the present application, the animal voiceprint monitoring device further includes: A data association storage module is used to control the shooting direction of the image acquisition device to be the direction where the activity sound is emitted and to shoot the image to obtain the animal image; Verify monitoring results based on animal images and animal information, and determine monitoring verification results; When the monitoring verification result is verification passed, data association is performed based on the animal image, target audio data and animal information, and stored in the animal information database.

[0057] In a possible implementation of the embodiment of the present application, the animal voiceprint monitoring device further includes: A manual confirmation module is used to associate data based on the animal image and the target audio data and store them in the animal information database when the monitoring verification result is a verification failure; In the animal information database, the data to be confirmed are marked, wherein the data marking facilitates manual animal voiceprint recognition of the data to be confirmed, and the data to be confirmed are the animal images, audio data and target audio data that failed verification; An animal voiceprint confirmation instruction carrying target animal information is obtained, and based on the animal voiceprint confirmation instruction, data to be confirmed is filled in an animal information database, wherein the data filling is used to enrich the animal information database information.

[0058] In a possible implementation of the embodiment of the present application, the animal voiceprint monitoring device further includes: The multidimensional analysis module is used to obtain animal information corresponding to different areas, perform multidimensional analysis based on the animal information corresponding to each area, and obtain animal activity analysis results, wherein the animal activity analysis results include: population distribution information, living habit information and activity range information.

[0059] In a possible implementation of the embodiment of the present application, the animal voiceprint monitoring device further includes: The sleep module is used to control the sound collection device to enter a sleep state when no active sound is detected in the monitoring area, so that the sound collection device can save energy and extend its service life.

[0060] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the animal voiceprint monitoring device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0061] An electronic device is provided in an embodiment of the present application, such as Figure 3 As shown, Figure 3 The electronic device 300 shown includes: a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, such as through a bus 302. Optionally, the electronic device 300 may also include a transceiver 304. It should be noted that in actual applications, the transceiver 304 is not limited to one, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present application.

[0062] The processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0063] The bus 302 may include a path to transmit information between the above components. The bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but it does not mean that there is only one bus or only one type of bus.

[0064] The memory 303 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0065] The memory 303 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the contents shown in the above method embodiment.

[0066] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server, etc. Figure 3 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0067] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer-readable storage medium is run on a computer, the computer can execute the corresponding content in the aforementioned method embodiment.

[0068] The embodiment of the present application provides a computer program product, which includes a computer program, which implements the method in any of the above embodiments when executed by a processor. Compared with the related art, in order to collect purer audio data that is less disturbed by background noise and irrelevant sounds, when the active sound is detected in the monitoring area, the active sound is oriented to the sound source, the sound collection direction of the sound collection device is controlled to be the direction of the active sound and audio collection is performed, and the sound source direction is accurately aligned to effectively capture the sound of animals in a wide monitoring area. Then, the audio data in the monitoring area collected by the sound collection device is obtained, and the audio data is preprocessed using an audio processing algorithm to obtain target audio data. The preprocessing is used to significantly improve the quality of the audio data, enhance the accuracy and processing efficiency of voiceprint recognition. Furthermore, voiceprint feature extraction is performed based on the target audio data to determine a multi-dimensional voiceprint feature vector, which can comprehensively and meticulously display the multi-faceted sound characteristics of the target audio data, and help improve the accuracy and robustness of subsequent voiceprint model matching. Finally, the animal information is determined by matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library. Audio data is collected through sound source localization, making the audio data less affected by background noise and irrelevant sounds, and preprocessing operations are performed to significantly improve the quality of audio data. Animal voiceprint monitoring is performed using high-quality target audio data, which improves the accuracy of animal voiceprint monitoring.

[0069] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0070] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for monitoring animal voiceprints, characterized in that: include: When active sound is detected in the monitoring area, the sound source of the active sound is oriented, the sound collection device is controlled to collect the sound in the direction of the active sound and the audio is collected; Acquire the audio data in the monitoring area collected by the sound collection device, and pre-process the audio data using an audio processing algorithm to obtain target audio data, wherein the pre-processing includes: denoising, filtering, audio segmentation, and mono separation; Extract voiceprint features based on the target audio data to determine a multidimensional voiceprint feature vector, wherein the multidimensional voiceprint feature vector is a multidimensional vector composed of frequency features, amplitude features, time domain features, and frequency domain features; An animal voiceprint model library is obtained, and animal information is determined by matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library, wherein the animal information includes: animal species and animal status.

2. The animal voiceprint monitoring method according to claim 1, characterized in that: The preprocessing of the audio data by using an audio processing algorithm to obtain target audio data includes: Performing denoising and filtering processing on the audio data to obtain first audio data, wherein the denoising and filtering processing is used to remove complex and changeable background noise in the audio data; Performing mono separation based on the first audio data to determine a target source audio signal; The continuous target source audio signal is segmented using an audio segmentation technique to obtain target audio data.

3. The animal voiceprint monitoring method according to claim 1, characterized in that: After matching the multi-dimensional voiceprint feature vector with the voiceprint model corresponding to each species in the animal voiceprint model library to determine the animal information, the method further includes: Controlling the shooting direction of the image acquisition device to the direction where the activity sound is emitted and taking images to obtain animal images; Verify the monitoring result based on the animal image and the animal information, and determine the monitoring verification result; When the monitoring verification result is verification passed, data association is performed based on the animal image, the target audio data and the animal information, and stored in an animal information database.

4. The animal voiceprint monitoring method according to claim 3, characterized in that: After verifying the monitoring result based on the animal image and the animal information and determining the monitoring verification result, the method further includes: When the monitoring verification result is a verification failure, data association is performed based on the animal image and the target audio data, and the data is stored in the animal information database; In the animal information database, the data to be confirmed are data-marked, wherein the data marking facilitates manual animal voiceprint recognition of the data to be confirmed, and the data to be confirmed are the animal image that failed verification, the audio data, and the target audio data; An animal voiceprint confirmation instruction carrying target animal information is obtained, and based on the animal voiceprint confirmation instruction, data to be confirmed in the animal information database is filled with data, wherein the data filling is used to enrich the animal information database information.

5. The animal voiceprint monitoring method according to claim 1, characterized in that: Also includes: The animal information corresponding to different areas is obtained, and multi-dimensional analysis is performed based on the animal information corresponding to each area to obtain animal activity analysis results, wherein the animal activity analysis results include: population distribution information, living habit information and activity range information.

6. The animal voiceprint monitoring method according to claim 1, characterized in that: Also includes: When no active sound is detected in the monitoring area, the sound collecting device is controlled to enter a dormant state, so that the sound collecting device can save energy and extend its service life.

7. An electronic device, characterized in that: include: at least one processor; Memory; At least one application, wherein the at least one application is stored in a memory and configured to be executed by at least one processor, and the at least one application is configured to: execute the animal voiceprint monitoring method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed in a computer, the computer is caused to execute the animal voiceprint monitoring method according to any one of claims 1 to 6.

9. A computer program product, characterized in that The method comprises a computer program, wherein the computer program is executed by a processor to implement the animal voiceprint monitoring method according to any one of claims 1 to 6.

Images