Bird feeder-based driving and feeding method, device and equipment and readable medium

Abstract

The embodiment of the invention discloses a bird feeder-based driving and feeding method, device and equipment and a readable medium. A specific embodiment of the method comprises the steps of collecting multi-modal monitoring data of a corresponding feeding area; on the basis of the multi-modal monitoring data, species identification processing is carried out on the feeding area, and species identification information is obtained; in response to determining that the identified species information represents that the identification result is a bird identifier and a confidence coefficient value represented by the confidence coefficient information is smaller than a preset value, performing auxiliary identification processing based on preset species appearing and disappearing information corresponding to geographic position information of an area where the preset feeder is located to obtain auxiliary identified species information; based on the auxiliary identification species information, a preset bird feeder is controlled to execute driving and feeding treatment; and in response to determining that a confidence coefficient value represented by the confidence coefficient information is greater than or equal to a preset value, controlling a preset bird feeder to execute driving and feeding treatment based on the identified species information. According to the embodiment, the driving and feeding accuracy of the bird feeder is improved.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technology, and more particularly to a bird feeder-based driving and feeding method, device, equipment, and readable medium. Background Art

[0002] Bird feeder-based driving and feeding technology uses a bird feeder to identify birds in the bird feeder's feeding area, feed the identified bird species, and drive away identified non-bird species. Currently, when using a bird feeder to identify birds in the bird feeder's feeding area, the common method is to only use the bird feeder's camera module to identify birds in the bird feeder's feeding area.

[0003] However, when using the above method to identify birds in the feeding area of ​​a bird feeder, the following technical problems often occur:

[0004] Bird identification in the bird feeder's feeding area is performed solely through the camera module on the bird feeder. This single visual recognition is easily affected by factors such as occlusion, lighting changes, and angle offset, resulting in poor bird identification accuracy. Furthermore, when the confidence level of the bird identification result is low, auxiliary identification processing based on the preset species presence information corresponding to the geographic location of the feeder area is not performed. Although an identification result (e.g., bird or non-bird) is obtained, the low confidence level leads to large identification errors, resulting in poor accuracy in driving and feeding the bird feeder.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the inventive concept and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0007] Some embodiments of the present disclosure provide a bird feeder-based driving and feeding method, device, electronic device, and computer-readable medium to solve one or more of the technical problems mentioned in the background technology section above.

[0008] In a first aspect, some embodiments of the present disclosure provide a bird feeder-based driving and feeding method, the method comprising: collecting multimodal monitoring data corresponding to a feeding area of ​​a preset bird feeder; performing species identification processing in the feeding area based on the multimodal monitoring data to obtain species identification information, wherein the species identification information includes identification species information and confidence information; in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, performing auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information; based on the auxiliary identification species information, controlling the preset bird feeder to perform driving and feeding processing; in response to determining that the confidence value represented by the confidence information is greater than or equal to the preset value, controlling the preset bird feeder to perform driving and feeding processing based on the identification species information.

[0009] In a second aspect, some embodiments of the present disclosure provide a bird feeder-based driving and feeding device, the device comprising: a collection unit, configured to collect multimodal monitoring data corresponding to a feeding area of ​​a preset bird feeder; an identification processing unit, configured to perform species identification processing in the feeding area based on the multimodal monitoring data to obtain species identification information, wherein the species identification information includes identification species information and confidence information; an auxiliary identification processing unit, configured to, in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information; a first control unit, configured to control the preset bird feeder to perform driving and feeding processing based on the auxiliary identification species information; and a second control unit, configured to control the preset bird feeder to perform driving and feeding processing based on the identification species information in response to determining that the confidence value represented by the confidence information is greater than or equal to the preset value.

[0010] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0012] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the bird feeder-based driving and feeding methods of some embodiments of the present disclosure, the accuracy of bird identification and bird feeder driving and feeding is improved. Specifically, the reason for the poor accuracy of bird identification and bird feeder driving and feeding is that only the camera module on the bird feeder is used to identify birds in the bird feeder feeding area. Single visual recognition is easily interfered by factors such as occlusion, lighting changes, and angle offset, resulting in poor bird identification accuracy. At the same time, when the confidence level of the bird identification result is low, auxiliary identification processing is not performed based on the preset species presence information corresponding to the geographical location information of the feeder area. Although the identification result (for example, bird or non-bird) is obtained, due to its low confidence level, the recognition error is large, resulting in poor accuracy of bird feeder driving and feeding. Based on this, the bird feeder-based driving and feeding methods of some embodiments of the present disclosure first collect multimodal monitoring data corresponding to the above-mentioned feeding area for the preset bird feeder. In this way, multimodal monitoring data of the feeding area can be collected. The multimodal detection data includes characteristic information of species in the feeding area captured from different dimensions. Then, based on the multimodal monitoring data, species identification processing is performed on the feeding area to obtain species identification information, wherein the species identification information includes identification species information and confidence information. Thus, species identification processing can be performed on the multimodal monitoring data including features from different dimensions, compensating for information loss caused by factors such as occlusion, illumination changes, and angle offset in single visual recognition, thereby improving the accuracy of species identification in the feeding area. Then, in response to determining that the identification species information represents the identification result as a bird identification and the confidence value represented by the confidence information is less than a preset value, auxiliary identification processing is performed based on the preset species presence information corresponding to the geographic location information of the area where the preset feeder is located, to obtain auxiliary identification species information. Thus, when the confidence level of the identification result is low, auxiliary identification processing can be performed based on the preset species presence information to obtain auxiliary identification species information. Thus, prior knowledge related to geographic location (i.e., preset species presence information) can be utilized to provide additional judgment basis for low-confidence identification results, effectively reducing identification errors. Next, based on the auxiliary species identification information, the preset bird feeder is controlled to execute the driving and feeding process. Thereafter, in response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, the preset bird feeder is controlled to execute the driving and feeding process based on the identified species information. This allows the bird feeder to be controlled to drive and feed birds based on identified species information with a small recognition error or a high confidence level, thereby improving the accuracy of the bird feeder's driving and feeding process. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0014] Figure 1 is a flow chart of some embodiments of the bird feeder-based driving and feeding method according to the present disclosure;

[0015] Figure 2 is a flow chart of other embodiments of the bird feeder-based driving and feeding method according to the present disclosure;

[0016] Figure 3 1 is a schematic structural diagram of some embodiments of a bird feeder-based driving and feeding device according to the present disclosure;

[0017] Figure 4 It is a structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0018] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0019] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0020] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0024] Figure 1 The process 100 of some embodiments of the bird feeder-based driving and feeding method according to the present disclosure is shown. The bird feeder-based driving and feeding method includes the following steps:

[0025] Step 101: For a feeding area of ​​a preset bird feeder, collect multimodal monitoring data corresponding to the feeding area.

[0026] In some embodiments, the execution entity (eg, a computing device) of the bird feeder-based driving and feeding method may collect multimodal monitoring data corresponding to a preset feeding area of ​​the bird feeder.

[0027] In some optional implementations of some embodiments, the execution entity may collect multimodal monitoring data corresponding to a feeding area of ​​a preset bird feeder by performing the following steps:

[0028] The first step is to collect the regional image of the feeding area after the preset time period of the preset bird feeder being turned on, and perform picture change detection processing on the regional image based on the pre-created feeding area background model to obtain picture detection information. The feeding area background model is a model (for example, a background image) obtained by modeling the background of the feeding area of ​​the preset bird feeder. In practice, the execution entity can collect regional images of the feeding area according to the preset collection frequency within the preset time period. Then, the execution entity can create a feeding area background model based on each collected regional image through a background modeling algorithm (for example, a Gaussian mixture model (GMM), a mean background method). The execution entity can compare the regional image with the feeding area background model to obtain a difference map. Then, the execution entity can determine the number of pixel points in the difference map whose pixel values ​​are greater than the preset value as the target number. Thereafter, in response to determining that the target number is greater than the preset number, the execution entity can determine the information representing the picture change as the picture detection information.

[0029] In the second step, in response to determining that the image detection information indicates an image change, or detecting that the sound trigger device emits a sound anomaly signal, or detecting that the weight trigger device emits a gravity anomaly signal, the camera, sound collection device, and gravity vibration collection device located in the feeding area are controlled to collect visual data, sound data, and gravity vibration data. The sound trigger device may be a device (e.g., a sound trigger) that uses a microphone to collect sound signals at a preset frequency, detects whether the sound signals contain short-term energy, and emits a sound anomaly signal when short-term energy is present. The sound anomaly signal is a signal (e.g., an electrical signal, a digital signal, etc.) emitted by the sound trigger device when it detects the presence of short-term energy in the collected sound. The weight trigger device may be a device (e.g., a force sensor) that collects the weight of the feeding area at a preset frequency and emits a gravity anomaly signal when the weight exceeds a preset weight. The camera may be a camera mounted on a preset bird feeder. The visual data may be data collected during a first preset collection period (i.e., a preset period after the camera device is triggered to begin collecting data at the preset collection frequency), including at least one image of the feeding area. The sound collection device may be a sound collector mounted on the preset bird feeder. The above-mentioned sound data is data collected within a second preset collection time period (i.e., a preset time period after the time point when the sound collection device is triggered to start collecting data), and may include digital audio data. The above-mentioned gravity vibration collection device may include a vibration sensor and a gravity sensor. The above-mentioned gravity vibration data includes average gravity and vibration signal information. The above-mentioned average gravity may be the average of various gravity values ​​collected at a preset collection frequency within a preset time period after the gravity vibration collection device is triggered. The above-mentioned vibration signal information may represent the vibration signal collected by the vibration collection device within a preset time period after being triggered.

[0030] The third step is to determine the collected visual data, sound data and gravity vibration data as multimodal monitoring data corresponding to the above-mentioned feeding area.

[0031] Step 102: Perform species identification processing in the feeding area based on the multimodal monitoring data to obtain species identification information.

[0032] In some embodiments, the above-mentioned execution entity may perform species identification processing in the feeding area based on the above-mentioned multimodal monitoring data to obtain species identification information. The above-mentioned species identification information includes species identification information and confidence information. In practice, the above-mentioned execution entity may input the above-mentioned multimodal monitoring data into a pre-trained multimodal species identification model to obtain species identification information. The above-mentioned species identification information includes species identification information and confidence information. The above-mentioned species identification information may characterize the classification of the species (for example, bird or non-bird). The above-mentioned confidence information may represent the confidence in the classification of the species. The above-mentioned multimodal species identification model may be a multimodal Transformer model that takes multimodal monitoring data as input and species identification information as output.

[0033] In some optional implementations of some embodiments, the execution entity may perform species identification processing in the feeding area based on the multimodal monitoring data to obtain species identification information through the following steps:

[0034] The first step is to generate visual species identification information, soundprint species identification information and gravity vibration species identification information based on the above multimodal monitoring data.

[0035] In the second step, species identification information is generated based on the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information. The visual species identification information includes a category label and a confidence level. The voiceprint species identification information includes a voiceprint identification category label and a confidence level. In response to determining that the category labels included in the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information are all bird identification, the execution entity may determine the average of the confidence levels included in the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information as the target confidence level. Next, the execution entity may determine the bird identification as the identified species information and the target confidence level as the confidence level information. Finally, the execution entity may determine the identified species information and the confidence level information as the species identification information. In response to determining that at least one of the category labels included in the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information is a non-bird identification, the execution entity may determine at least one confidence level corresponding to at least one classification label corresponding to the at least one non-bird identification as at least one negative confidence level. Afterwards, the above-mentioned execution entity may determine at least one confidence level corresponding to at least one category label for bird identification in the visual species identification information, the above-mentioned voiceprint species identification information, and the above-mentioned gravity vibration species identification information as at least one positive confidence level. Afterwards, the above-mentioned execution entity may determine the sum of at least one positive confidence level as a first numerical value. The above-mentioned execution entity may determine the sum of at least one negative confidence level as a second numerical value. Then, the above-mentioned execution entity may determine the difference between the first numerical value and the second numerical value as a target numerical value. In response to determining that the above-mentioned target numerical value is less than or equal to 0, the above-mentioned execution entity may determine the non-bird identification as the identification species information, and determine the absolute value of the target numerical value as the confidence information. In response to determining that the above-mentioned target numerical value is greater than 0, the above-mentioned execution entity may determine the bird identification as the identification species information, and determine the target numerical value as the confidence information. Finally, the above-mentioned execution entity may determine the identification species information and the confidence information as species identification information.

[0036] In some optional implementations of some embodiments, the execution entity may generate visual species identification information, voiceprint species identification information, and gravity vibration species identification information based on the multimodal monitoring data through the following steps:

[0037] The first step is to perform identification detection processing on the visual data included in the above-mentioned multimodal monitoring data through a pre-trained species recognition model to obtain visual species recognition information. In practice, the above-mentioned execution entity can randomly take out an image from at least one image of the feeding area included in the above-mentioned visual data as the feeding area image to be identified. Then, the above-mentioned execution entity can input the above-mentioned feeding area image to be identified into the pre-trained species recognition model to obtain visual species recognition information. The above-mentioned visual species recognition information includes category labels and confidence levels. The above-mentioned species recognition model can be a neural network model (for example, a YOLOv8 Nano model or an EfficientDet-D0 model) that takes the feeding area image to be identified as input and outputs visual species recognition information.

[0038] The second step is to perform voiceprint feature recognition processing on the sound data included in the above-mentioned multimodal monitoring data to obtain voiceprint species identification information. In practice, the above-mentioned execution entity can use MFCC (Mel Frequency Cepstral Coefficient) extraction technology to extract frequency domain feature information corresponding to the above-mentioned sound data from the digital audio data included in the above-mentioned sound data. The above-mentioned frequency domain feature information can be an MFCC vector sequence of audio. Then, the above-mentioned execution entity can input the above-mentioned MFCC vector sequence into a pre-trained lightweight sound event detection model (for example, a CRNN model, i.e., a convolutional recurrent neural network) to obtain voiceprint species identification information. The above-mentioned voiceprint species identification information includes a voiceprint identification category label (for example, a bird identification or a non-bird identification) and a confidence level.

[0039] The third step is to generate gravity vibration species identification information based on the gravity vibration data included in the above-mentioned multimodal monitoring data. In practice, the above-mentioned execution entity may determine the RMS (reflecting average energy) of the vibration signal included in the above-mentioned vibration data as the vibration energy. Then, in response to determining that the average gravity included in the above-mentioned gravity vibration data is less than the preset gravity and the above-mentioned vibration energy is less than the preset vibration energy, the above-mentioned execution entity may determine the bird identification and the preset confidence as the gravity vibration species identification information. In response to determining that the average gravity included in the above-mentioned gravity vibration data is greater than or equal to the preset gravity and the above-mentioned vibration energy is greater than or equal to the preset vibration energy, the above-mentioned execution entity may determine the non-bird identification and the preset confidence as the gravity vibration species identification information.

[0040] In some optional implementations of some embodiments, the execution entity may perform voiceprint feature recognition processing on the sound data included in the multimodal monitoring data to obtain voiceprint species identification information through the following steps:

[0041] The first step is to perform framing on the digital audio data included in the sound data to obtain an initial audio frame sequence. In practice, the execution entity can perform framing on the digital audio signal, discretizing the continuous signal into overlapping short-duration segments using a sliding window to generate an audio frame sequence. Each audio frame in the audio frame sequence can represent a windowed segment of a fixed duration (e.g., 25ms) in the original signal, i.e., the digital audio signal.

[0042] In the second step, for each initial audio frame in the above-mentioned initial audio frame sequence, in response to determining that the high-frequency and low-frequency energy ratio of the above-mentioned initial audio frame is greater than the preset ratio, the above-mentioned initial audio frame in the initial audio frame sequence is framed again to update the initial audio frame sequence. In practice, the above-mentioned execution subject can convert the initial audio frame into spectrum amplitude information through Fourier transform technology. Afterwards, the above-mentioned execution subject can use frequency band mask technology to determine the ratio of the sum of the squares of the high-frequency band spectrum amplitude to the sum of the squares of the low-frequency band spectrum amplitude as the high-frequency and low-frequency energy ratio. For example, the high frequency band is a frequency range of (>2kHz) and the low frequency band is a frequency range of (<500Hz). As an example, the initial audio frame sequence can be (F1, F2, F3). If the high-frequency and low-frequency energy ratio of F2 is greater than the preset ratio. Then the updated initial audio frame sequence can be (F1, F2-1, F2-2, F3).

[0043] In the third step, the updated initial audio frame sequence is input into the pre-trained noise suppression model to obtain an audio frame sequence.

[0044] The fourth step is to perform speech enhancement processing on the audio frame sequence to obtain an enhanced audio frame sequence. In practice, the execution subject may use spectral masking technology to perform speech enhancement processing on the audio frame sequence to obtain an enhanced audio frame sequence.

[0045] Step 5: For each audio frame in the audio frame sequence, extract the frequency domain feature information of the audio frame to obtain the audio frame frequency domain feature information. In practice, the execution entity may use MFCC (Mel-Frequency Cepstral Coefficient) extraction technology to extract the frequency domain feature information of the audio frame. The audio frame frequency domain feature information may be the MFCC vector of the audio frame.

[0046] In the sixth step, the obtained frequency domain feature information of each audio frame is arranged according to the order of the corresponding audio frames in the audio frame sequence to obtain an audio frame frequency domain feature information sequence.

[0047] In the seventh step, the above audio frame frequency domain feature information sequence is input into a pre-trained lightweight sound event detection model to obtain voiceprint species identification information.

[0048] The above-described technical solution and its related contents, as an inventive feature of an embodiment of the present disclosure, address the technical problem of "low accuracy in voiceprint species identification." Factors contributing to low accuracy in voiceprint species identification are often as follows: In voiceprint species identification, high-frequency components (such as birdsong and insect wingbeats) often carry species-specific information. Fixed-length framing may not fully capture the time-varying characteristics of high-frequency components. Furthermore, the original audio may contain a large amount of background noise (such as wind and ambient noise), which interferes with feature extraction, resulting in low accuracy in voiceprint species identification. If these factors are addressed, the accuracy of voiceprint species identification can be improved. To achieve this, the digital audio data included in the sound data is first framed to obtain an initial audio frame sequence. Second, for each initial audio frame in the initial audio frame sequence, in response to determining that the proportion of high-frequency and low-frequency energy in the initial audio frame is greater than a preset ratio, the initial audio frame in the initial audio frame sequence is further framed to update the initial audio frame sequence. Therefore, by analyzing the high-frequency and low-frequency energy ratios for each audio frame in the initial audio frame sequence, frames with high high-frequency energy ratios can be further divided into smaller subframes to more accurately capture the time-varying characteristics of high-frequency components. Third, the updated initial audio frame sequence is input into a pre-trained noise suppression model to obtain an audio frame sequence. The updated initial audio frame sequence can then be input into the pre-trained noise suppression model to remove background noise. Fourth, speech enhancement is performed on the audio frame sequence to obtain an enhanced audio frame sequence. Spectral masking techniques are used to further suppress non-target components. Noise suppression and speech enhancement improve the signal-to-noise ratio (SNR) and reduce noise interference. Fifth, for each audio frame in the audio frame sequence, frequency domain feature information is extracted to obtain audio frame frequency domain feature information. Sixth, the obtained audio frame frequency domain feature information is arranged according to the order of the corresponding audio frames in the audio frame sequence to obtain an audio frame frequency domain feature information sequence. Thus, the audio frame frequency domain feature information sequence corresponding to the audio frame sequence can be extracted. In the seventh step, the aforementioned audio frame frequency domain feature information sequence is input into a pre-trained lightweight sound event detection model to obtain voiceprint species identification information. This is also because, before extracting the frequency domain feature information, the digital audio data is framed to obtain an initial audio frame sequence. By analyzing the high-frequency and low-frequency energy ratios of each audio frame in the initial audio frame sequence, frames with high high-frequency energy ratios are further framed to generate shorter subframes. This allows for more accurate capture of the time-varying characteristics of the high-frequency components. This, in turn, makes the time-varying characteristics of the high-frequency components included in the audio frame frequency domain feature information sequence more accurate, thereby improving the accuracy of voiceprint species identification.At the same time, the initial audio frame sequence is denoised, and the audio frame sequence is subjected to speech enhancement processing after denoising, which reduces the interference of background noise on feature extraction, improves the accuracy of the extracted audio frame frequency domain feature information sequence, and further improves the accuracy of voiceprint species recognition.

[0049] Step 103, in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, auxiliary identification processing is performed based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information.

[0050] In some embodiments, the execution subject may, in response to determining that the identification result represented by the above-mentioned identification species information is a bird identification and the confidence value represented by the above-mentioned confidence information is less than a preset value, perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the above-mentioned preset feeder is located to obtain auxiliary identification species information. Among them, the above-mentioned preset species occurrence information includes a species occurrence time probability model. The above-mentioned species occurrence probability model is a mathematical model for predicting and quantifying the possibility of different animals appearing at preset times and places. The above-mentioned species occurrence time probability model can use a weighted time window (such as a Gaussian distribution fitting for time periods) to represent the probability of different animals appearing at different time points. As an example, the above-mentioned species occurrence time probability model can be:

[0051] "Distribution probability density function:

[0052] Among them, P(t) represents the probability of occurrence of a species in different time periods, reflecting the possibility of detecting the species at a specific time point t. σ is the standard deviation, which measures the degree of dispersion of the data. In the species appearance time probability model, σ determines the "width" of the probability distribution. The larger σ is, the more dispersed the time range of species appearance is. The smaller σ is, the more concentrated the time of species appearance is around the mean μ. e is the base of the natural logarithm, which is approximately equal to 2.71828. μ is the mean, and in the species appearance time probability model, it represents the time point with the highest probability of species appearance. For example, if a certain bird is usually most active around 12 o'clock in the morning, then μ can be set to 12 (assuming time is in hours).

[0053] Birds: active time μ = 12 (noon), σ = 2 (active period is 10:00-14:00).

[0054] Rodents: active time μ = 23 (midnight), σ = 3 (active period is 20:00-02:00).

[0055] Lizard: Active time μ = 10 (morning), σ = 5 (active period 05:00-15:00).

[0056] In some optional implementations of some embodiments, the execution entity may perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located through the following steps to obtain auxiliary identification species information:

[0057] The first step is to determine the time for collecting multimodal monitoring data as the collection time. The collection time can be the time corresponding to the time when the camera device, sound collection device, and gravity vibration collection device located in the feeding area start collecting data.

[0058] The second step is to obtain a list of candidate species corresponding to the geographic location information. The geographic location information indicates the geographic location of the preset bird feeder. The candidate species list may include a set of species identifiers of various species that occur in the area corresponding to the geographic location.

[0059] The third step is to generate the identification of each occurring species and each occurrence probability information corresponding to the above-mentioned collection time based on the above-mentioned collection time, the above-mentioned candidate species list and the above-mentioned species occurrence time probability model. In practice, the above-mentioned execution entity can determine the identification of each occurring species corresponding to each active time period including the above-mentioned collection time through the species occurrence time probability model. Afterwards, the above-mentioned execution entity can determine the intersection of the above-mentioned each occurring species identification and the set of species identifications included in the candidate species list as the target species identification set. Then, the above-mentioned execution entity can determine the each target species identification in the target species identification set as each occurring species identification. Then, for each target species identification in the above-mentioned target species identification set, the above-mentioned execution entity can determine the probability of occurrence of the above-mentioned target species identification at the above-mentioned collection time as the occurrence probability information through the above-mentioned species occurrence time probability model. Among them, each of the above-mentioned each occurring species identification corresponds to one of the occurrence probability information in the above-mentioned each occurrence probability information.

[0060] The fourth step is to generate auxiliary species identification information based on the aforementioned species identifications and occurrence probability information. In practice, the execution entity may determine the occurrence probability information with the highest occurrence probability among the occurrence probability information as the reference occurrence probability information. Finally, the execution entity may determine the occurrence species identification among the occurrence species identifications that corresponds to the reference occurrence probability information as the reference occurrence species identification. The reference occurrence species identification is determined as the auxiliary species identification information.

[0061] Step 104: Based on the auxiliary species identification information, control the preset bird feeder to perform driving and feeding processing.

[0062] In some embodiments, the execution entity may control the preset bird feeder to perform driving away and feeding processes based on the auxiliary species identification information.

[0063] In some optional implementations of some embodiments, the execution entity may control the preset bird feeder to perform driving and feeding processing based on the auxiliary species identification information through the following steps:

[0064] In the first step, in response to determining that the species identifier included in the auxiliary species identification information is a bird identifier, the preset bird feeder is activated to open the feeder compartment and generate feeder opening record information. In practice, the execution entity may drive a motor to open the feeder compartment door via a gear or slide mechanism. The feeder opening record information includes the identifier indicating the feeder compartment was opened and the time it was opened.

[0065] In the second step, in response to determining that the infested species identification included in the auxiliary species identification information is a non-bird identification, the following driving-away steps are performed:

[0066] The first sub-step is to control a sonic repelling device mounted on the preset bird feeder to play sound waves of a preset frequency. The sonic repelling device may be a speaker or a transducer. For example, the preset frequency may be 15-20 kHz.

[0067] The second sub-step is to control the vibration driving device installed on the preset bird feeder to start, wherein the vibration driving device can be a vibration generator.

[0068] The third sub-step is to control a light flashing driving device installed on the preset bird feeder to flash, wherein the light flashing driving device can be an LED light, a flash light, or a laser light.

[0069] The fourth sub-step is to generate driving record information corresponding to the sonic driving device, the vibration driving device, and the light flashing driving device, wherein the driving record information may include the activation time of the sonic driving device, the vibration driving device, and the light flashing driving device.

[0070] The third step is to send the generated food bin opening record information and the above-mentioned expulsion record information to a preset user terminal. The above-mentioned preset user terminal can be a mobile phone or computer used by the user.

[0071] Step 105 : In response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, based on the identified species information, controlling the preset bird feeder to perform driving and feeding processing.

[0072] In some embodiments, the execution entity may control the preset bird feeder to perform driving and feeding processing based on the identified species information in response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value.

[0073] In some optional implementations of some embodiments, the execution entity may control the preset bird feeder to perform driving and feeding processing based on the identified species information through the following steps:

[0074] In the first step, in response to determining that the species identification included in the species identification information is a bird identification, the preset bird feeder is driven to open the feeding bin and generate feeding bin opening record information.

[0075] In the second step, in response to determining that the species identification included in the species identification information is a non-bird identification, the following driving-away steps are performed:

[0076] The first sub-step is to control the sonic repelling device installed on the above-mentioned preset bird feeder to play sound waves of a preset frequency.

[0077] The second sub-step is to control the vibration driving device installed on the above-mentioned preset bird feeder to start.

[0078] The third sub-step is to control the light flashing driving device installed on the above-mentioned preset bird feeder to flash.

[0079] The fourth sub-step is to generate driving record information corresponding to the sound wave driving device, the vibration driving device and the light flashing driving device.

[0080] The third step is to send the generated food bin opening record information and the above-mentioned expulsion record information to the preset user terminal.

[0081] In the process of adopting technical solutions to solve the problems mentioned in the background technology, the following problems often arise:

[0082] The visual data included in the multimodal monitoring data is identified, detected, and processed using a pre-selected species recognition model, without introducing an online learning optimization mechanism for the model. Since species themselves are dynamic and environmental conditions are constantly changing (such as climate change and habitat destruction), this may lead to changes in the distribution, behavior, or characteristics of species (for example, some birds may migrate to new areas due to habitat loss, and their behavior may change due to environmental pressure). If the pre-trained species recognition model is used continuously without continuous optimization, the species recognition model may perform poorly in certain scenarios, such as misidentifying similar species or missing rare species. The species recognition model's generalization ability is poor, and the recognition accuracy will gradually deteriorate.

[0083] Faced with the above technical problems, the inventors decided to adopt the following solutions:

[0084] In some optional implementations of some embodiments, the above execution entity may further perform the following steps:

[0085] The first step is, in response to determining that the confidence level included in the above-mentioned visual species identification information is greater than or equal to the target confidence level, determining the feeding area image to be identified, the above-mentioned species identification mark and the above-mentioned confidence level included in the above-mentioned multimodal monitoring data as updated training samples, and storing the above-mentioned updated training samples in a preset training database. The above-mentioned preset training database includes a training sample set, and the training samples in the above-mentioned training sample set have corresponding storage times. The above-mentioned training samples are updated training samples or user feedback updated training samples. The above-mentioned user feedback updated training samples can be training samples manually annotated by users. The above-mentioned user feedback updated training samples include feeding area images, species identification marks and confidence levels (here the confidence level can be 1).

[0086] In the second step, the following online learning optimization steps are performed according to the preset frequency:

[0087] The first sub-step is to obtain a training sample set from the above-mentioned preset training database.

[0088] The second sub-step is to determine the number of training samples included in the training sample set as the updated training sample data number.

[0089] The third sub-step is to generate a target acquisition quantity based on a preset ratio and the updated number of training samples. The preset ratio may be 30% (i.e., the number of original training samples accounts for 30% of the total number of training samples in the training sample set). In practice, the target acquisition quantity of the execution entity may be: target acquisition quantity = updated number of training samples ÷ (1-30%) × 30%.

[0090] The fourth sub-step is to obtain a target number of original training samples from a preset original training sample set.

[0091] In a fifth sub-step, the target number of original training samples is added to the training sample set to update the training sample set to obtain an updated training sample set.

[0092] The sixth sub-step is to train the above-mentioned species recognition model based on the above-mentioned updated training sample set to obtain a trained updated species recognition model to replace the pre-trained species recognition model. As an example, the feeding area image included in each new training sample in the updated training sample set can first be used as input, and the species identification identifier (i.e., the label or name of the species) and the confidence level corresponding to the feeding area image in the updated training sample set can be used as the expected output to train an updated species recognition model. Afterwards, the above-mentioned execution entity can replace the updated species recognition model with the pre-trained species recognition model to update the pre-trained species recognition model.

[0093] The above technical solution and its related contents, as an inventive point of an embodiment of the present disclosure, solve the technical problem that "the generalization ability of the species recognition model is poor, and the recognition accuracy will gradually deteriorate". The factors that lead to the poor generalization ability of the species recognition model and the gradual deterioration of the recognition accuracy are often as follows: the visual data included in the multimodal monitoring data is recognized and detected by a pre-selected species recognition model, and no model online learning optimization mechanism is introduced. Since species themselves are dynamically changing and environmental conditions are constantly changing (such as climate change and habitat destruction), the distribution, behavior or characteristics of species may change (for example, some birds may migrate to new areas due to habitat loss, and their behavior may change due to environmental pressure). If the pre-trained species recognition model is always used and the species recognition model is not continuously optimized, the species recognition model may perform poorly in certain scenarios, such as misidentifying similar species or missing rare species. The generalization ability of the species recognition model is poor, and the recognition accuracy will gradually deteriorate. If the above factors are solved, the effect of improving the generalization ability and recognition accuracy of the species recognition model can be achieved. To achieve this effect, first, in response to determining that the confidence level of the visual species recognition information is greater than or equal to the target confidence level, the multimodal monitoring data, including the image of the feeding area to be identified, the species identification mark, and the confidence level, is determined as an updated training sample. The updated training sample is then stored in a preset training database. The preset training database includes a training sample set, wherein each training sample in the training sample set is associated with a storage time, and each training sample is an updated training sample or a user-feedback updated training sample. This allows high-confidence recognition results to be retained as updated training samples, and the updated training samples are stored in the preset training database for online learning optimization of the species recognition model. Then, the following online learning optimization steps are performed at a preset frequency: First, a training sample set is obtained from the preset training database. This obtains a training sample set containing updated training samples or user-feedback updated training samples that retain high-confidence recognition results. Second, the number of training samples included in the training sample set is determined as the number of updated training sample data. Third, a target acquisition number is generated based on a preset ratio and the number of updated training samples. In the fourth sub-step, a target number of original training samples are obtained from a preset original training sample set. In the fifth sub-step, the target number of original training samples are added to the training sample set to update the training sample set, thereby obtaining an updated training sample set. This results in an updated training sample set that includes both the original samples and the updated training samples, ensuring that the model does not completely forget old features and that the addition of new samples enhances its adaptability to changes. In the sixth sub-step, the species recognition model is trained based on the updated training sample set to obtain a trained species recognition model, thereby optimizing the species recognition model.This allows the species recognition model to be retrained based on the updated training sample set, resulting in a newly trained model that replaces the old one. Furthermore, thanks to the introduction of online learning optimization mechanisms in these steps, the species recognition model can continuously learn from new samples to adapt to changes in species characteristics in different scenarios, reducing misidentifications and missed identifications, and improving the model's generalization and accuracy.

[0094] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the bird feeder-based driving and feeding methods of some embodiments of the present disclosure, the accuracy of bird identification and bird feeder driving and feeding is improved. Specifically, the reason for the poor accuracy of bird identification and bird feeder driving and feeding is that only the camera module on the bird feeder is used to identify birds in the bird feeder feeding area. Single visual recognition is easily interfered by factors such as occlusion, lighting changes, and angle offset, resulting in poor bird identification accuracy. At the same time, when the confidence level of the bird identification result is low, auxiliary identification processing is not performed based on the preset species presence information corresponding to the geographical location information of the feeder area. Although the identification result (for example, bird or non-bird) is obtained, due to its low confidence level, the recognition error is large, resulting in poor accuracy of bird feeder driving and feeding. Based on this, the bird feeder-based driving and feeding methods of some embodiments of the present disclosure first collect multimodal monitoring data corresponding to the above-mentioned feeding area for the preset bird feeder. In this way, multimodal monitoring data of the feeding area can be collected. The multimodal detection data includes characteristic information of species in the feeding area captured from different dimensions. Then, based on the multimodal monitoring data, species identification processing is performed on the feeding area to obtain species identification information, wherein the species identification information includes identification species information and confidence information. Thus, species identification processing can be performed on the multimodal monitoring data including features from different dimensions, compensating for information loss caused by factors such as occlusion, illumination changes, and angle offset in single visual recognition, thereby improving the accuracy of species identification in the feeding area. Then, in response to determining that the identification species information represents the identification result as a bird identification and the confidence value represented by the confidence information is less than a preset value, auxiliary identification processing is performed based on the preset species presence information corresponding to the geographic location information of the area where the preset feeder is located, to obtain auxiliary identification species information. Thus, when the confidence level of the identification result is low, auxiliary identification processing can be performed based on the preset species presence information to obtain auxiliary identification species information. Thus, prior knowledge related to geographic location (i.e., preset species presence information) can be utilized to provide additional judgment basis for low-confidence identification results, effectively reducing identification errors. Next, based on the auxiliary species identification information, the preset bird feeder is controlled to execute the driving and feeding process. Thereafter, in response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, the preset bird feeder is controlled to execute the driving and feeding process based on the identified species information. This allows the bird feeder to be controlled to drive and feed birds based on identified species information with a small recognition error or a high confidence level, thereby improving the accuracy of the bird feeder's driving and feeding process.

[0095] Further references Figure 2, which shows a process 200 of another embodiment of the method for driving and feeding birds based on a bird feeder. The process 200 of the method for driving and feeding birds based on a bird feeder includes the following steps:

[0096] Step 201 : For a feeding area of ​​a preset bird feeder, multimodal monitoring data corresponding to the feeding area is collected.

[0097] In some embodiments, the execution entity (eg, a computing device) of the bird feeder-based driving and feeding method may collect multimodal monitoring data corresponding to a preset feeding area of ​​the bird feeder.

[0098] Step 202: Generate visual species identification information, voiceprint species identification information, and gravity vibration species identification information based on the multimodal monitoring data.

[0099] In some embodiments, the execution entity may generate visual species identification information, voiceprint species identification information, and gravity vibration species identification information based on the multimodal monitoring data.

[0100] Step 203: Generate species identification information based on the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information.

[0101] In some embodiments, the execution entity may generate species identification information based on the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information.

[0102] In step 204 , in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, auxiliary identification processing is performed based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information.

[0103] In some embodiments, the above-mentioned execution entity can, in response to determining that the above-mentioned identification species information represents an identification result as a bird identification and the confidence value represented by the above-mentioned confidence information is less than a preset value, perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the above-mentioned preset feeder is located to obtain auxiliary identification species information.

[0104] Step 205: Based on the auxiliary species identification information, control the preset bird feeder to perform driving and feeding processing.

[0105] In some embodiments, the execution entity may control the preset bird feeder to perform driving away and feeding processes based on the auxiliary species identification information.

[0106] Step 206 : In response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, based on the identified species information, controlling the preset bird feeder to perform driving and feeding processing.

[0107] In some embodiments, the execution entity may control the preset bird feeder to perform driving and feeding processing based on the identified species information in response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value.

[0108] from Figure 2 It can be seen that Figure 1 Compared with the description of some corresponding embodiments, Figure 2 In some embodiments, the process 200 of the bird feeder-based repelling and feeding method utilizes multimodal monitoring data to generate visual species identification information, soundprint species identification information, and gravity vibration species identification information. Species identification is performed based on visual, soundprint, and gravity vibration information. Species identification information is generated based on the visual, soundprint, and gravity vibration information, integrating identification results from multiple dimensions and improving species identification accuracy.

[0109] Further references Figure 3 As an implementation of the methods shown in the figures, the present disclosure provides some embodiments of a bird feeder-based driving and feeding device. Figure 1 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0110] like Figure 3As shown, the bird feeder-based driving and feeding device 300 of some embodiments includes: a collection unit 301 , an identification processing unit 302 , an auxiliary identification processing unit 303 , a first control unit 304 and a second control unit 305 . Among them, the collection unit 301 is configured to collect multimodal monitoring data corresponding to the feeding area of ​​the preset bird feeder; the identification processing unit 302 is configured to perform species identification processing in the feeding area based on the above multimodal monitoring data to obtain species identification information, wherein the above species identification information includes identification species information and confidence information; the auxiliary identification processing unit 303 is configured to, in response to determining that the identification result represented by the above identification species information is a bird identification and the confidence value represented by the above confidence information is less than a preset value, perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the above preset feeder is located to obtain auxiliary identification species information; the first control unit 304 is configured to control the above preset bird feeder to perform driving and feeding processing based on the above auxiliary identification species information; the second control unit 305 is configured to control the above preset bird feeder to perform driving and feeding processing based on the above identification species information in response to determining that the confidence value represented by the above confidence information is greater than or equal to the preset value.

[0111] It is understood that the units described in the device 300 are similar to those in the reference Figure 1 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the device 300 and the units included therein, and will not be repeated here.

[0112] Reference below Figure 4 , which shows a structural diagram of an electronic device 400 suitable for implementing some embodiments of the present disclosure. Figure 4 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0113] like Figure 4 As shown, the electronic device 400 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 into a random access memory (RAM) 403. Various programs and data required for the operation of the electronic device 400 are also stored in the RAM 403. The processing device 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0114] Typically, the following devices may be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 408 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 409. The communication device 409 may allow the electronic device 400 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 400 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 4 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0115] In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from a network via the communication device 409, or installed from the storage device 408, or installed from the ROM 402. When the computer program is executed by the processing device 401, the functions defined in the methods of some embodiments of the present disclosure are performed.

[0116] It should be noted that the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In some embodiments of the present disclosure, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device. In some embodiments of the present disclosure, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0117] In some embodiments, the client and server can communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0118] The computer-readable medium may be included in the electronic device, or may exist independently and not be incorporated into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: collect multimodal monitoring data corresponding to the feeding area of ​​a preset bird feeder; perform species identification processing for the feeding area based on the multimodal monitoring data to obtain species identification information, wherein the species identification information includes identification species information and confidence information; in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, perform auxiliary identification processing based on the presence information of the preset species corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information; based on the auxiliary identification species information, control the preset bird feeder to execute driving and feeding processing; in response to determining that the confidence value represented by the confidence information is greater than or equal to the preset value, control the preset bird feeder to execute driving and feeding processing based on the identification species information.

[0119] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0120] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0121] The units described in some embodiments of the present disclosure may be implemented in software or in hardware. The described units may also be provided in a processor. For example, they may be described as follows: a processor including an acquisition unit, an identification processing unit, an auxiliary identification processing unit, a first control unit, and a second control unit. The names of these units do not, in some cases, constitute limitations on the units themselves. For example, the acquisition unit may also be described as a "unit for collecting multimodal monitoring data corresponding to a feeding area of ​​a preset bird feeder."

[0122] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.

[0123] The above descriptions are merely some preferred embodiments of the present disclosure and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of technical features, but should also encompass other technical solutions formed by any combination of technical features or their equivalents without departing from the inventive concept. For example, a technical solution formed by replacing a feature with a technical feature having similar functions as disclosed in the embodiments of the present disclosure (but not limited to) can be formed.

Claims

1. A method for driving away and feeding birds based on a bird feeder, comprising: For a feeding area of ​​a preset bird feeder, collecting multimodal monitoring data corresponding to the feeding area; Performing species identification processing on the feeding area based on the multimodal monitoring data to obtain species identification information, wherein the species identification information includes identification species information and confidence information; In response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, performing auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located to obtain auxiliary identification species information; Based on the auxiliary species identification information, controlling the preset bird feeder to perform driving and feeding processing; In response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, the preset bird feeder is controlled to perform driving and feeding processing based on the identified species information.

2. The method according to claim 1, wherein The collecting of multimodal monitoring data corresponding to the feeding area of ​​the preset bird feeder includes: After a preset time period of the bird feeder being turned on, a regional image of the feeding area is collected, and image change detection processing is performed on the regional image based on a pre-created background model of the feeding area to obtain image detection information; In response to determining that the image detection information indicates an image change or detecting that the sound trigger device emits an abnormal sound signal or detecting that the weight trigger device emits an abnormal gravity signal, controlling a camera device, a sound collection device, and a gravity vibration collection device located in the feeding area to collect visual data, sound data, and gravity vibration data; The collected visual data, sound data and gravity vibration data are determined as multimodal monitoring data corresponding to the feeding area.

3. The method according to claim 1, wherein The processing of species identification in the feeding area based on the multimodal monitoring data to obtain species identification information includes: generating visual species identification information, voiceprint species identification information, and gravity vibration species identification information based on the multimodal monitoring data; Species identification information is generated based on the visual species identification information, the voiceprint species identification information, and the gravity vibration species identification information.

4. The method according to claim 1, wherein The preset species appearance information includes a species appearance time probability model, and the auxiliary identification processing is performed on the preset species appearance information corresponding to the geographical location information of the area where the preset feeder is located to obtain the auxiliary identification species information, including: The time for collecting the multimodal monitoring data is determined as the collection time; Obtaining a list of candidate species corresponding to the geographic location information; Based on the collection time, the candidate species list and the species appearance time probability model, generating each appearance species identifier and each appearance probability information corresponding to the collection time; Based on the identification of each appearing species and the information on each appearing probability, auxiliary species identification information is generated.

5. The method according to claim 1, wherein The auxiliary species identification information includes an identification of the species that has appeared, and the controlling of the preset bird feeder to perform driving and feeding processing based on the auxiliary species identification information includes: In response to determining that the infested species identifier included in the auxiliary species identification information is a bird identifier, driving the preset bird feeder to open a feeding bin and generating feeding bin opening record information; In response to determining that the infested species identifier included in the auxiliary species identification information is a non-bird identifier, the following driving-away steps are performed: controlling a sound wave driving device installed on the preset bird feeder to play sound waves of a preset frequency; Controlling the vibration driving device installed on the preset bird feeder to start; controlling a light flashing repelling device installed on the preset bird feeder to flash; generating driving record information corresponding to the sound wave driving device, the vibration driving device, and the light flashing driving device; The generated food bin opening record information and the expulsion record information are sent to a preset user terminal.

6. The method according to claim 1, wherein The identified species information includes an identified species identifier, and the controlling the preset bird feeder to perform driving and feeding processing based on the identified species information includes: In response to determining that the species identification included in the species identification information is a bird identification, driving the preset bird feeder to open a feeding bin, and generating feeding bin opening record information; In response to determining that the species identification included in the species identification information is a non-bird identification, the following driving-away step is performed: controlling a sound wave driving device installed on the preset bird feeder to play sound waves of a preset frequency; Controlling the vibration driving device installed on the preset bird feeder to start; controlling a light flashing repelling device installed on the preset bird feeder to flash; generating driving record information corresponding to the sound wave driving device, the vibration driving device, and the light flashing driving device; The generated food bin opening record information and the expulsion record information are sent to a preset user terminal.

7. The method according to claim 3, wherein: The generating of visual species identification information, voiceprint species identification information, and gravity vibration species identification information based on the multimodal monitoring data includes: Using a pre-trained species recognition model, the visual data included in the multimodal monitoring data is subjected to recognition detection processing to obtain visual species recognition information; Performing voiceprint feature recognition processing on the sound data included in the multimodal monitoring data to obtain voiceprint species identification information; Gravity vibration species identification information is generated based on the gravity vibration data included in the multimodal monitoring data.

8. A bird feeder-based driving and feeding device, comprising: a collecting unit configured to collect multimodal monitoring data corresponding to a feeding area of ​​a preset bird feeder; an identification processing unit configured to perform species identification processing in the feeding area based on the multimodal monitoring data to obtain species identification information, wherein the species identification information includes identification species information and confidence information; an auxiliary identification processing unit configured to, in response to determining that the identification result represented by the identification species information is a bird identification and the confidence value represented by the confidence information is less than a preset value, perform auxiliary identification processing based on the preset species occurrence information corresponding to the geographical location information of the area where the preset feeder is located, to obtain auxiliary identification species information; a first control unit configured to control the preset bird feeder to perform driving and feeding processing based on the auxiliary species identification information; The second control unit is configured to, in response to determining that the confidence value represented by the confidence information is greater than or equal to a preset value, control the preset bird feeder to perform driving and feeding processing based on the identified species information.

9. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images