Bird identification method, device, equipment and storage medium

Abstract

The present invention relates to the field of biometric identification technology, and more particularly to a bird identification method, apparatus, device, and storage medium. The method comprises: extracting bird foreground features and environmental background features from an image of a bird to be identified; determining a preset segmentation ratio parameter based on the bird foreground features and environmental background features; performing image segmentation on the image of the bird to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image; extracting features from the head image, torso image, and leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information; and identifying the species of the bird to be identified based on the head feature information, torso feature information, and leg feature information. Because the present invention performs identification after image segmentation, the present invention improves identification accuracy compared to existing methods that perform overall identification of an entire bird image.

Description

Technical Field

[0001] The present invention relates to the field of biometric identification technology, and in particular to a bird identification method, device, equipment and storage medium. Background Art

[0002] At present, with the progress and development of human society, human beings’ awareness of protecting the natural ecological environment is constantly increasing. Birds, as friends of mankind, have been listed as protected objects. Therefore, the study of birds is of great significance.

[0003] In the existing bird identification, drones are used to take bird images and perform overall identification of the entire bird image. Since overall identification involves a lot of image information, the identification accuracy is low.

[0004] The above content is only used to assist in understanding the technical solution of the present invention and does not constitute an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of the present invention is to provide a bird identification method, device, equipment and storage medium, aiming to solve the technical problem of low accuracy in overall identification of an entire bird image in the prior art.

[0006] To achieve the above object, the present invention provides a bird identification method, which comprises the following steps:

[0007] Extract bird foreground features and environmental background features from the bird image to be identified;

[0008] Determining a preset segmentation ratio parameter according to the bird foreground characteristics and the environmental background characteristics;

[0009] Performing image segmentation on the bird image to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image;

[0010] Performing feature extraction on the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information;

[0011] The species of the bird to be identified is identified according to the head feature information, the torso feature information and the leg feature information.

[0012] Optionally, the step of determining a preset segmentation ratio parameter according to the bird foreground features and the environmental background features includes:

[0013] Performing a first rough identification of the birds to be identified based on the bird foreground features to obtain a first category set;

[0014] performing a second rough identification on the birds to be identified according to the background environment characteristics, and screening the first category set based on the identification result to obtain a second category set;

[0015] A preset mapping relationship table is searched according to the second category set, and a preset segmentation ratio parameter is determined according to the search result.

[0016] Optionally, after the step of searching a preset mapping relationship table according to the second category set and determining a preset segmentation ratio parameter according to the search result, the method further includes:

[0017] Collecting environmental sounds from the habitat of the bird to be identified, and preprocessing the environmental sounds to obtain processed environmental sounds;

[0018] Performing sound extraction on the processed environmental sound based on the Mel-frequency cepstrum method to obtain bird sounds;

[0019] Determining a rough type of the bird to be identified based on the bird sounds, and adjusting the preset segmentation ratio parameter based on the rough type;

[0020] Accordingly, the step of performing image segmentation on the bird image to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image includes:

[0021] The image of the bird to be identified is segmented based on the adjusted preset segmentation ratio parameters to obtain a head image, a torso image, and a leg image.

[0022] Optionally, after the step of identifying the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information, the method further includes:

[0023] Divide the habitat of the bird to be identified into regions and determine the location for placing food;

[0024] Determining target food according to the habitat of the bird to be identified, and placing the target food at the food placement location according to a preset food placement amount;

[0025] The current remaining food amount of the food delivery location is obtained, and the number of birds in the to-be-identified bird habitat is counted based on the current remaining food amount and the preset food delivery amount.

[0026] Optionally, the step of determining target food according to the habitat of the bird to be identified includes:

[0027] Acquiring corresponding geographical location information and climate information according to the habitat of the bird to be identified;

[0028] determining a target bird species in the bird habitat to be identified based on the geographical location information and the climate information;

[0029] A target food type is determined according to the target bird type, and a target food is determined based on the target food type.

[0030] Optionally, after the step of obtaining the current remaining amount of food at the food delivery location and counting the number of birds in the to-be-identified bird habitat based on the current remaining amount of food and the preset food delivery amount, the method further includes:

[0031] Obtaining historical bird population information corresponding to the habitat of the bird to be identified;

[0032] Determining an evolution trend of the habitat of the bird to be identified based on the historical bird population information and statistical results;

[0033] A corresponding maintenance strategy is generated according to the evolution trend.

[0034] Optionally, after the step of identifying the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information, the method further includes:

[0035] Acquiring a flight path of the bird to be identified, and determining a current nest location of the bird to be identified based on the flight path;

[0036] Collecting a current bird's nest image corresponding to the current bird's nest position, and determining a structure type according to the current bird's nest image;

[0037] The recognition result is verified based on the structure type.

[0038] In addition, to achieve the above-mentioned purpose, the present invention further provides a bird identification device, comprising:

[0039] A feature extraction module is used to extract bird foreground features and environmental background features in the bird image to be identified;

[0040] A parameter determination module, configured to determine a preset segmentation ratio parameter based on the bird foreground characteristics and the environmental background characteristics;

[0041] An image segmentation module is used to segment the image of the bird to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image;

[0042] an information extraction module, configured to extract features from the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information;

[0043] The species identification module is used to identify the species of the bird to be identified based on the head feature information, the torso feature information and the leg feature information.

[0044] In addition, to achieve the above-mentioned purpose, the present invention also proposes a bird identification device, which includes: a memory, a processor, and a bird identification program stored in the memory and executable on the processor, wherein the bird identification program is configured to implement the steps of the bird identification method described above.

[0045] In addition, to achieve the above-mentioned purpose, the present invention further proposes a storage medium, on which a bird identification program is stored. When the bird identification program is executed by a processor, the steps of the bird identification method described above are implemented.

[0046] The present invention extracts bird foreground features and environmental background features from a bird image to be identified; determines a preset segmentation ratio parameter based on the bird foreground features and the environmental background features; performs image segmentation on the bird image to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image; extracts features from the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information; and identifies the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information. Because the present invention determines the preset segmentation ratio parameter based on the bird foreground features and environmental background features in the captured bird image, then performs image segmentation on the bird image based on the preset segmentation ratio parameter, and finally performs species identification based on the head image, torso image, and leg image obtained by segmentation, compared to existing methods of identifying the entire bird image as a whole, the present invention can perform identification after image segmentation, thereby improving identification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of the structure of a bird identification device in the hardware operating environment involved in an embodiment of the present invention;

[0048] Figure 2 This is a flow chart of a first embodiment of the bird identification method of the present invention;

[0049] Figure 3 This is a flow chart of a second embodiment of the bird identification method of the present invention;

[0050] Figure 4 This is a flow chart of a third embodiment of the bird identification method of the present invention;

[0051] Figure 5 This is a structural block diagram of the first embodiment of the bird identification device of the present invention.

[0052] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a bird identification device in the hardware operating environment involved in an embodiment of the present invention.

[0055] like Figure 1 As shown, the bird identification device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0056] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation of the bird identification device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0057] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a bird identification program.

[0058] exist Figure 1In the bird identification device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the bird identification device of the present invention can be set in the bird identification device. The bird identification device calls the bird identification program stored in the memory 1005 through the processor 1001 and executes the bird identification method provided by the embodiment of the present invention.

[0059] The embodiment of the present invention provides a bird identification method, referring to Figure 2 , Figure 2 Schematic diagram of the flow of the first embodiment of the bird identification method of the present invention.

[0060] In this embodiment, the bird identification method includes the following steps:

[0061] Step S10: extracting bird foreground features and environmental background features in the bird image to be identified.

[0062] It should be noted that the method of this embodiment can be applied in scenarios where bird identification is required, or in other scenarios where identification is required. The execution subject of this embodiment can be a bird identification device with data processing, network communication, and program execution functions, such as a bird detection device, or other devices that can perform the same or similar functions. This embodiment and the following embodiments are specifically described using the above-mentioned bird identification device (hereinafter referred to as the device).

[0063] It is understandable that the above-mentioned bird image to be identified may be an image containing a bird, which may be obtained by shooting with a camera or by other means, and this embodiment does not limit this.

[0064] It should be understood that the above-mentioned bird image to be identified may be composed of a bird foreground image and an environmental background image. The above-mentioned bird foreground image may include bird foreground features, that is, the above-mentioned bird foreground image only includes the bird part image in the bird image to be identified. The above-mentioned environmental background image may include background environment features, that is, the above-mentioned environmental background image may include part of the image other than the bird image part, for example, background environments such as the sky and the woods, which are not limited in this embodiment.

[0065] It should be emphasized that the method of extracting the above-mentioned bird foreground features and environmental background features from the above-mentioned bird images to be identified can be to extract them through the region of interest method, that is, to use the bird foreground features in the bird images to be identified as the region of interest, or to extract them through other methods. The specific extraction method can be set according to actual conditions.

[0066] In a specific implementation, the above-mentioned device may first obtain an image of the bird to be identified taken by a camera, and extract the bird foreground features and background environment features from the above-mentioned image of the bird to be identified.

[0067] Step S20: determining a preset segmentation ratio parameter according to the bird foreground features and the environmental background features.

[0068] It should be noted that the above-mentioned preset segmentation ratio parameters can be parameters used to perform image segmentation on the above-mentioned bird images to be identified. In this embodiment, since the body structures and sizes of different birds are different, the segmentation ratio parameters corresponding to different birds can be stored in the above-mentioned device.

[0069] In a specific implementation, the above-mentioned device can roughly identify the birds to be identified in the above-mentioned bird foreground features through bird foreground features and background environment features, determine the approximate species, and then select the corresponding preset segmentation ratio parameters from the mapping relationship table according to the determined species.

[0070] Step S30: performing image segmentation on the bird image to be identified based on the preset segmentation ratio parameters to obtain a head image, a torso image, and a leg image.

[0071] It is understandable that the head image may be a partial image of the head of the bird to be identified, the torso image may be a partial image of the torso of the bird to be identified, and the leg image may be a partial image of the leg of the bird to be identified.

[0072] It should be emphasized that due to the shooting angle, a part of the image of the bird foreground feature may not be captured completely. For example, due to the shooting angle, the bird foreground image a1 of the bird A to be identified may not contain a leg image, or the legs of the bird A to be identified may be blocked by obstacles, which may affect the image segmentation results.

[0073] In this embodiment, the bird A to be identified may be tracked and photographed. For example, a leg image may be present in the bird foreground image a2 obtained after tracking and photographing the bird A to be identified. The leg image in the bird foreground image a2 may be used as a segmentation result.

[0074] In a specific implementation, the above-mentioned device can perform image segmentation on the bird image to be identified based on preset segmentation ratio parameters to obtain a head image, a torso image and a leg image.

[0075] Step S40: extracting features from the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information.

[0076] It should be understood that the above-mentioned preset feature recognition models can be divided into a head feature recognition model, a torso feature recognition model and a leg feature recognition model. The above-mentioned preset feature recognition models can be obtained by training based on bird feature samples. The above-mentioned bird feature samples contain corresponding labels. The above-mentioned preset feature recognition models can be obtained by training through bird feature samples and corresponding labels.

[0077] It should be noted that the above-mentioned head feature information may include head color feature information, eye feature information, mouth feature information and other information used for identification; the above-mentioned torso feature information may include torso color feature information, tail feature information and other information used for identification; the above-mentioned leg feature information may include leg structure feature information, leg length feature information, leg color feature information and other information used for identification. The specific feature information mentioned above is not limited in this embodiment.

[0078] In a specific implementation, the above-mentioned device can extract features of the head image, torso image and leg image through a preset adjustment recognition model to obtain corresponding head feature information, torso feature information and leg feature information.

[0079] Step S50: Identify the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information.

[0080] It is understandable that the above-mentioned device may store a database of bird species corresponding to each characteristic information. Since there may be multiple bird species for a certain characteristic information, in this embodiment, the above-mentioned device can filter from the database according to the head characteristic information to obtain a first category set corresponding to the head characteristic information, filter from the database according to the torso characteristic information to obtain a second category set corresponding to the torso characteristic information, filter from the database according to the leg characteristic information to obtain a third category set corresponding to the leg characteristic information, and then find the intersection of the first category set, the second category set and the third category set to obtain the species of the bird to be identified.

[0081] In a specific implementation, the above-mentioned device can obtain a first category set based on the head feature information, obtain a second category set based on the torso feature information, and obtain a third category set based on the leg feature information, and then perform category identification on the bird to be identified based on the first category set, the second category set, and the third category set.

[0082] Furthermore, in order to verify the recognition result and further improve the recognition accuracy, in this embodiment, after the above step S50, the following steps are further included:

[0083] Step S51: obtaining the flight path of the bird to be identified, and determining the current nest location of the bird to be identified based on the flight path;

[0084] Step S52: collecting a current bird's nest image corresponding to the current bird's nest position, and determining a structure type according to the current bird's nest image.

[0085] It should be noted that since the bird nests built by different birds are inconsistent, the structural types may be inconsistent. The above structural types may include the structure of the bird nest and the building materials, etc. For example, some birds can use materials such as branches and feathers to build, and some birds can use materials such as grass and leaves to build. Therefore, in this embodiment, the above recognition results can be verified based on the structural type of the bird nest to further improve the recognition accuracy.

[0086] In a specific implementation, the above-mentioned device can track and shoot the bird to be identified, determine the flight path of the bird based on the tracking and shooting, shoot the bird's nest of the bird to be identified based on the flight path, obtain the current bird's nest image, and then determine the structure type based on the current bird's nest image.

[0087] Step S53: verifying the recognition result based on the structure type.

[0088] It should be emphasized that if the bird species determined based on the structural type is inconsistent with the bird species in the identification result, the above-mentioned device can obtain the structural type of the bird's nest structure of the bird species in the identification result, and compare the similarity between the structural type of the bird's nest structure of the bird species in the identification result and the structural type determined by the current bird's nest image. If the similarity exceeds the preset threshold, the identification result can be determined to be correct. For ease of understanding, for example, the bird species determined according to the identification result is Class C, and the bird species determined according to the structural type is Class D, then the above-mentioned device can determine whether the structural type of the bird's nest of Class D bird and the structural type of the bird's nest of Class C bird exceed the preset threshold. If so, the species of the bird to be identified can be determined to be Class C.

[0089] The above-mentioned device of this embodiment can first obtain the bird image to be identified taken by the camera, and extract the bird foreground features and background environment features from the above-mentioned bird image to be identified; roughly identify the bird to be identified in the above-mentioned bird foreground features through the bird foreground features and background environment features, determine the approximate species, and then select the corresponding preset segmentation ratio parameters from the mapping relationship table according to the determined species; perform image segmentation on the bird image to be identified based on the preset segmentation ratio parameters to obtain the head image, torso image and leg image; perform feature extraction on the head image, torso image and leg image through the preset adjustment recognition model to obtain the corresponding head feature information, torso feature information and leg feature information; obtain the corresponding head feature information according to the head feature information A first category set is obtained, a second category set is obtained based on the torso feature information, a third category set is obtained based on the leg feature information, and then the species of the bird to be identified is identified based on the first category set, the second category set and the third category set; compared with the existing overall identification of an entire bird image, this embodiment can segment the image and then perform identification, thereby improving the accuracy of identification; further, this embodiment can also track and shoot the bird to be identified, determine the flight path of the bird based on the tracking shooting, shoot the bird's nest to be identified based on the flight path, obtain the current bird's nest image, and then determine the structure type based on the current bird's nest image, and finally verify the identification result based on the structure type, thereby further improving the recognition accuracy.

[0090] refer to Figure 3 , Figure 3 FIG. 4 is a flow chart of the second embodiment of the bird identification method of the present invention.

[0091] In order to accurately obtain the preset segmentation ratio parameters, such as Figure 3 As shown, in this embodiment, the above step S20 includes:

[0092] Step S21: performing a first rough identification on the birds to be identified based on the bird foreground features to obtain a first category set.

[0093] It should be noted that the first rough recognition may be performed based on obvious features in the foreground features of the bird. The first rough recognition is fast but not accurate.

[0094] It is understandable that the obvious features mentioned above may be features with a larger image area or higher definition. The first rough recognition mentioned above may be implemented based on a relatively simple recognition algorithm, and the specific algorithm is not limited in this embodiment.

[0095] In a specific implementation, the above-mentioned device can perform a first rough identification on the birds to be identified according to the bird foreground features, determine the bird species that may be involved, and use them as the first category set.

[0096] Step S22: performing a second rough identification on the birds to be identified according to the background environment characteristics, and filtering the first category set based on the identification result to obtain a second category set.

[0097] It should be understood that the second rough identification can be roughly identified based on obvious features in the background environment features. Since different types of birds live in different environments, auxiliary judgment can be made based on the background environment features.

[0098] It should be noted that the above second coarse recognition can also be implemented based on a relatively simple recognition algorithm.

[0099] In a specific implementation, the above-mentioned device can perform a second coarse recognition on the birds to be identified according to the background environment characteristics, and filter the above-mentioned first category set according to the recognition result to obtain the second category set.

[0100] Step S23: searching a preset mapping relationship table according to the second category set, and determining a preset segmentation ratio parameter according to the search result.

[0101] In this embodiment, a mapping relationship table between the second type set and the corresponding preset segmentation ratio parameters may be preset in the above device, and the corresponding preset segmentation ratio parameters may be determined according to the second type set.

[0102] Furthermore, since both the first rough recognition and the second rough recognition are rough recognition, in order to determine the accuracy of the preset segmentation ratio parameter, in this embodiment, after the above step S23, the following steps are further included:

[0103] Step S24: collecting environmental sounds of the habitat of the bird to be identified, and preprocessing the environmental sounds to obtain processed environmental sounds.

[0104] It is understandable that the habitat of the bird to be identified may be the current location of the bird to be identified, which can be obtained by a sound collection module carried by a drone.

[0105] It should be understood that, since the above-mentioned environmental sound contains more noise, the above-mentioned preprocessing may include a series of steps such as pre-emphasis, framing and windowing to eliminate noise such as aliasing and high-order harmonic distortion caused by other factors.

[0106] In a specific implementation, the above-mentioned device can collect the environmental sounds of the habitat of the bird to be identified, pre-process the environmental sounds, and use the sounds after removing noise as the processed environmental sounds.

[0107] Step S25: performing sound extraction on the processed environmental sound based on the Mel-frequency cepstrum method to obtain bird sounds.

[0108] It should be noted that the above-mentioned Mel-frequency cepstrum method may be a method based on the linear change of the nonlinear Mel-scale logarithmic energy spectrum of the sound frequency, and may be used for sound extraction.

[0109] Step S26: determining a rough type of the bird to be identified based on the bird sounds, and adjusting the preset segmentation ratio parameters according to the rough type.

[0110] It is understandable that the above-mentioned device may store sound data corresponding to various types of birds. By comparing the extracted sound with the sound data, a rough type of the bird to be identified can be obtained.

[0111] Accordingly, the above step S30 includes:

[0112] Step S30 ′: performing image segmentation on the bird image to be identified based on the adjusted preset segmentation ratio parameters to obtain a head image, a torso image, and a leg image.

[0113] In a specific implementation, the above-mentioned device can perform sound extraction on the processed environmental sound based on the Mel-frequency cepstrum method to obtain the bird sound of the bird to be identified, and determine the rough type based on the bird sound. According to the rough type, the corresponding segmentation ratio parameter is searched in the preset mapping relationship table to adjust the preset segmentation ratio parameters obtained according to the bird foreground characteristics and background environment characteristics. Finally, the image is segmented using the adjusted preset segmentation ratio parameters, thereby improving the accuracy of the preset segmentation ratio parameters.

[0114] In this embodiment, the above-mentioned device can perform a first rough identification of the birds to be identified based on the foreground characteristics of the birds, determine the possible bird species involved, and use them as a first category set; perform a second rough identification of the birds to be identified based on the background environment characteristics, and filter the above-mentioned first category set according to the identification results to obtain a second category set; search the preset mapping relationship table according to the second category set, and determine the preset segmentation ratio parameters based on the search results; at the same time, collect the environmental sounds of the habitat of the birds to be identified, pre-process the environmental sounds, and use the sounds after removing noise as the processed environmental sounds; adjust the preset segmentation ratio parameters based on the environmental sounds to improve the accuracy of the preset segmentation ratio parameters, thereby improving the accuracy of the image segmentation.

[0115] refer to Figure 4 , Figure 4 2 is a flow chart of the third embodiment of the bird identification method of the present invention.

[0116] In order to facilitate the counting of bird numbers and facilitate subsequent management by staff, Figure 4 As shown, in this embodiment, after the above step S50, the following steps are further included:

[0117] Step S60: Divide the habitat of the bird to be identified into regions and determine the location for placing food.

[0118] It should be noted that the above-mentioned area division can be performed according to a preset density, or according to the number of appearances of the birds to be identified. For example, if the number of appearances of the birds to be identified in a certain area is relatively large, the area can be divided into a food placement location.

[0119] Step S70: determining target food according to the habitat of the bird to be identified, and placing the target food at the food placement location according to a preset food placement amount.

[0120] It is understandable that, since birds of the same species eat different foods due to differences in their habitats, the above-mentioned device can use the food that is edible by birds in the habitat of the bird to be identified as the target food.

[0121] It should be understood that the above-mentioned preset food delivery amount can be set according to actual conditions and is not limited in this embodiment.

[0122] Step S80: obtaining the current remaining food amount at the food delivery location, and counting the number of birds in the to-be-identified bird habitat based on the current remaining food amount and the preset food delivery amount.

[0123] It should be explained that the above-mentioned device can calculate the amount of food reduced in one day after the food is released in units of half a day or a day, and count the number of birds in the habitat of the bird to be identified based on the amount of food reduced.

[0124] In a specific implementation, the above device can place target food at the food placement location, calculate the amount of food reduction, and determine the number of birds based on the amount of food reduction.

[0125] Furthermore, in order to further determine the target food type so as to improve the accuracy of subsequent statistics, in this embodiment, the step of determining the target food based on the habitat of the bird to be identified includes:

[0126] Obtain corresponding geographic location information and climate information based on the bird habitat to be identified; determine the target bird species of the bird habitat to be identified based on the geographic location information and the climate information; determine the target food species based on the target bird species, and determine the target food based on the target food species.

[0127] It should be noted that the above-mentioned target bird species may be bird species that may appear in the above-mentioned bird habitats to be identified. The possible bird species can be determined based on the geographical location information and climate information of the bird habitats to be identified, thereby improving the accuracy of the target food.

[0128] Furthermore, in order to facilitate the subsequent maintenance of the habitat of the bird to be identified, in this embodiment, after the above step S80, the following steps are further included:

[0129] Acquire historical bird population information corresponding to the to-be-identified bird habitat; determine an evolution trend of the to-be-identified bird habitat based on the historical bird population information and statistical results; and generate a corresponding maintenance strategy according to the evolution trend.

[0130] It is understandable that the above-mentioned device can store historical bird population information within a preset time period, and the above-mentioned preset time period can be set according to actual conditions.

[0131] It should be understood that the above-mentioned historical bird population information may be historical population change information corresponding to different types of birds, and the above-mentioned device may determine the evolution trend of the habitat of the bird to be identified based on the population change information of various types of birds.

[0132] It should be noted that the above-mentioned maintenance strategy can be a strategy to return the evolution trend to normal. For example, in the above-mentioned evolution trend, the number of Class F birds gradually increases and exceeds the preset bird number threshold. Then the above-mentioned device can determine that the number of Class F birds needs to be reduced, and can generate a corresponding maintenance strategy to return the number of Class F birds to normal.

[0133] It should be emphasized that the above maintenance strategy may be a maintenance method such as manual capture, and this embodiment does not limit this.

[0134] The above-mentioned equipment in this embodiment can divide the habitat of the bird to be identified into areas, obtain the food placement location, determine the target bird species of all birds that may appear in the habitat to be identified based on the geographical location information and climate information of the habitat of the bird to be identified, and then determine the target food species based on the target bird species, and determine the target food for placement based on the target food species; record the current remaining food amount, and determine the number of birds in the habitat based on the preset food placement amount and the current remaining food amount, so as to facilitate subsequent management and maintenance by staff; at the same time, the above-mentioned equipment can also determine the evolution trend of the habitat to be identified based on historical bird population information and statistical results, and generate corresponding maintenance strategies in time for maintenance, thereby improving maintenance efficiency.

[0135] In addition, an embodiment of the present invention further provides a storage medium, on which a bird identification program is stored. When the bird identification program is executed by a processor, the steps of the bird identification method described above are implemented.

[0136] In addition, refer to Figure 5 , Figure 5 This is a structural block diagram of a first embodiment of a bird identification device according to the present invention. The present invention further provides a bird identification device, comprising:

[0137] A feature extraction module 501 is used to extract bird foreground features and environmental background features in the bird image to be identified;

[0138] A parameter determination module 502 is configured to determine a preset segmentation ratio parameter based on the bird foreground characteristics and the environmental background characteristics;

[0139] An image segmentation module 503 is configured to segment the image of the bird to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image;

[0140] An information extraction module 504 is configured to extract features from the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information;

[0141] The species identification module 505 is configured to identify the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information.

[0142] The above-mentioned device of this embodiment can first obtain the bird image to be identified taken by the camera, and extract the bird foreground features and background environment features from the above-mentioned bird image to be identified; roughly identify the bird to be identified in the above-mentioned bird foreground features through the bird foreground features and background environment features, determine the approximate species, and then select the corresponding preset segmentation ratio parameters from the mapping relationship table according to the determined species; perform image segmentation on the bird image to be identified based on the preset segmentation ratio parameters to obtain the head image, torso image and leg image; perform feature extraction on the head image, torso image and leg image through the preset adjustment recognition model to obtain the corresponding head feature information, torso feature information and leg feature information; obtain the corresponding head feature information according to the head feature information A first category set is obtained, a second category set is obtained based on the torso feature information, a third category set is obtained based on the leg feature information, and then the species of the bird to be identified is identified based on the first category set, the second category set and the third category set; compared with the existing overall identification of an entire bird image, this embodiment can segment the image and then perform identification, thereby improving the accuracy of identification; further, this embodiment can also track and shoot the bird to be identified, determine the flight path of the bird based on the tracking shooting, shoot the bird's nest to be identified based on the flight path, obtain the current bird's nest image, and then determine the structure type based on the current bird's nest image, and finally verify the identification result based on the structure type, thereby further improving the recognition accuracy.

[0143] Other embodiments or specific implementations of the bird identification device of the present invention can refer to the above-mentioned method embodiments and will not be described in detail here.

[0144] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0145] The serial numbers of the embodiments of the present invention are for description only and do not represent the advantages and disadvantages of the embodiments. Through the description of the above implementation modes, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0146] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A bird identification method, characterized in that: The method comprises the following steps: Extract bird foreground features and environmental background features from the bird image to be identified; Determining a preset segmentation ratio parameter according to the bird foreground characteristics and the environmental background characteristics; Performing image segmentation on the bird image to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image; Performing feature extraction on the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information; Identify the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information; The step of determining a preset segmentation ratio parameter according to the bird foreground feature and the environmental background feature includes: Performing a first rough identification of the birds to be identified based on the bird foreground features to obtain a first category set; performing a second rough identification on the birds to be identified according to the environmental background characteristics, and screening the first category set based on the identification result to obtain a second category set; Searching a preset mapping relationship table according to the second category set, and determining a preset segmentation ratio parameter according to the search result; After the step of searching the preset mapping relationship table according to the second category set and determining the preset segmentation ratio parameter according to the search result, the method further includes: Collecting environmental sounds from the habitat of the bird to be identified, and preprocessing the environmental sounds to obtain processed environmental sounds; Performing sound extraction on the processed environmental sound based on the Mel-frequency cepstrum method to obtain bird sounds; Determining a rough type of the bird to be identified based on the bird sounds, and adjusting the preset segmentation ratio parameter based on the rough type; Accordingly, the step of performing image segmentation on the bird image to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image includes: The image of the bird to be identified is segmented based on the adjusted preset segmentation ratio parameters to obtain a head image, a torso image, and a leg image.

2. The bird identification method according to claim 1, wherein: After the step of identifying the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information, the method further includes: Divide the habitat of the bird to be identified into regions and determine the location for placing food; Determining target food according to the habitat of the bird to be identified, and placing the target food at the food placement location according to a preset food placement amount; The current remaining amount of food at the food delivery location is obtained, and the number of birds in the to-be-identified bird habitat is counted based on the current remaining amount of food and the preset food delivery amount.

3. The bird identification method according to claim 2, wherein: The step of determining target food according to the habitat of the bird to be identified includes: Acquiring corresponding geographical location information and climate information according to the habitat of the bird to be identified; determining a target bird species in the bird habitat to be identified based on the geographical location information and the climate information; A target food type is determined according to the target bird type, and a target food is determined based on the target food type.

4. The bird identification method according to claim 2, wherein: After the step of obtaining the current remaining amount of food at the food placement location and counting the number of birds in the to-be-identified bird habitat based on the current remaining amount of food and the preset food placement amount, the method further includes: Obtaining historical bird population information corresponding to the habitat of the bird to be identified; Determining an evolution trend of the habitat of the bird to be identified based on the historical bird population information and statistical results; A corresponding maintenance strategy is generated according to the evolution trend.

5. The bird identification method according to any one of claims 1 to 4, characterized in that: After the step of identifying the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information, the method further includes: Acquiring a flight path of the bird to be identified, and determining a current nest location of the bird to be identified based on the flight path; Collecting a current bird's nest image corresponding to the current bird's nest position, and determining a structure type according to the current bird's nest image; The recognition result is verified based on the structure type.

6. A bird identification device, characterized in that: The device comprises: A feature extraction module is used to extract bird foreground features and environmental background features in the bird image to be identified; A parameter determination module, configured to determine a preset segmentation ratio parameter based on the bird foreground characteristics and the environmental background characteristics; An image segmentation module is used to segment the image of the bird to be identified based on the preset segmentation ratio parameter to obtain a head image, a torso image, and a leg image; an information extraction module, configured to extract features from the head image, the torso image, and the leg image using a preset feature recognition model to obtain corresponding head feature information, torso feature information, and leg feature information; a species identification module, configured to identify the species of the bird to be identified based on the head feature information, the torso feature information, and the leg feature information; The parameter determination module is further configured to perform a first coarse identification of the birds to be identified based on the bird foreground features to obtain a first category set; perform a second coarse identification of the birds to be identified based on the environmental background features, and filter the first category set based on the identification results to obtain a second category set; search a preset mapping relationship table based on the second category set, and determine a preset segmentation ratio parameter based on the search result; The parameter determination module is further configured to collect environmental sounds from the habitat of the bird to be identified, pre-process the environmental sounds to obtain processed environmental sounds, perform sound extraction on the processed environmental sounds based on the Mel-frequency cepstrum method to obtain bird sounds, determine the rough species of the bird to be identified based on the bird sounds, and adjust the preset segmentation ratio parameter based on the rough species; The image segmentation module is further configured to perform image segmentation on the bird image to be identified based on the adjusted preset segmentation ratio parameters to obtain a head image, a torso image, and a leg image.

7. A bird identification device, characterized in that: The device includes: a memory, a processor, and a bird identification program stored in the memory and executable on the processor, wherein the bird identification program is configured to implement the steps of the bird identification method according to any one of claims 1 to 5.

8. A storage medium, characterized in that: The storage medium stores a bird identification program, which, when executed by a processor, implements the steps of the bird identification method according to any one of claims 1 to 5.

Images