Bird management method and management system based on electric power image detection

Abstract

The invention relates to the technical field of bird management, in particular to a bird management method and management system based on electric power image detection, and aims to solve the problems of how to accurately identify various birds and how to manage the number of the birds in the bird attracting process. The embodiment of the invention provides a bird management method based on electric power image detection, and the method comprises the steps: arranging a plurality of image detectors on an island, carrying out the image detection of a target region around the island in a target time period, judging whether a target bird exists in a target group or not according to the similarity, and if yes, executing the step 1; if yes, specific audio is sent to the bird flock, target birds are attracted to go to the island, in the process that the target birds fly to the island, directional recognition is conducted on the target birds, video images in the island are periodically obtained, and the target number of the target birds in the island is calculated according to the directional recognition result; and adjusting the sending mode of the specific audio according to the target number.

Description

Technical Field

[0001] The present invention relates to the technical field of bird management, and in particular to a bird management method and management system based on power image detection. Background Art

[0002] Island bird detection methods focus on using various scientific and technological means to detect, record and study wild birds. They integrate multiple disciplines such as ecology, ornithology, computer science and environmental science, aiming to gain a more comprehensive understanding of bird behavior, migration patterns, ecological and environmental needs, etc. Traditional bird detection methods mainly rely on visual observation and manual recording by human observers, which is limited by the observer's subjective judgment and working status, and the observation time is also very limited. In addition, traditional detection methods often focus on the identification of bird species, and cannot take into account both bird identification and attraction, and do not manage the number of birds during the attraction process. Therefore, how to accurately identify various birds and manage their numbers in the process of attracting birds is one of the technical problems that urgently need to be solved in this field. Summary of the invention

[0003] The problem solved by the invention is: how to accurately identify various birds and manage their number in the process of attracting birds.

[0004] To solve the above problems, an embodiment of the present invention provides a bird management method based on power image detection, the management method comprising: setting multiple image detectors on an island, and performing image detection on a target area around the island within a target time period; when a flock of birds passes through the target area, obtaining video data of the flock of birds, and judging whether the flock of birds belongs to a target group based on the video data; when the flock of birds belongs to the target group, obtaining local features of each individual bird in the target group to obtain local feature results; calculating the similarity between each individual bird and the target bird based on the local feature results, and judging whether there is a target bird in the target group based on the similarity; if so, sending specific audio to the flock of birds to attract the target birds to the island; when the target birds are successfully attracted, performing direction identification on the target birds during the process of the target birds flying to the island to obtain the direction identification results of the target birds; periodically obtaining video images inside the island, and calculating the target number of target birds inside the island based on the direction identification results; adjusting the sending mode of the specific audio according to the target number.

[0005] Compared with the prior art, the technical effects achieved by adopting this technical solution are as follows: the setting of the target time period and the target area improves the working time and working range of the image detector and makes it more targeted. In the image detection process, there is no need to perform frame-by-frame image analysis on each individual bird, which improves the working efficiency of the image detector. The determination process of the target group fully considers the actual situation during the bird migration process, making the acquisition of the target group more reasonable. The similarity between each individual bird and the target bird is calculated based on the local feature results, fully considering the errors caused by video judgment, and improving the accuracy of target group identification. The directional identification process realizes the marking of each individual bird, and can accurately obtain the number of target birds inside the island. By setting the target number, it helps to control the target birds to develop to a number that is more suitable for their own survival.

[0006] In one embodiment of the present invention, when a flock of birds belongs to a target group, local features are acquired for each individual bird in the target group to obtain local feature results, specifically including: acquiring the body contour of each individual bird in the target group, and screening out the body feature points corresponding to the target group in the body contour; determining the local feature results of each individual bird according to the position of the body feature points.

[0007] Compared with the existing technology, the technical effect achieved by adopting this technical solution is as follows: the acquisition of the body contour of each individual bird fully considers the possible impact of the actual shooting environment, further improves the accuracy of the target group, and determines the local feature results according to the position of the body feature points, fully considering the situation that birds of the same family have relatively similar local feature points when flying, thereby improving the accuracy of target bird identification.

[0008] In one embodiment of the present invention, the body contour of each individual bird in a target group is acquired, and the body feature points corresponding to the target group are screened out from the body contour, specifically including: judging the flight path of the target group according to the flight mode of the target group, recording the image detector that captures the target group as the initial detector, and screening the supplementary detector corresponding to the initial detector according to the flight path; acquiring an initial image of the body contour according to the initial detector, and when the initial image is not clear, controlling the supplementary detector to change the shooting angle to acquire a supplementary image of the body contour; and determining the body feature points corresponding to the target group according to the initial image and the supplementary image.

[0009] Compared with the existing technology, the technical effect achieved by adopting this technical solution is as follows: the initial image and the supplementary image are obtained according to the initial detector and the supplementary detector, which fully considers the actual situation of the target group during flight, avoids unclear images caused by single-angle shooting, and reduces the probability of misjudgment due to incomplete shooting.

[0010] In one embodiment of the present invention, the similarity between each individual bird and the target bird is calculated based on the local feature results, and whether the target bird exists in the target group is determined based on the similarity, specifically including: when the similarity is greater than a first result threshold, marking the individual bird as the target bird; when the similarity is less than or equal to the first result threshold, not marking the individual bird; when the target bird appears, stopping the calculation of the similarity between the individual bird and the target bird.

[0011] Compared with the prior art, the technical effect achieved by adopting this technical solution is as follows: the setting of the first result threshold fully takes into account the error caused by the image detector, while reducing the missed detection rate of the target bird, the efficiency of image recognition is improved, and the calculation of similarity is stopped when the target bird appears, which reduces unnecessary recognition operations of the image detector and reduces invalid recognition workload.

[0012] In one embodiment of the present invention, if yes, specific audio is sent to the flock of birds to attract the target birds to the island, specifically including: when there are target birds in the target group, calculating the first distance for the target birds to reach the target position and the flying speed of the target birds; calculating the sending time and the decibel of the sending volume of the specific audio according to the first distance and the flight speed; calculating the duration of the impact of the specific audio on the target birds according to the flying speed of the target birds and the flight path of the target birds, and controlling the sending mode of the specific audio according to the impact duration.

[0013] Compared with the existing technology, the technical effect achieved by adopting this technical solution is as follows: the sending time and the decibel of the sending volume of the specific audio are calculated according to the first distance and the flight speed, the situations encountered by the sound in the actual propagation process are taken into account, the sending mode of the specific audio is controlled according to the impact duration, and the possible reactions of the target group when hearing the specific audio during the flight are fully considered, thereby increasing the possibility of attracting the target birds.

[0014] In one embodiment of the present invention, after the target bird is successfully attracted, the target bird is directionally identified during the process of the target bird flying to the island to obtain a direction identification result of the target bird, which specifically includes: after the target bird is successfully attracted, the shooting angle of the image detector is determined according to the flight posture of the target bird, and the captured image after the shooting angle is adjusted is recorded as the shooting result; the target bird is directionally identified according to the shooting result to obtain the direction identification result of the target bird, and the target bird is encoded according to the direction identification result.

[0015] Compared with the existing technology, the technical effect achieved by adopting this technical solution is as follows: the shooting angle of the image detector is determined according to the flight posture of the target bird, which enriches the details of the target bird's appearance features and makes directional identification easier. After the target bird is successfully attracted, the image detector can capture more details of the target bird in the process of flying to the island, and the target bird is encoded according to the directional identification results, making the subsequent management of the target bird more convenient.

[0016] In one embodiment of the present invention, video images of the interior of the island are periodically acquired, and the target number of target birds in the island is calculated based on the directional recognition results, specifically including: periodically performing detections in the target area of ​​the island based on the living habits of the target birds; detecting the target birds in the island, directional recognition of the uncoded target birds, and acquiring the coding of the directional recognition results, and obtaining the current number of target birds in the island through the coding; and obtaining the target number of target birds based on the ecological conditions of the island and the current number of target birds.

[0017] Compared with the existing technology, the technical effect achieved by adopting this technical solution is: periodically detecting target birds on the island in the target area on the island, which improves the accuracy of the number of target birds, is conducive to the dynamic detection of the number of target birds, and facilitates the subsequent management of the number of target birds.

[0018] In one embodiment of the present invention, the sending mode of specific audio is adjusted according to the target number, specifically including: when the difference between the existing number and the target number is less than or equal to a first threshold, the sending mode of the specific audio is not changed; when the difference between the existing number and the target number is greater than the first threshold, the sending mode of the specific audio is controlled according to the flight mode of the target group.

[0019] Compared with the existing technology, the technical effect achieved by adopting this technical solution is as follows: when the difference between the existing number and the target number is greater than the first threshold, the sending mode of the specific audio is controlled according to the flight pattern of the target group, which fully considers the impact of the target bird population and the subsequent possible increase in the number on the survival of the target birds, which is conducive to the ideal state of the target bird population on the island.

[0020] In one embodiment of the present invention, the present invention also provides a bird management system located on an island, and the management method described in the above embodiment is applied to the management system, and the management system includes: an image capture module: the image capture module is used to obtain video data of a flock of birds; an image recognition module: the image recognition module is used to perform directional identification of target birds; a calculation module: the calculation module is used to calculate the similarity between each individual bird and the target bird; an audio sending module: the audio sending module is used to adjust the sending mode of a specific audio. The management system has all the technical features of the above management method, which will not be repeated here one by one. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings to be used in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. Figure 1 This is one of the flow charts of the bird management method based on power image detection of the present invention; Figure 2 This is the second flow chart of the bird management method based on power image detection of the present invention; Figure 3 This is the third flow chart of the bird management method based on power image detection of the present invention; Figure 4 This is the fourth flow chart of the bird management method based on power image detection of the present invention; Figure 5 It is a system schematic diagram of the management system of the present invention.

[0022] Description of reference numerals: 100 - management system; 110 - image capturing module; 120 - image recognition module; 130 - computing module; 140 - audio sending module. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] [First embodiment] See also Figure 1 In a specific embodiment, the present invention provides an optical cable management method based on artificial intelligence, the management method comprising: S100, setting up multiple image detectors on the island, and performing image detection on a target area around the island within a target time period; S200, when a flock of birds passes through a target area, obtaining video data of the flock of birds, and determining whether the flock of birds belongs to the target group according to the video data; S300, when the bird flock belongs to the target group, local features are acquired for each individual bird in the target group to obtain local feature results; S400, calculating the similarity between each individual bird and the target bird according to the local feature results, and judging whether the target bird exists in the target group according to the similarity; S500, if yes, send a specific audio to the flock of birds to attract the target birds to the island; when the target birds are successfully attracted, perform directional identification on the target birds during the process of the target birds flying to the island, and obtain the directional identification results of the target birds; S600, periodically acquiring video images of the interior of the island, calculating the target number of target birds in the island according to the directional recognition result, and adjusting the sending mode of the specific audio according to the target number.

[0025] In step S100, migratory birds usually migrate every year, and the target time is set according to the migration time of the migratory birds. In addition, when the external conditions remain unchanged, the migration route of the migratory birds is relatively fixed. Therefore, the target area can be set according to the migration route of the migratory birds.

[0026] In step S200, the species of migrating bird flocks can usually be determined by combining multiple characteristics such as flight formation, chirping sounds, flight patterns, wing shapes, wing flapping frequencies, and group behavior. Data on these characteristics of the bird flock can be obtained through video data of the bird flock. The data obtained can be used to determine whether the species of the bird flock belongs to the target group. In addition, some birds, due to their small numbers, will mix with other birds during the migration process for safe migration. For example, Chinese Crested Terns will take advantage of the clustering protection habits of Great Crested Terns and Lesser Steller Terns, and mix with Great Crested Terns and Lesser Steller Terns to roost, breed, and migrate.

[0027] In the above situation, in order to obtain video data of rare bird species in migration, in addition to obtaining video data of such birds, it is usually necessary to obtain video information of bird populations with which they often mix. For example, in order to obtain video data of the migration of Chinese Crested Terns, it is necessary to obtain video data of the migration of Chinese Crested Terns, Great Crested Terns and Lesser Crested Terns.

[0028] In step S300 and step S400, under normal circumstances, the type of bird can be determined by the bird's appearance features, head features, flight patterns, etc., and the unique landmarks of the target bird are recorded as local features of the target bird. In addition, an image detector that can capture each individual bird in the target group is selected based on the flight path of the target group, and the local features of the bird are extracted based on the video image data captured by the image detector to obtain local feature results. The obtained local feature results are compared with the local features of the target bird to obtain the similarity between the two. For example, an adult Chinese Gull often has a tuft of black or dark brown feathers on its head, which are relatively obvious and often upright. This crest is an important mark that distinguishes it from other terns.

[0029] It should be noted that the feathers of some birds may vary with age, season or breeding period. For example, adult Chinese Crested Terns have a distinct black head, white abdomen and bright black wingtips during the breeding season, while in the non-breeding season, the color of the head of adult Chinese Crested Terns will fade.

[0030] In step S500, in nature, birds can communicate through chirping, so the target birds can be called and attracted through specific audio. In addition, the sending mode of the specific audio is managed according to the flight pattern of the target group where the target birds are located to improve the success rate of attracting the target birds. There are still certain differences between different individuals of the same bird. By acquiring the differences between individuals, the directional recognition results of the target birds can be obtained. For example, different individuals can be directional recognized according to the color depth of the feathers, the distribution of spots or the shape of the patterns between individual birds.

[0031] In step S600, the target birds are encoded according to the directional recognition results. In addition, when the external environment is relatively stable, the breeding season and hunting area of ​​the target birds are relatively fixed. The target birds are periodically detected according to the living habits of the target birds to detect whether there are target birds that have not been encoded. When target birds that have not been encoded appear, they are encoded, and the current number of target birds is obtained according to the encoding. The target number of target birds is calculated according to the current number and the ecological environment of the island, and the sending mode of the specific audio is adjusted according to the difference between the target number and the current number. It should be noted that the target number refers to the optimal value that can accommodate the target birds on the island, and the target number is calculated based on the environment of the island and the survival mode of the target birds.

[0032] The setting of target time period and target area improves the working time and working range of image detectors and makes them more targeted. In the process of image detection, there is no need to perform frame-by-frame image analysis on each individual bird, which improves the working efficiency of image detectors. The determination process of target groups fully considers the actual situation of bird migration, making the acquisition of target groups more reasonable. The similarity between each individual bird and the target bird is calculated based on the local feature results, which fully considers the errors caused by video judgment and improves the accuracy of target group identification. The process of directional identification realizes the marking of each individual bird and can accurately obtain the number of target birds inside the island. By setting the target number, it helps to control the target birds to develop to a number that is more suitable for their own survival.

[0033] [Second embodiment] See also Figure 2In a specific embodiment, when the flock of birds belongs to a target group, local features are acquired for each individual bird in the target group to obtain local feature results, which specifically include: S310, acquiring the body contour of each individual bird in the target group, and selecting body feature points corresponding to the target group in the body contour; S320: Determine the local feature result of each bird individual according to the position of the body feature point.

[0034] In step S310, the body contours of birds are rich and varied. Usually, the category of birds can be judged according to the body contours of birds. However, in the process of shooting a flock of flying birds, it may be encountered that the body contours of the photographed birds are not comprehensive due to incomplete shooting angles and individual birds being blocked. Therefore, it is necessary to analyze the target group video captured by the image detector frame by frame to obtain the body contour of each individual bird. According to the body contour of each individual bird, the target group is further screened, and the individual birds that do not belong to the target group are eliminated. The body feature points such as body shape, wing shape, tail shape, foot features, head features and body color of each individual bird in the screened target group are obtained, and the body feature points of each individual bird are screened and summarized to obtain the body feature points corresponding to the target group. For example, when there are multiple birds in the target group, the individual birds that do not belong to the target group are eliminated, thereby improving the recognition efficiency, avoiding the frame-by-frame recognition of each individual bird in the image, and reducing the workload required for image recognition.

[0035] In step S320, when birds of the same family have similar local feature points during flight, the local feature results of the individual bird can be determined based on the positions of the body feature points. For example, during flight, the wing shapes of the Chinese Crested Tern and the Great Crested Tern are similar, but there are certain differences in their body sizes. The local feature results of the individual bird can be obtained by measuring the size of the wingspan.

[0036] The acquisition of the body contour of each individual bird fully considers the possible impact of the actual shooting environment, further improving the accuracy of the target group. The local feature results are determined according to the position of the body feature points, fully considering the situation that birds of the same family have relatively similar local feature points when flying, thereby improving the accuracy of target bird identification.

[0037] [Third embodiment] In a specific embodiment, the body contour of each bird individual in the target group is obtained, and the body feature points corresponding to the target group are screened out from the body contour, specifically including: S311, judging the flight path of the target group according to the flight mode of the target group, recording the image detector that photographs the target group as the initial detector, and selecting the supplementary detector corresponding to the initial detector according to the flight path; S312, obtaining an initial image of the body contour according to the initial detector, and when the initial image is unclear, controlling the supplementary detector to change the shooting angle, obtaining a supplementary image of the body contour, and determining body feature points corresponding to the target group according to the initial image and the supplementary image.

[0038] In step S311, the initial detector is the first image detector to capture the target group. After the initial detector captures the corresponding target group, other image detectors do not perform image detection on the target group before the initial detector issues a detection result.

[0039] Usually, there is a certain connection between the flight pattern of birds and their flight paths. For example, when some seagulls fly along the coastline, they will rise with the help of the sea breeze and then glide along the coastline. Their flight path is roughly parallel to the coastline and fluctuates with the direction and strength of the sea breeze. When they need to rise again or change direction, they will flap their wings, and the flight path will rise or turn slightly. Therefore, the flight path of the target group can be judged by factors such as the flight pattern of birds, climate, and group behavior.

[0040] In step S312, usually, when birds are flying, the videos and images captured of the birds may be blurry, or the individual birds may be obscured. When these situations occur, the shooting angle of the supplementary detector is adjusted in time according to the flight path of the target group. Video images of individual birds at different angles can be obtained based on the initial image and the supplementary image. The number of body feature points of individual birds can be increased based on the video images at different angles, making the obtained data more accurate.

[0041] It should be noted that when the angle of photographing individual birds is relatively single and the species of the individual birds cannot be determined, and it is necessary to increase the shooting angle, the shooting angle of the supplementary detector can also be adjusted to enrich the body feature points of the photographed individual birds.

[0042] The initial image and the supplementary image are obtained according to the initial detector and the supplementary detector, which fully considers the actual situation of the target group during flight, avoids unclear images caused by single-angle shooting, and reduces the probability of misjudgment due to incomplete shooting.

[0043] [Fourth embodiment] See also Figure 3In a specific embodiment, the similarity between each individual bird and the target bird is calculated based on the local feature results, and whether the target bird exists in the target group is determined based on the similarity, which specifically includes: S410, when the similarity is greater than the first result threshold, marking the individual bird as a target bird; when the similarity is less than or equal to the first result threshold, not marking the individual bird; S420: When the target bird appears, stop calculating the similarity between the individual bird and the target bird.

[0044] In step S410, generally, in order to determine whether there is a target bird in the target group, it is necessary to perform image recognition on each individual bird. Therefore, the local feature results of the individual bird can be compared with the local features of the target bird through an image detector to obtain the similarity between the two. For example, a deep neural network is used to extract a high-dimensional feature vector of the image, and then the similarity between the feature vectors is calculated.

[0045] For example, when the first result threshold is 70% and the similarity of a certain bird individual is 80%, then the bird individual is marked as a target bird.

[0046] In step S420, usually, when performing image recognition on individual birds, it is necessary to perform it one by one. When the target bird appears, it can be determined that the target bird exists in the target group. At this time, there is no need to continue to perform subsequent recognition on each individual bird in the target group, and the subsequent attraction operation can be directly performed on the target group.

[0047] The setting of the first result threshold fully considers the errors brought by the image detector, while reducing the missed detection rate of the target bird and improving the efficiency of image recognition. When the target bird appears, the similarity calculation is stopped, which reduces unnecessary recognition operations of the image detector and reduces invalid recognition workload.

[0048] [Fifth embodiment] See also Figure 4 In a specific embodiment, if yes, a specific audio is sent to the flock of birds to attract the target birds to the island, specifically including: S510, when there is a target bird in the target group, calculating a first distance from the target bird to the target position and a flight speed of the target bird; S520, calculating the sending time and the decibel of the sending volume of the specific audio according to the first distance and the flight speed; S530: Calculate the duration of the influence of the specific audio on the target bird according to the flight speed of the target bird and the flight path of the target bird, and control the sending mode of the specific audio according to the influence duration.

[0049] In step S510, the flight speed of the target birds can usually be calculated by instrument measurement, direct observation comparison or empirical estimation. For example, a high-speed camera is aimed at a flock of flying birds to record the flight images of the flock of birds over a period of time. The changes in the position of the flock of birds in each frame are analyzed by software. The actual flight speed of the flock of birds can be calculated based on the shooting frame rate and the pixel distance of the flock of birds moving in the picture, combined with parameters such as the focal length of the camera. For example, a radar is used to measure the flight speed of a flock of birds. The electromagnetic waves emitted by the radar will be reflected back after encountering the flock of birds. The speed of the flock of birds, the flight direction, and the position from the radar can be accurately calculated based on the frequency changes of the reflected waves.

[0050] In step S520, under normal circumstances, the sending time of the specific audio can be calculated according to the first distance, the flight speed of the target group and the propagation speed of the sound. For example, it is known that the flight speed of the target group is v1, the propagation speed of the specific audio is v2, the first distance is d1, the target position refers to the position where the target group flies to a certain position where the specific audio can be heard, the distance between the island and the target position is d2, the sending time of the specific audio is t, the time required for the sound to propagate to the target position is t2, and the time required for the target group to fly to the target position is t1. There is the following relationship between them: t=t1-t2=d1÷v1-d2÷v2.

[0051] In addition, the intensity of sound propagates in the air at a decreasing rate, so it is necessary to set the decibel size of a specific audio according to the distance. For example, the sound intensity is inversely proportional to the square of the distance. Assuming that the initial intensity of the sound is I0, when the sound intensity decays to I1, it is considered that it no longer affects the target. Assuming that the distance of sound propagation at this time is r1, according to the attenuation formula, we can get: I0÷I1=r1 2 .

[0052] In step S530, under normal circumstances, the duration of the impact of specific audio on the target birds is limited, so it is necessary to calculate the duration of the impact of the specific audio on the target birds, determine the next target position based on the impact duration and the flight path of the target birds, and adjust the direction of the specific audio sending device according to the orientation of the new target position. For example, assuming that the target group can hear the specific audio when flying to the target position, the position of the island is connected to the target position, and the angle formed by the line segment between the two and the ground is θ1. The angle between the specific audio sending device and the ground is adjusted according to θ1, and the direction of the specific audio sending device is adjusted according to the direction of the target position.

[0053] The duration of the impact of a specific audio frequency on the target birds is calculated according to the attenuation formula. For example, the sound intensity at the target position is 100 units, and it is considered negligible when it decays to 1 unit. According to the attenuation formula, r1=√100=10 meters, so the duration of the impact from the emission of the sound to the decay to the negligible intensity is t3=r1÷v2=10÷343≈0.029 seconds.

[0054] It should be noted that the propagation of sound is affected by the characteristics of the sound source and environmental factors. In order to avoid attracting target birds already on the island, loudspeakers or directional sound technology can be used to make the sound propagate mainly in a specific direction.

[0055] The sending time and decibel volume of the specific audio are calculated based on the first distance and flight speed, taking into account the situations encountered by the sound during the actual propagation process. The sending mode of the specific audio is controlled according to the impact duration, and the possible reactions of the target group when hearing the specific audio during flight are fully considered, thereby increasing the possibility of attracting target birds.

[0056] [Sixth embodiment] See also Figure 4 In a specific embodiment, after the target bird is successfully attracted, the target bird is directional identified during the process of flying to the island, and the directional identification result of the target bird is obtained, which specifically includes: S540, when the target bird is successfully attracted, determining the shooting angle of the image detector according to the flight posture of the target bird, and recording the captured image after adjusting the shooting angle as the shooting result; S550: perform directional identification on the target bird according to the shooting result, obtain the directional identification result of the target bird, and encode the target bird according to the directional identification result.

[0057] In step S540, when an individual bird flies towards the island, the individual bird is identified again through the image detector. Generally speaking, the appearance characteristics of different individuals of the same species vary to a certain extent due to genetic variation or the influence of the acquired environment. When the individual bird belongs to the target bird, in order to capture more details of the appearance characteristics of the target bird, it is necessary to adjust the shooting angle of the image detector according to the flight posture of the target bird. When the target bird is attracted, the target bird will leave the target group. At this time, it is easier to capture the characteristics of each bird when shooting the target bird.

[0058] In step S550, according to the results captured in step S540, the target birds are encoded according to the identification results, and the characteristics of each target bird are stored according to its own physical characteristics. When the image detector detects the relevant birds later, it can be quickly locked. For some target birds, they can be identified by manually adding marks on the island.

[0059] The shooting angle of the image detector is determined according to the flight posture of the target bird, which enriches the details of the target bird's appearance features and makes directional identification easier. After the target bird is successfully attracted, the image detector can capture more details of the target bird during its flight to the island. The target bird is encoded according to the directional identification results, making the subsequent management of the target bird more convenient.

[0060] [Seventh embodiment] In a specific embodiment, periodically acquiring video images of the interior of the island and calculating the number of target birds in the interior of the island according to the directional recognition result specifically includes: S610, periodically performing detection in the target area within the island according to the living habits of the target birds; S620, detect the target birds on the island, perform directional identification on the target birds that have not been coded, and obtain the coding of the directional identification results, obtain the current number of target birds on the island through coding, and obtain the target number of target birds based on the ecological conditions of the island and the current number of target birds.

[0061] In step S610, in an ecosystem, each bird has its own ecological niche. The bird's living habits and activity areas can be understood based on the bird's ecological niche. In order to detect the target bird in a targeted manner, the target bird can be detected based on its living habits and activity areas.

[0062] In step S620, generally, there may be omissions in the process of directional identification of target birds when they fly to the island, so it is necessary to periodically monitor the target birds on the island. In addition, periodic detection can ensure that when new target birds are born, they can be discovered in time and protected.

[0063] The target birds on the island are detected periodically within the target area on the island, which improves the accuracy of the target bird population, facilitates the dynamic detection of the target bird population, and facilitates the subsequent management of the target bird population.

[0064] [Eighth embodiment] In a specific embodiment, adjusting the sending mode of a specific audio according to the target quantity specifically includes: S630: when the difference between the current number and the target number is less than or equal to the first threshold, the sending mode of the specific audio is not changed; S640: When the difference between the current number and the target number is greater than a first threshold, control the sending mode of the specific audio according to the flight mode of the target group.

[0065] In steps S630 and S640, under normal circumstances, a target population size to be achieved can be formulated based on multiple factors such as the habitat conditions, population status, and reproductive capacity of the target birds. The subsequent number of target birds can be predicted by the natural growth rate of the target birds. Therefore, the first threshold can be set according to the natural growth rate of the target birds. In step S640, when the existing number is close to the target number, the sending frequency, sending direction, and sending volume of the specific audio are set according to the flight pattern of the target group. For example, when there are other islands near the island and the other islands are also suitable for the survival of the target birds, the target birds can be attracted by the specific audio sending devices on the other islands. It should be noted that when there are no specific audio sending devices on the nearby islands, the sending direction of the specific audio sending devices on the islands can be appropriately adjusted according to the geographical locations between the islands to achieve the purpose of attracting the target birds.

[0066] It should be noted that when the number of target birds on an island reaches or exceeds the target number, and there are no other islands nearby that are suitable for the survival of the target birds or the number of target birds on other islands has also reached the target number, the target number can be expanded by improving the island environment, or by reporting to the relevant departments.

[0067] When the difference between the current number and the target number is greater than the first threshold, the sending mode of the specific audio is controlled according to the flight pattern of the target group, which fully considers the impact of the target bird population and the subsequent possible increase in the number on the survival of the target birds, which is conducive to the ideal number of target birds on the island.

[0068] [Ninth embodiment] See also Figure 5 The present invention also provides a management system 100. The management method described in the above embodiment is applied to the management system 100. The management system 100 includes: an image capture module 110: the image capture module 110 is used to obtain video data of a flock of birds; an image recognition module 120: the image recognition module 120 is used to perform directional recognition of target birds; a calculation module 130: the calculation module 130 is used to calculate the similarity between each individual bird and the target bird; an audio sending module 140: the audio sending module 140 is used to adjust the sending mode of a specific audio. The management system 100 has all the technical features of the above management method, which will not be described one by one here.

[0069] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A bird management method based on power image detection, characterized in that: The management method comprises: Arrange a plurality of image detectors on the island, and perform image detection on a target area around the island within a target time period; When a flock of birds passes through the target area, obtaining video data of the flock of birds, and determining whether the flock of birds belongs to the target group according to the video data; When the flock of birds belongs to the target group, local features are acquired for each individual bird in the target group to obtain local feature results; Calculating the similarity between each of the individual birds and the target bird based on the local feature results, and judging whether the target bird exists in the target group based on the similarity; If yes, a specific audio frequency is sent to the flock of birds to attract the target birds to the island; When the target bird is successfully attracted, during the process of the target bird flying to the island, the target bird is oriented and identified, and an oriented identification result of the target bird is obtained; Periodically acquiring video images of the interior of the island, and calculating the target number of the target birds within the island according to the directional recognition result; The sending mode of the specific audio is adjusted according to the target number.

2. The management method according to claim 1, characterized in that: When the flock of birds belongs to the target group, local features are acquired for each individual bird in the target group to obtain local feature results, specifically including: Acquiring the body contour of each individual bird in the target group, and selecting body feature points corresponding to the target group in the body contour; The local feature result of each of the individual birds is determined according to the position of the body feature points.

3. The management method according to claim 2, characterized in that: The step of acquiring the body contour of each individual bird in the target group and selecting body feature points corresponding to the target group in the body contour specifically includes: According to the flight mode of the target group, the flight path of the target group is determined, the image detector that photographs the target group is recorded as an initial detector, and a supplementary detector corresponding to the initial detector is selected according to the flight path; Acquire an initial image of the body contour according to the initial detector, and when the initial image is unclear, control the supplementary detector to change the shooting angle to acquire a supplementary image of the body contour; The body feature points corresponding to the target group are determined according to the initial image and the supplementary image.

4. The management method according to claim 3, characterized in that: The calculating the similarity between each individual bird and the target bird according to the local feature result, and judging whether the target bird exists in the target group according to the similarity, specifically includes: When the similarity is greater than a first result threshold, marking the bird individual as the target bird; When the similarity is less than or equal to the first result threshold, the bird individual is not marked; When the target bird appears, the calculation of the similarity between the individual bird and the target bird is stopped.

5. The management method according to claim 4, characterized in that: If so, sending a specific audio to the flock of birds to attract the target birds to the island, specifically including: When the target bird exists in the target group, calculating a first distance from the target bird to a target position and a flight speed of the target bird; Calculate the sending time and the decibel of the sending volume of the specific audio according to the first distance and the flight speed; The duration of influence of the specific audio on the target bird is calculated according to the flight speed of the target bird and the flight path of the target bird, and the sending mode of the specific audio is controlled according to the influence duration.

6. The management method according to claim 5, characterized in that: After the target bird is successfully attracted, during the process of the target bird flying to the island, the target bird is directional identified to obtain the directional identification result of the target bird, which specifically includes: When the target bird is successfully attracted, the shooting angle of the image detector is determined according to the flight posture of the target bird, and the image captured after adjusting the shooting angle is recorded as a shooting result; The target bird is oriented and identified according to the shooting result, the oriented identification result of the target bird is obtained, and the target bird is encoded according to the oriented identification result.

7. The management method according to claim 6, characterized in that: The periodic acquisition of the video image inside the island and the calculation of the target number of the target birds inside the island according to the directional recognition result specifically include: According to the living habits of the target birds, periodically conducting detection in the target area within the island; Detecting the target birds on the island, performing the directional identification on the target birds that have not been coded, obtaining the codes of the directional identification results, and obtaining the current number of the target birds on the island through the codes; The target number of the target birds is obtained according to the ecological conditions of the island and the existing number of the target birds.

8. The management method according to claim 7, characterized in that: The adjusting the sending mode of the specific audio according to the target quantity specifically includes: When the difference between the existing number and the target number is less than or equal to a first threshold, the sending mode of the specific audio is not changed; When the difference between the existing number and the target number is greater than a first threshold, the sending mode of the specific audio is controlled according to the flight mode of the target group.

9. A management system for birds on an island, characterized in that: The management method according to any one of claims 1 to 8 is applied to the management system, and the management system comprises: Image capture module: the image capture module is provided with the image detector for acquiring the image data of the flock of birds; Image recognition module: The image recognition module is used to perform directional recognition of the target bird; Calculation module: The calculation module is used to calculate the similarity between each of the bird individuals and the target bird; Audio sending module: The audio sending module is used to adjust the sending mode of the specific audio.

Images