Diversified biological collection and recognition system

Abstract

The invention innovatively provides a diversified organism collection and recognition system, an unmanned aerial vehicle platform is provided with a high-definition camera and a microphone, a database comprises a plurality of groups of organism types and life habits, the organism types are in one-to-one correspondence with the life habits, and the life habits comprise sound characteristics, life traces, inhabiting environments and the like; the habit analysis module is used for acquiring animal traces in a to-be-analyzed image in an image analysis technology, and calculating to obtain a habit influence value through life trace reliability of all marked animal traces; the sound recognition module is used for acquiring animal sounds or activity sounds in the sound data as audio clips to be analyzed, and calculating to obtain a sound influence value through the sound similarity of the marked audio clips to be analyzed; and the comprehensive evaluation module is used for acquiring the quantity, the habit influence value and the sound influence value of the to-be-analyzed organisms, and acquiring the predicted quantity of the to-be-analyzed organisms in a specified area through a statistical algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of biometric collection, and more particularly to a diversity biometric collection and recognition system. Background Art

[0002] The purpose of biodiversity collection is to fully understand the basic information such as the distribution, quantity, and species of biodiversity. It is of great significance to maintain the balance of the ecosystem, ensure food safety and human survival. For areas that are inconvenient for people to enter, such as forests and wild areas, collection is mainly carried out through drones. In the process of biological quantity counting, the number of organisms photographed by drones is currently the main basis. For organisms that are not photographed, it is impossible to count the number. In the process of biological collection, the statistical deviation of the number of organisms is large. Summary of the invention

[0003] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a diverse biometric collection and identification system to overcome the above-mentioned defects in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A diverse biological collection and identification system includes an unmanned aerial vehicle platform equipped with a high-definition camera and a microphone, and also includes an identification system, wherein the identification system includes a data collection module, an image analysis module, a habit analysis module, a sound recognition module, and a comprehensive evaluation module;

[0006] The database includes several groups of biological species and living habits, wherein the biological species correspond to the living habits one by one, and the living habits include sound characteristics, living traces, habitats, etc.;

[0007] The data acquisition module is used to control the UAV to collect image data and sound data in a specified area according to a preset flight strategy;

[0008] The image analysis module is used to obtain image data in real time as the image to be analyzed, determine the organisms to be analyzed, and obtain the number of organisms to be analyzed in the image to be analyzed by image recognition technology;

[0009] The habit analysis module is used to obtain animal traces in the image to be analyzed in the image analysis technology, and compare all animal traces with the life traces of the organism to be analyzed, respectively, to obtain the reliability of the life traces, and preset a reliability threshold. If the reliability of the life trace is greater than the reliability threshold, the animal trace is marked, and the habit influence value is calculated through the reliability of the life traces of all marked animal traces;

[0010] The sound recognition module is used to obtain animal calls or activity sounds in the sound data as audio segments to be analyzed, compare all audio segments to be analyzed with the sound features of the organism to be analyzed, and obtain sound similarity, and a similarity threshold is preset. If the sound similarity is greater than the similarity threshold, the audio to be analyzed is marked, and the sound influence value is calculated based on the sound similarity of the marked audio to be analyzed;

[0011] The comprehensive evaluation module is used to obtain the number of organisms to be analyzed, the influence value of habits and the influence value of sound, and obtain the estimated number of organisms to be analyzed in the specified area through statistical algorithms;

[0012] The data storage module is used to obtain and store the types and quantities of the organisms to be analyzed, and to construct an organism distribution model according to the distribution of the organisms to be analyzed.

[0013] Preferably, the living traces include activity paths, nest locations, foraging areas, etc.

[0014] Preferably, a scratch analysis unit is provided in the habit analysis module, and the scratch analysis unit is used to obtain scratch image information in the image to be analyzed, and compare the scratch image information with the life traces in the database.

[0015] Preferably, a nest recognition unit is provided in the habit analysis module, and a target detection model is preset in the nest recognition unit. The target detection module is used to analyze and identify the image to be analyzed, obtain the target point, and obtain the nest credibility, and a credibility threshold is preset. If the nest credibility is greater than the credibility threshold, the habit impact value is calculated based on the living habit reliability and the nest credibility.

[0016] Preferably, the habit analysis module is further provided with an identification calibration unit. When the nest credibility is less than the credibility threshold, the identification calibration unit is used to obtain the nest credibility and compare it with the minimum threshold in the identification calibration unit. If the nest credibility is greater than the minimum threshold, the target point is obtained and an execution command is generated. The drone is provided with a blowing component and also includes an execution module. The execution module is used to obtain the execution command and the target point in the identification calibration unit, control the drone to move to the target point, and drive the blowing component to move, so as to move the obstruction on the surface of the nest and control the nest identification unit to re-identify.

[0017] Preferably, the sound recognition module is further provided with a sound simulation unit, in which the audio information of the organism to be analyzed is preset. When the current volume in the sound data is too low, the sound simulation unit can be used to randomly play the audio information of the organism to be analyzed.

[0018] Preferably, the sound recognition module is also preset with an environment judgment unit, which is used to collect environmental characteristics, including sampling time, ambient temperature and ambient humidity, and obtain the habitat of the organism to be analyzed, obtain standard characteristics based on the habitat, and the standard characteristics reflect the habitat suitable for the life of the organism to be analyzed, compare the environmental characteristics with the standard characteristics, and generate a judgment value, and obtain the sound impact value through the judgment value and the sound similarity.

[0019] Preferably, a flight control module is also included, and the flight control module is used to obtain the living habits of the organisms to be analyzed and dynamically adjust the flight path of the drone according to the image data collected in real time.

[0020] Preferably, the determining of the organism in the image to be analyzed includes obtaining a moving object in the image to be analyzed as a feature object to be analyzed, obtaining a type of the feature object to be analyzed based on an image recognition technology, and comparing the type of the feature object to be analyzed with the type of the organism to be analyzed in the data storage module; if the type of the feature object to be analyzed is the same as the type of the organism to be analyzed in the data storage module, re-determining the organism to be analyzed; if the type of the feature object to be analyzed is the same as the type of the organism to be analyzed in the data storage module, marking the feature object as the organism to be analyzed.

[0021] The beneficial effects of the present invention are as follows: using a drone as a carrier and formulating a targeted flight strategy according to the living habits of the organisms to be analyzed, so as to fully detect the common activity range of the organisms to be analyzed; using a habit analysis module, comparing the animal traces in the image to be analyzed with the living traces of the organisms in the database, and evaluating the reliability of the living traces, so as to calculate the habit influence value, and through the habit analysis, a more comprehensive understanding of the living habits and activity range of the organisms to be analyzed can be obtained; and through a sound recognition module, analyzing the animal calls or activity sounds in the sound data, and comparing them with the sound characteristics of the organisms in the database, so as to obtain the sound influence value; by obtaining the number of organisms to be analyzed captured by the image, the sound influence value and the habit influence value, a comprehensive analysis is performed to obtain the estimated number of organisms to be analyzed, and the accuracy of the number in the biological collection process is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is the overall structural diagram of the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a component centered. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a component centered. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a component centered. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0026] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0027] like Figure 1 As shown, the present invention provides a diverse biological collection and identification system, characterized in that it includes an unmanned aerial vehicle platform, which is equipped with a high-definition camera and a microphone, and uses the high-definition camera to take real-time photos of the forest, and uses the microphone to collect forest sounds. At the same time, the microphone can be used to play sounds to attract biological organisms; it also includes an identification system, which includes a data collection module, an image analysis module, a habit analysis module, a sound recognition module, and a comprehensive evaluation module;

[0028] The database includes several groups of biological species and living habits. Biological species correspond to living habits one by one. Living habits include sound characteristics, living traces, habitats, etc. Living traces include activity paths, nest locations, foraging areas, etc. Habitats include the temperature, humidity and activity time that are suitable and preferred by biological life;

[0029] The data acquisition module is used to control the drone to collect image data and sound data in the specified area according to the preset flight strategy. The data acquisition module is used to control the drone to set the specified flight strategy and collect data within the specified area. According to the different living habits of animals and the environment information, different flight strategies are adjusted. The flight strategy includes the adjustment of parameters such as flight path, altitude, speed, etc., and controls the drone to fly stably in the specified area, including the execution of flight actions such as hovering, straight flight, turning, and landing. Through the preset flight strategy, every corner of the specified area can be covered more comprehensively, so as to collect more comprehensive and detailed information.

[0030] The image analysis module is used to obtain image data as the image to be analyzed in real time, determine the organisms to be analyzed, and obtain the number of organisms to be analyzed in the image to be analyzed through image recognition technology; the image analysis module implements the reception of image data collected by the drone to ensure the timeliness of data collection and analysis, and pre-processes the acquired image data to obtain the image to be analyzed, and extracts features and classifies the image to be analyzed based on computer vision and machine learning algorithms, determines the species of organisms to be analyzed in the image, and further counts the number of organisms to be analyzed based on the identification of the species of organisms to be analyzed;

[0031] It also includes a flight control module, which is used to obtain the living habits of the organisms to be analyzed and dynamically adjust the flight path of the drone based on the image data collected in real time. The flight path of the drone is dynamically adjusted by obtaining the living habits of the organisms to be analyzed and based on the image data collected in real time.

[0032] The habit analysis module is used to obtain animal traces in the image to be analyzed in the image analysis technology. Animal traces include footprints, feces, nests, etc., and compare all animal traces with the life traces of the organism to be analyzed. The database contains the life trace samples and feature descriptions of the organism to be analyzed, and performs accurate matching. According to the comparison results, the reliability of life traces of different animal traces is calculated. The reliability of life traces involves the sum of multiple factors such as matching degree, trace freshness and environmental factors. Through the reliability of life traces, the existence and distribution of the organism to be analyzed in the specified area can be further determined, which is convenient for the ecological analysis of the organism to be analyzed. Detection, and the reliability of life traces reflects the activity level of organisms in the area, which helps to understand the ecological habits and behavioral patterns of organisms, facilitates the demarcation of protected areas for them, and specifies protection measures; and a reliability threshold is preset. If the reliability of life traces is greater than the reliability threshold, the animal traces are marked, and the habit impact value is calculated through the reliability of life traces of all marked animal traces; by comparing the reliability of life traces with the reliability threshold, some animal traces that are not the noise of the organism are screened, and the life traces that may be produced by the organism to be analyzed are retained, and the habit impact value of the specified area is calculated through the reliability of all marked life traces;

[0033] The habit analysis module is provided with a scratch analysis unit, which is used to obtain scratch image information in the image to be analyzed and compare the scratch image information with the life traces in the database.

[0034] The habit analysis module is provided with a nest recognition unit, and the nest recognition unit is preset with a target detection model. The target detection module is used to analyze and identify the image to be analyzed, obtain the target point, and obtain the nest credibility, and a credibility threshold is preset. If the nest credibility is greater than the credibility threshold, the habit impact value is calculated based on the reliability of living habits and the credibility of the nest; the target detection model is such as a convolutional neural network model that focuses on deep learning. After being trained with a large amount of labeled data, the model can quickly identify the nest characteristics of the organism to be analyzed in the image. After collecting image data using a drone, the animal nest recognition unit uses the target detection model to identify the image to be analyzed, extracts feature information from the image to be analyzed, and compares it with the pre-trained feature library to determine whether there is a nest in the image, and obtains its target point, that is, the specific location of the nest. In the recognition process, the feasibility of the nest is calculated based on the clarity, completion type and feature matching of the nest, and a credibility threshold is preset to judge the reliability of the recognition result. If the credibility is greater than the threshold, the recognition result is considered reliable; if it is less than the threshold, further processing or calibration is required.

[0035] The habit analysis module is also provided with an identification calibration unit. When the nest credibility is less than the credibility threshold, the identification calibration unit intervenes to improve the accuracy of the identification. The identification calibration unit is used to obtain the nest credibility and compare it with the minimum threshold in the identification calibration unit. If the nest credibility is greater than the minimum threshold, the target point is obtained and an execution command is generated. The identification calibration unit is provided with a minimum threshold to determine whether further processing is required. If the animal nest credibility is greater than the minimum threshold but less than the credibility threshold, it means that although the identification result is not reliable enough, it is still possible to improve the credibility through calibration, that is, there may be obstructions on the surface of the nest, which affects the judgment. The drone is provided with a blowing component and also includes an execution module. The execution module is used to obtain the execution command and the target point in the identification calibration unit, and control the drone to move to the target point, and drive the blowing component to move, so as to move the obstruction on the surface of the nest, and control the nest identification unit to re-identify; the execution module controls the drone to move to the target point (i.e., the nest position) and drives the blowing component on the drone to move. The blowing component attempts to remove obstructions (such as fallen leaves, weeds, etc.) on the surface of the nest by blowing air, so as to more clearly expose the characteristics of the nest; after the blowing operation is completed, the animal nest recognition unit re-identifies the target point and recalculates the credibility of the animal nest.

[0036] The sound recognition module is used to obtain the animal calls or activity sounds in the sound data as the audio segments to be analyzed, compare all the audio segments to be analyzed with the sound features of the organism to be analyzed, and obtain the sound similarity. A similarity threshold is preset. If the sound similarity is greater than the similarity threshold, the audio to be analyzed is marked, and the sound impact value is calculated by the sound similarity of the marked audio to be analyzed; each audio segment to be analyzed will be compared with multiple samples in the sound feature library of the organism to be analyzed, and the degree of match between the two will be evaluated by calculating the similarity of audio features (such as frequency, pitch, length, timbre, etc.); the preset similarity threshold is used to determine the audio segment and the organism to be analyzed. Whether the similarity between the sound features of the organism to be analyzed is high enough to identify the audio clip as the call or activity sound of the organism to be analyzed; each audio clip to be analyzed will be compared with multiple samples in the sound feature library of the organism to be analyzed, and the degree of match between the two will be evaluated by calculating the similarity of audio features (such as frequency, pitch, length, timbre, etc.); the preset similarity threshold is used to determine whether the similarity between the audio clip and the sound features of the organism to be analyzed is high enough to identify the audio clip as the call or activity sound of the organism to be analyzed; by obtaining the sound similarity of all marked clips to be analyzed and obtaining the sound influence value, the possible number of organisms to be analyzed in the specified area is obtained using the sound influence value.

[0037] The sound recognition module is also equipped with a sound simulation unit, which is preset with the audio information of the creature to be analyzed. When the current volume in the sound data is too low, the sound simulation unit can be used to randomly play the audio information of the creature to be analyzed; when the volume of the environmental figure data is detected to be too low, the sound effects can be played randomly to simulate the real call of the creature to be analyzed and the activity sound to attract the call of the creature to be analyzed, so as to facilitate the collection of the call of the creature to be analyzed.

[0038] The sound recognition module is also preset with an environmental judgment unit, which is used to collect environmental features, including sampling time, ambient temperature and ambient humidity, and obtain the habitat of the organism to be analyzed, obtain standard features based on the habitat, and the standard features reflect the habitat suitable for the organism to be analyzed. Compare the environmental features with the standard features, and generate a judgment value. The sound impact value is obtained through the judgment value and the sound similarity; the environmental judgment unit is responsible for collecting and analyzing environmental features, including sampling time, ambient temperature, ambient humidity, etc., to evaluate whether the current environment is suitable for the life of the organism to be analyzed; the environmental judgment unit compares the current environmental features with the standard features and generates a judgment value. The high or low judgment value reflects the degree of fit between the current environment and the standard environment; the judgment value combined with the sound similarity can further refine the calculation of the sound impact value. For example, even if the sound similarity is high, if the current environment is not suitable for the survival of the organism, the sound impact value may be reduced accordingly.

[0039] The comprehensive evaluation module is used to obtain the number of organisms to be analyzed, the influence value of habits and the influence value of sound, and obtain the estimated number of organisms to be analyzed in the specified area through statistical algorithms; the comprehensive evaluation module uses a series of statistical algorithms (such as regression analysis, machine learning models, Bayesian networks, etc.) to integrate the above data sources and predict the estimated number of organisms to be analyzed in the specified area; the comprehensive evaluation module uses a series of statistical algorithms (such as regression analysis, machine learning models, Bayesian networks, etc.) to integrate the above data sources and predict the estimated number of organisms to be analyzed in the specified area.

[0040] The data storage module is used to obtain and store the species and quantity of the organisms to be analyzed, and to construct a biological distribution model based on the distribution of the organisms to be analyzed; the data storage module is used to collect the species, quantity and distribution of different organisms to be analyzed, and based on the collected data on the distribution of the organisms to be analyzed, the data storage module can construct a biological distribution model. These models usually use geographic information system (GIS) technology to combine biological distribution data with geographic spatial information to graphically display the distribution of organisms in different regions; the data storage module is used to realize rapid retrieval, query and analysis of data, providing necessary data support for the comprehensive evaluation module and other related systems.

[0041] Determining the organism in the image to be analyzed includes obtaining a moving object in the image to be analyzed as a feature object to be analyzed, obtaining a type of the feature object to be analyzed according to an image recognition technology, and comparing the type of the feature object to be analyzed with the type of the organism to be analyzed in the data storage module. If the type of the feature object to be analyzed is the same as the type of the organism to be analyzed in the data storage module, the organism to be analyzed is re-determined; if the type of the feature object to be analyzed is different from the type of the organism to be analyzed in the data storage module, the feature object is marked as the organism to be analyzed; obtaining a moving object in the image to be analyzed: detecting and extracting all moving objects from the image as candidate feature objects to be analyzed, using image recognition technology (such as convolutional neural network, etc.) to identify the type of each candidate feature object, and comparing each identified feature object type with the species of the organism to be analyzed pre-stored in the data storage module one by one; if there is a match (that is, the type of the feature object is the same as a species of a certain organism to be analyzed), ignoring the feature object, re-determining the feature object, or marking it as an analyzed feature object; if there is no match (that is, the type of the feature object is different from all pre-stored species of the organism to be analyzed), marking the feature object as the organism to be analyzed.

[0042] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A diversity biometric collection and recognition system, characterized in that: It includes a drone platform, which is equipped with a high-definition camera and a microphone, and also includes a recognition system, which includes a data acquisition module, an image analysis module, a habit analysis module, a sound recognition module and a comprehensive evaluation module; The database includes several groups of biological species and living habits, wherein the biological species correspond to the living habits one by one, and the living habits include sound characteristics, living traces, habitats, etc.; The data acquisition module is used to control the UAV to collect image data and sound data in a specified area according to a preset flight strategy; The image analysis module is used to obtain image data in real time as the image to be analyzed, determine the organisms to be analyzed, and obtain the number of organisms to be analyzed in the image to be analyzed by image recognition technology; The habit analysis module is used to obtain animal traces in the image to be analyzed in the image analysis technology, and compare all animal traces with the life traces of the organism to be analyzed, respectively, to obtain the reliability of the life traces, and preset a reliability threshold. If the reliability of the life trace is greater than the reliability threshold, the animal trace is marked, and the habit influence value is calculated through the reliability of the life traces of all marked animal traces; The sound recognition module is used to obtain animal calls or activity sounds in the sound data as audio segments to be analyzed, compare all audio segments to be analyzed with the sound features of the organism to be analyzed, and obtain sound similarity, and a similarity threshold is preset. If the sound similarity is greater than the similarity threshold, the audio to be analyzed is marked, and the sound influence value is calculated based on the sound similarity of the marked audio to be analyzed; The comprehensive evaluation module is used to obtain the number of organisms to be analyzed, the influence value of habits and the influence value of sound, and obtain the estimated number of organisms to be analyzed in the specified area through statistical algorithms; The data storage module is used to obtain and store the types and quantities of the organisms to be analyzed, and to construct an organism distribution model according to the distribution of the organisms to be analyzed.

2. A diverse biological collection and recognition system according to claim 1, characterized in that: The living traces include activity paths, nest locations, foraging areas, etc.

3. The diverse biological collection and recognition system according to claim 1, characterized in that: The habit analysis module is provided with a scratch analysis unit, which is used to obtain scratch image information in the image to be analyzed and compare the scratch image information with the life traces in the database.

4. The diverse biological collection and recognition system according to claim 1, characterized in that: The habit analysis module is provided with a nest recognition unit, and the nest recognition unit is preset with a target detection model. The target detection module is used to analyze and identify the image to be analyzed, obtain the target point, and obtain the nest credibility, and a credibility threshold is preset. If the nest credibility is greater than the credibility threshold, the habit impact value is calculated based on the living habit reliability and the nest credibility.

5. A diverse biological collection and recognition system according to claim 4, characterized in that: The habit analysis module is also provided with an identification calibration unit. When the nest credibility is less than the credibility threshold, the identification calibration unit is used to obtain the nest credibility and compare it with the minimum threshold in the identification calibration unit. If the nest credibility is greater than the minimum threshold, the target point is obtained and an execution command is generated. The drone is provided with a blowing component and also includes an execution module. The execution module is used to obtain the execution command and the target point in the identification calibration unit, control the drone to move to the target point, and drive the blowing component to move, so as to move the obstruction on the surface of the nest, and control the nest identification unit to re-identify.

6. The diverse biological collection and recognition system according to claim 1, characterized in that: The sound recognition module is also provided with a sound simulation unit, in which the audio information of the organism to be analyzed is preset. When the current volume in the sound data is too low, the sound simulation unit can be used to randomly play the audio information of the organism to be analyzed.

7. The diverse biological collection and recognition system according to claim 1, characterized in that: The sound recognition module is also preset with an environment judgment unit, which is used to collect environmental characteristics, including sampling time, ambient temperature and ambient humidity, and obtain the habitat of the organism to be analyzed, obtain standard characteristics based on the habitat, and the standard characteristics reflect the habitat suitable for the organism to be analyzed. The environmental characteristics are compared with the standard characteristics, and a judgment value is generated. The sound impact value is obtained through the judgment value and the sound similarity.

8. The diverse biological collection and recognition system according to claim 1, characterized in that: It also includes a flight control module, which is used to obtain the living habits of the organisms to be analyzed and dynamically adjust the flight path of the drone according to the image data collected in real time.

9. The diverse biological collection and recognition system according to claim 1, characterized in that: The determining of the organism in the image to be analyzed includes obtaining a moving object in the image to be analyzed as a feature object to be analyzed, obtaining a type of the feature object to be analyzed according to an image recognition technology, and comparing the type of the feature object to be analyzed with the type of the organism to be analyzed in the data storage module; if the type of the feature object to be analyzed is the same as the type of the organism to be analyzed in the data storage module, re-determining the organism to be analyzed; if the type of the feature object to be analyzed is the same as the type of the organism to be analyzed in the data storage module, marking the feature object as the organism to be analyzed.

Images