Biological diversity detection system based on image recognition

Abstract

The invention discloses a biodiversity detection system based on image recognition, and relates to the technical field of biological recognition, and the system comprises an image collection module which is used for collecting biological image data in a monitoring area and preprocessing the biological image data to generate recognizable data, and the recognizable data comprises recognizable images and recognizable contents; the first recognition module is used for inputting the recognizable image into a preset first analysis model to obtain a first biological variety; the second recognition module is used for inputting the recognizable content into a preset second analysis model to obtain a second biological variety; the variety comparison module is used for comparing the first biological variety with the second biological variety to obtain a biological consistency label, and outputting a biological variety result according to the biological consistency label. Through multi-view and multi-resolution biological image acquisition and preprocessing, the comprehensiveness and accuracy of image data are improved, and the accuracy of the biological variety is improved. And biological capture at different time points is ensured.

Description

Technical Field

[0001] The present invention relates to the field of biometric identification technology, and in particular to a biodiversity detection system based on image recognition. Background Art

[0002] With the increasing importance of environmental protection and biodiversity monitoring, accurate monitoring and classification of biological populations has become an important task in various fields of research. Traditional biodiversity monitoring methods mainly rely on manual on-site observation and recording, which is not only time-consuming and labor-intensive, but also easily interfered by human factors, resulting in incomplete and inaccurate monitoring results. Especially in large-scale monitoring areas, biological populations are diverse and widely distributed, and manual detection is difficult to meet high-frequency and comprehensive monitoring needs. In addition, changes in light and climate under different environmental conditions may also affect the capture and identification of organisms, further increasing the complexity and uncertainty of monitoring work.

[0003] In recent years, with the development of image recognition and machine learning technologies, automated monitoring systems have gradually become an effective means to address the challenges of biodiversity detection. However, existing systems still face some difficulties in practical applications. For example, image acquisition at a single angle and fixed resolution can easily lead to omission or distortion of biological features, which in turn affects the accuracy of recognition. In addition, there are many misjudgments in the automatic recognition system and the lack of manual correction, which limits the reliability of the recognition results. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a biodiversity detection system based on image recognition.

[0005] A biodiversity detection system based on image recognition, the system comprising: An image acquisition module, used to acquire biological image data in the monitoring area and pre-process the biological image data to generate recognizable data, wherein the recognizable data includes a recognizable image and recognizable content; A first recognition module, used for inputting the recognizable image into a preset first analysis model to obtain a first biological species; A second identification module, used for inputting the identifiable content into a preset second analysis model to obtain a second biological species; The variety comparison module is used to compare the first biological variety with the second biological variety to obtain a biological consistency label, and output a biological variety result according to the biological consistency label.

[0006] Furthermore, the biological image data includes biological images of multiple viewing angles, high-resolution biological images and low-resolution biological images, and the collection cycle of the biological image data is once every n minutes.

[0007] Furthermore, the logic of preprocessing biological image data is: After each shooting cycle, synchronize the biological images of different perspectives according to the timestamp to ensure that the biological image data of multiple perspectives are aligned with the same time point, detect the difference in clarity and exposure in the biological images of multiple synchronized perspectives, select the biological image of the main perspective with the highest clarity, mark the images of other perspectives, and output the biological image of the main perspective as a recognizable image; The high-resolution biological images and low-resolution biological images are adaptively processed by using preset resolution differentiation rules to generate recognizable content.

[0008] Furthermore, the logic of the resolution differentiation rule is: The high-resolution biological image is sampled to a suitable size to reduce the amount of redundant data, and is combined with the low-resolution biological image to generate a multi-resolution biological image pyramid. Each resolution layer of the multi-resolution biological image pyramid is used to process different required content, and the processed different required content is output as recognizable content.

[0009] Furthermore, the construction logic of the first analysis model is: Acquire historical image recognition training data, and divide the historical image recognition training data into an image recognition training set and an image recognition test set, wherein the historical image recognition training data includes identifiable images and their corresponding estimated first biological species; Constructing a first regression network, using the recognizable images in the image recognition training set as input data of the first regression network, and using the estimated first biological species in the image recognition training set as output data of the first regression network, training the first regression network, and obtaining an initial first machine learning network; The initial first machine learning network is model verified using an image recognition test set, and an initial first machine learning network that is less than or equal to a preset first test error threshold is output as a pre-constructed first analysis model.

[0010] Furthermore, the construction logic of the second analysis model is: Acquire historical content recognition training data, and divide the historical content recognition training data into a content recognition training set and a content recognition test set, wherein the historical content recognition training data includes identifiable content and its corresponding estimated second biological species; Constructing a second regression network, using the identifiable content in the content recognition training set as input data of the second regression network, and using the estimated second biological species in the content recognition training set as output data of the second regression network, training the second regression network, and obtaining an initial second machine learning network; The initial second machine learning network is model verified using the content recognition test set, and an initial second machine learning network that is less than or equal to a preset second test error threshold is output as a pre-constructed second analysis model.

[0011] Furthermore, the bio-consistency label includes a variety-consistent label and a variety-different label.

[0012] Furthermore, the logic of comparing the first biological species and the second biological species to obtain a biological consistency label is: If the first biological variety and the second biological variety are the same, a variety-consistent label is generated, and the biological variety corresponding to the first biological variety and the second biological variety is output as the biological variety result; if the first biological variety and the second biological variety are different, a variety-different label is generated, and a recognizable image is output for artificial biological variety determination.

[0013] A biodiversity detection method based on image recognition, applied to the biodiversity detection system based on image recognition, the method comprising: Collecting biological image data in the monitoring area and preprocessing the biological image data to generate identifiable data, wherein the identifiable data includes identifiable images and identifiable content; Inputting the recognizable image into a preset first analysis model to obtain a first biological species; Inputting the identifiable content into a preset second analysis model to obtain a second biological species; The first biological variety is compared with the second biological variety to obtain a biological consistency label, and a biological variety result is output according to the biological consistency label.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention improves the comprehensiveness and accuracy of image data through multi-view and multi-resolution biological image acquisition and preprocessing, ensuring the clear presentation and feature capture of organisms at different time points and angles. Through timestamp synchronization and exposure comparison, the main view high-quality image is selected for analysis, which not only reduces redundant information, but also optimizes the quality of the detection model input, thereby improving the stability and accuracy of the detection; In addition, the present invention constructs independent first and second analysis models to perform biological species identification based on differentiated processing of images and content, respectively, and cross-validates through a species comparison module, thereby significantly reducing the probability of misjudgment. By generating consistent labels, the system not only ensures the reliability of the identification results, but also provides an opportunity for manual intervention in abnormal situations, thereby achieving an organic combination of automation and manual verification, and improving the accuracy and interpretability of biodiversity detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 A module diagram of a biodiversity detection system based on image recognition provided in Example 1 of the present invention; Figure 2 This is a flow chart of a biodiversity detection method based on image recognition provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0018] Example 1

[0019] See also Figure 1 As shown, this embodiment discloses a biodiversity detection system based on image recognition, and the method includes: The image acquisition module S110 is used to acquire biological image data in the monitoring area and pre-process the biological image data to generate identifiable data, wherein the identifiable data includes identifiable images and identifiable content; Specifically, the biological image data includes biological images of multiple viewing angles, high-resolution biological images and low-resolution biological images, and the collection cycle of the biological image data is once every n minutes; It should be noted that n minutes is determined according to the geographical scope of the monitoring area. The geographical scope is specifically the forest coverage of the monitoring area. The larger the geographical scope is, the smaller n needs to be set to prevent missed collection due to the large geographical scope. If the geographical scope is smaller, the larger n needs to be set. Image data of the monitoring area with different viewing angles, resolutions and time points are obtained through a variety of image acquisition devices. Specifically, the logic of preprocessing biological image data is: After each shooting cycle, synchronize the biological images of different perspectives according to the timestamp to ensure that the biological image data of multiple perspectives are aligned with the same time point, detect the difference in clarity and exposure in the biological images of multiple synchronized perspectives, select the biological image of the main perspective with the highest clarity, mark the images of other perspectives, and output the biological image of the main perspective as a recognizable image; Performing resolution adaptive processing on high-resolution biological images and low-resolution biological images through preset resolution differentiation rules to generate recognizable content; Specifically, the logic of the resolution differentiation rule is: The high-resolution biological image is sampled to a suitable size to reduce the amount of redundant data, and is combined with the low-resolution biological image to generate a multi-resolution biological image pyramid, each resolution layer of the multi-resolution biological image pyramid is used to process different required contents, and the processed different required contents are output as recognizable contents; It should be noted that the different processing requirements are determined according to the implementation of historical data, so as to achieve a balance between detail enhancement and overall shape, and the image registration algorithm includes but is not limited to affine and perspective transformation.

[0020] A first recognition module S120, for inputting a recognizable image into a preset first analysis model to obtain a first biological species; Specifically, the construction logic of the first analysis model is: Acquire historical image recognition training data, and divide the historical image recognition training data into an image recognition training set and an image recognition test set, wherein the historical image recognition training data includes identifiable images and their corresponding estimated first biological species; Constructing a first regression network, using the recognizable images in the image recognition training set as input data of the first regression network, and using the estimated first biological species in the image recognition training set as output data of the first regression network, training the first regression network, and obtaining an initial first machine learning network; The initial first machine learning network is model verified using an image recognition test set, and an initial first machine learning network that is less than or equal to a preset first test error threshold is output as a pre-constructed first analysis model.

[0021] The second recognition module S130 is used to input the identifiable content into a preset second analysis model to obtain a second biological species; Specifically, the construction logic of the second analysis model is: Acquire historical content recognition training data, and divide the historical content recognition training data into a content recognition training set and a content recognition test set, wherein the historical content recognition training data includes identifiable content and its corresponding estimated second biological species; Constructing a second regression network, using the identifiable content in the content recognition training set as input data of the second regression network, and using the estimated second biological species in the content recognition training set as output data of the second regression network, training the second regression network, and obtaining an initial second machine learning network; The initial second machine learning network is model verified using the content recognition test set, and an initial second machine learning network that is less than or equal to a preset second test error threshold is output as a pre-constructed second analysis model.

[0022] A variety comparison module S140 is used to compare the first biological variety with the second biological variety to obtain a biological consistency label, and output a biological variety result according to the biological consistency label; Specifically, the bio-consistent labels include species-consistent labels and species-different labels; Specifically, the logic of comparing the first biological species and the second biological species to obtain a biological consistency label is: If the first biological variety and the second biological variety are the same, a variety-consistent label is generated, and the biological variety corresponding to the first biological variety and the second biological variety is output as the biological variety result; if the first biological variety and the second biological variety are different, a variety-different label is generated, and a recognizable image is output for artificial biological variety determination.

[0023] The beneficial effect of this technology is that through multi-view and multi-resolution biological image acquisition and preprocessing, the comprehensiveness and accuracy of image data are improved, ensuring the clear presentation and feature capture of organisms at different time points and angles. Through timestamp synchronization and exposure comparison, the main view high-quality image is selected for analysis, which not only reduces redundant information, but also optimizes the quality of the detection model input, thereby improving the stability and accuracy of the detection; In addition, this system constructs independent first and second analysis models to identify biological species based on differentiated processing of images and content, respectively, and cross-validates through species comparison modules, significantly reducing the probability of misjudgment. By generating consistent labels, the system ensures the reliability of identification results while providing opportunities for manual intervention in abnormal situations, thus achieving an organic combination of automation and manual verification and improving the accuracy and interpretability of biodiversity detection.

[0024] Example 2

[0025] See also Figure 2 As shown, based on the unified inventive concept, this embodiment discloses a method for detecting biodiversity based on image recognition, the method comprising: S210: collecting biological image data in the monitoring area and preprocessing the biological image data to generate identifiable data, wherein the identifiable data includes identifiable images and identifiable content; Specifically, the biological image data includes biological images of multiple viewing angles, high-resolution biological images and low-resolution biological images, and the collection cycle of the biological image data is once every n minutes; Specifically, the logic of preprocessing biological image data is: After each shooting cycle, synchronize the biological images of different perspectives according to the timestamp to ensure that the biological image data of multiple perspectives are aligned with the same time point, detect the difference in clarity and exposure in the biological images of multiple synchronized perspectives, select the biological image of the main perspective with the highest clarity, mark the images of other perspectives, and output the biological image of the main perspective as a recognizable image; Performing resolution adaptive processing on high-resolution biological images and low-resolution biological images through preset resolution differentiation rules to generate recognizable content; Specifically, the logic of the resolution differentiation rule is: The high-resolution biological image is sampled to a suitable size to reduce the amount of redundant data, and is combined with the low-resolution biological image to generate a multi-resolution biological image pyramid. Each resolution layer of the multi-resolution biological image pyramid is used to process different required content, and the processed different required content is output as recognizable content.

[0026] S220: Inputting the recognizable image into a preset first analysis model to obtain a first biological species; Specifically, the construction logic of the first analysis model is: Acquire historical image recognition training data, and divide the historical image recognition training data into an image recognition training set and an image recognition test set, wherein the historical image recognition training data includes identifiable images and their corresponding estimated first biological species; Constructing a first regression network, using the recognizable images in the image recognition training set as input data of the first regression network, and using the estimated first biological species in the image recognition training set as output data of the first regression network, training the first regression network, and obtaining an initial first machine learning network; The initial first machine learning network is model verified using an image recognition test set, and an initial first machine learning network that is less than or equal to a preset first test error threshold is output as a pre-constructed first analysis model.

[0027] S230: Inputting the identifiable content into a preset second analysis model to obtain a second biological species; The construction logic of the second analysis model is: Acquire historical content recognition training data, and divide the historical content recognition training data into a content recognition training set and a content recognition test set, wherein the historical content recognition training data includes identifiable content and its corresponding estimated second biological species; Constructing a second regression network, using the identifiable content in the content recognition training set as input data of the second regression network, and using the estimated second biological species in the content recognition training set as output data of the second regression network, training the second regression network, and obtaining an initial second machine learning network; The initial second machine learning network is model verified using the content recognition test set, and an initial second machine learning network that is less than or equal to a preset second test error threshold is output as a pre-constructed second analysis model.

[0028] S240: Compare the first biological variety and the second biological variety to obtain a biological consistency label, and output a biological variety result according to the biological consistency label; Specifically, the bio-consistent labels include species-consistent labels and species-different labels; Specifically, the logic of comparing the first biological species and the second biological species to obtain a biological consistency label is: If the first biological variety and the second biological variety are the same, a variety-consistent label is generated, and the biological variety corresponding to the first biological variety and the second biological variety is output as the biological variety result; if the first biological variety and the second biological variety are different, a variety-different label is generated, and a recognizable image is output for artificial biological variety determination.

[0029] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired network or a wireless network. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD) or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0030] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0031] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0032] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a biodiversity detection system based on image recognition. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0033] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0034] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0035] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0036] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A biodiversity detection system based on image recognition, characterized in that: The system comprises: An image acquisition module, used to acquire biological image data in the monitoring area and pre-process the biological image data to generate recognizable data, wherein the recognizable data includes a recognizable image and recognizable content; A first recognition module, used for inputting the recognizable image into a preset first analysis model to obtain a first biological species; A second identification module, used for inputting the identifiable content into a preset second analysis model to obtain a second biological species; The variety comparison module is used to compare the first biological variety with the second biological variety to obtain a biological consistency label, and output a biological variety result according to the biological consistency label.

2. The biodiversity detection system based on image recognition according to claim 1, characterized in that: The biological image data includes biological images of multiple viewing angles, high-resolution biological images and low-resolution biological images, and the collection cycle of the biological image data is once every n minutes.

3. A biodiversity detection system based on image recognition according to claim 2, characterized in that: The logic of preprocessing biological image data is: After each shooting cycle, synchronize the biological images of different perspectives according to the timestamp to ensure that the biological image data of multiple perspectives are aligned with the same time point, detect the difference in clarity and exposure in the biological images of multiple synchronized perspectives, select the biological image of the main perspective with the highest clarity, mark the images of other perspectives, and output the biological image of the main perspective as a recognizable image; The high-resolution biological images and low-resolution biological images are adaptively processed by using preset resolution differentiation rules to generate recognizable content.

4. The biodiversity detection system based on image recognition according to claim 3, characterized in that: The logic of the resolution differentiation rule is: The high-resolution biological image is sampled to a suitable size to reduce the amount of redundant data, and is combined with the low-resolution biological image to generate a multi-resolution biological image pyramid. Each resolution layer of the multi-resolution biological image pyramid is used to process different required content, and the processed different required content is output as recognizable content.

5. The biodiversity detection system based on image recognition according to claim 4, characterized in that: The construction logic of the first analysis model is: Acquire historical image recognition training data, and divide the historical image recognition training data into an image recognition training set and an image recognition test set, wherein the historical image recognition training data includes identifiable images and their corresponding estimated first biological species; Constructing a first regression network, using the recognizable images in the image recognition training set as input data of the first regression network, and using the estimated first biological species in the image recognition training set as output data of the first regression network, training the first regression network, and obtaining an initial first machine learning network; The initial first machine learning network is model verified using an image recognition test set, and an initial first machine learning network that is less than or equal to a preset first test error threshold is output as a pre-constructed first analysis model.

6. The biodiversity detection system based on image recognition according to claim 5, characterized in that: The construction logic of the second analysis model is: Acquire historical content recognition training data, and divide the historical content recognition training data into a content recognition training set and a content recognition test set, wherein the historical content recognition training data includes identifiable content and its corresponding estimated second biological species; Constructing a second regression network, using the identifiable content in the content recognition training set as input data of the second regression network, and using the estimated second biological species in the content recognition training set as output data of the second regression network, training the second regression network, and obtaining an initial second machine learning network; The initial second machine learning network is model verified using the content recognition test set, and an initial second machine learning network that is less than or equal to a preset second test error threshold is output as a pre-constructed second analysis model.

7. The biodiversity detection system based on image recognition according to claim 6, characterized in that: The bio-consistency labels include species-consistent labels and species-different labels.

8. The biodiversity detection system based on image recognition according to claim 7, characterized in that: The logic of comparing the first biological species and the second biological species to obtain a biological consistency label is: If the first biological variety and the second biological variety are the same, a variety-consistent label is generated, and the biological variety corresponding to the first biological variety and the second biological variety is output as the biological variety result; if the first biological variety and the second biological variety are different, a variety-different label is generated, and a recognizable image is output for artificial biological variety determination.

9. A method for biodiversity detection based on image recognition, applied to a biodiversity detection system based on image recognition as claimed in claim 1, characterized in that: The method comprises: Collecting biological image data in the monitoring area and preprocessing the biological image data to generate identifiable data, wherein the identifiable data includes identifiable images and identifiable content; Inputting the recognizable image into a preset first analysis model to obtain a first biological species; Inputting the identifiable content into a preset second analysis model to obtain a second biological species; The first biological variety is compared with the second biological variety to obtain a biological consistency label, and a biological variety result is output according to the biological consistency label.

Images