Benthonic animal community structure identification method and system based on image identification

Abstract

The invention discloses a benthonic animal community structure recognition method and system based on image recognition, and relates to the technical field of benthonic animal community structure recognition, and the method comprises the steps: carrying out the image collection of a target water area through an underwater camera, and transmitting the collected image to a computer. According to the method, the image recognition technology is adopted, rapid and accurate recognition and classification of the benthonic animals are achieved, the working efficiency is greatly improved, recognition is carried out by utilizing the image recognition technology and combining features such as shapes, textures and colors of the benthonic animals, the confusion rate and the error rate in the classification and counting process are effectively reduced, and the classification accuracy is improved. The accuracy and reliability of the result are improved, the abnormal change of the zoobenthos community structure can be found in time by utilizing a real-time data analysis technology, timely early warning and response are provided for protecting the ecological environment of the water body, targeted suggestions and measures are automatically provided, a user is helped to better improve and protect the ecological environment of the water body, and the user experience is improved. And the scientificity and effectiveness of decision making are improved.

Description

Technical Field

[0001] The invention relates to the technical field of benthic animal community structure recognition, and in particular to a benthic animal community structure recognition method and system based on image recognition. Background Art

[0002] Benthic animals are a general term for animals in benthic organisms, including coelenterates, sponges, flatworms, nematodes, annelids, arthropods and other types. They are an important ecological type in aquatic organisms, mainly living at the bottom of the ocean or inland waters. Since benthic animals live in the bottom mud for a long time, they usually have little ability to avoid environmental pollution and changes. Therefore, the destruction and reconstruction of their communities require a relatively long time. It is of great significance to identify the community structure of benthic animals. The community structure of benthic animals is one of the important indicators for assessing the biodiversity of aquatic ecosystems. Changes in the community structure of benthic animals can directly reflect the health of aquatic ecosystems. Benthic animals have a long life history and stable living habits. Changes in their community structure can explain the cumulative effects of pollutants in water bodies. By analyzing the community structure of benthic animals, the trend of environmental quality changes in water bodies can be predicted. Identifying the community structure of benthic animals can provide a basis for water ecological restoration. Understanding the community structure of benthic animals in damaged water bodies can select appropriate ecological restoration measures to promote the recovery of aquatic ecosystems.

[0003] As an important part of aquatic ecosystems, benthic animals, whose community structure can directly reflect the ecological environment quality of water bodies, have always been a challenging task to identify and monitor benthic animals because they live at the bottom of water bodies. Although traditional benthic animal identification and community structure analysis methods can identify benthic animals, they do not have early warning functions and are difficult to detect abnormal changes in benthic animal community structure in a timely manner and provide effective countermeasures. Therefore, it is necessary to propose a benthic animal community structure identification method and system based on image recognition to solve the problems in the existing technology. Summary of the invention

[0004] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a method and system for identifying benthic animal community structure based on image recognition. It has the functions of early warning and providing targeted suggestions and measures based on analysis results. It can timely detect abnormal changes in the structure of benthic animal communities and provide effective response measures to protect the ecological environment of water bodies.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for identifying benthic animal community structure based on image recognition, the method comprising:

[0006] Capture images of the target water area through an underwater camera, and transmit the captured images to a computer;

[0007] Preprocessing the acquired images, including denoising, contrast enhancement, and color correction steps;

[0008] Using image recognition technology, the pre-processed images are used to identify benthic animals, and the types and quantities of benthic animals are obtained;

[0009] According to the identification results, the benthic animal community structure is analyzed, the number and proportion of various benthic animals are counted, and the ecological environment quality of the water body is evaluated;

[0010] Monitor changes in benthic animal community structure in real time, detect abnormal situations, and issue early warning notifications in a timely manner;

[0011] Provide targeted suggestions and measures based on the analysis results.

[0012] Furthermore, the method includes collecting images of the target water area by an underwater camera and transmitting the collected images to a computer, and comprising:

[0013] First, choose the underwater video equipment including waterproof camera, diving shell, and tripod;

[0014] Adjust the equipment settings according to the actual environment and install a tripod to stabilize the camera;

[0015] Place underwater cameras into target waters and control the cameras to shoot images at different angles, distances, and times to obtain more diverse data;

[0016] The acquired image data is transferred from the camera to the computer.

[0017] Furthermore, the method of preprocessing the collected image includes:

[0018] First, identify the noise type in the image, and select a denoising algorithm from mean filtering, median filtering, and Gaussian filtering according to the noise type to perform denoising processing;

[0019] Enhance contrast, calculate the histogram and cumulative histogram of the image, count the number of pixels and cumulative number of gray levels at each gray level, calculate the adjusted gray level, and apply it to the original image to obtain the equalized image, divide the image into small areas, and perform histogram equalization on each area to better adapt to areas with different brightness changes, calculate the minimum and maximum pixel values ​​in the original image, and map the pixel value range in the original image to the new range through linear mapping to enhance the contrast of the image;

[0020] Color correction, adjust the proportion of red, green and blue in the image to correct the color of the image and adjust the hue and saturation of the image.

[0021] Furthermore, the use of image recognition technology to identify benthic animals on the pre-processed image includes:

[0022] Collect positive sample images containing benthic animals and negative sample images not containing benthic animals, and select features that can represent benthic animals;

[0023] The edge detection image processing algorithm is used to extract the morphological, texture and color features of benthic animals. The edge detection algorithm mainly involves Gaussian filtering, gradient calculation, non-maximum suppression and double threshold detection. Gradient direction: Gradient Magnitude: G x and G y are the gradients of the image in the x and y directions respectively;

[0024] Classifier training, using labeled benthic animal images for training;

[0025] The image to be identified is input into the trained classifier, and the model outputs the species and quantity of benthic animals.

[0026] Furthermore, the specific steps of training a classifier for identifying benthic animals using image recognition technology include:

[0027] Divide the positive and negative samples into training sets, validation sets, and test sets;

[0028] The classification model of the convolutional neural network CNN is used. The structure of the convolutional neural network CNN includes a convolutional layer, a pooling layer, and a fully connected layer. The convolutional layer operation formula is: S(i,j)=(I*K)(i,j)=Σ m Σ n I(im,jn)K(m,n), where I is the input image, K is the convolution kernel, and S is the feature map after convolution;

[0029] Use the training set data to train the model and adjust the model parameters to optimize the model performance. During the training process, use the validation set data to monitor the performance of the model so that timely adjustments can be made when the model is overfitting or underfitting.

[0030] Furthermore, according to the identification results, the benthic animal community structure is analyzed, the number and proportion of various benthic animals are counted, and the ecological environment quality of the water body is evaluated. The method includes:

[0031] First, the benthic animal species and quantity data obtained by image recognition were organized into tables and database forms;

[0032] Count the number of each benthic animal, including the number of individuals and the frequency of occurrence;

[0033] Based on the statistical results, calculate the proportion of various benthic animals in the community and understand the relative importance of different species in the community;

[0034] Analyze the stability and diversity of the community, and use the complexity and diversity community structure parameters of the biological community to reflect the stability of the biological community;

[0035] Assess the diversity, number of species, biomass, productivity and food chain length of benthic animal communities to understand the complexity and stability level of the communities.

[0036] Furthermore, the changes in the benthic animal community structure are monitored in real time, abnormal conditions are detected, and early warning notifications are issued in a timely manner. The method includes:

[0037] Apply data analysis technology to monitor abnormalities in benthic animal community structure in real time, and use statistical methods to set thresholds. If the number of benthic animal species and the frequency of occurrence exceed the normal range, it will be judged as abnormal;

[0038] N i is defined as the number of the i-th benthic animal, N is the total number of all benthic animals, then the proportion p of the i-th benthic animal i for: By comparing the p i , to determine whether the community composition has changed significantly. When the number of benthic animal species and the frequency of occurrence exceed the normal range, it is considered abnormal;

[0039] The Shannon-Wiener diversity index was used to assess the diversity of benthic animal communities. Where S is the number of species in the community, p i is the proportion of the i-th species. When the diversity index changes significantly, it can also be judged as an abnormal situation;

[0040] When the above abnormal situations are detected, early warning notifications will be issued in a timely manner. The early warning information will be conveyed through mobile phone text messages, emails and system pop-up windows to ensure that relevant personnel can receive and respond in time.

[0041] Furthermore, targeted suggestions and measures are provided based on the analysis results. The method includes:

[0042] Analyze the diversity, number of species, biomass, productivity and food chain length indicators of benthic fauna;

[0043] Based on the analysis results, determine the ecological environment quality of the water body, including the pollution level and nutrient status of the water body;

[0044] Based on the above judgments, corresponding suggestions are generated, including adjusting water ecological restoration measures, strengthening water pollution control and protecting benthic animal habitats;

[0045] Based on the generated recommendations, specific measures and action plans are recommended. The measures and action plans include determining specific pollution control measures based on the pollution level of the water body and the relative proportion of sensitive and pollution-tolerant species of benthic animals, including key areas for control, control methods and required resources, and recommending specific measures to protect benthic animal habitats, including demarcating protected areas and improving habitat environments, in combination with the spatial distribution pattern of benthic animal communities and environmental factors;

[0046] Evaluate and predict the effects of measures and action plans taken, simulate the effects of measures after implementation by building models, and predict the changing trends of benthic animal community structure and the improvement of water ecological environment quality based on historical data;

[0047] Based on the evaluation results, timely adjust and optimize suggestions and measures.

[0048] A benthic animal community structure recognition system based on image recognition, the system comprising:

[0049] An image acquisition module is used to control the underwater camera to acquire images and transmit the acquired images to a computer;

[0050] An image preprocessing module is used to preprocess the collected images, including denoising, contrast enhancement and color correction steps;

[0051] The benthic animal recognition module uses image recognition technology to identify benthic animals in the preprocessed images and obtain the types and quantities of benthic animals;

[0052] The community structure analysis module is used to analyze the benthic animal community structure based on the identification results, count the number and proportion of various benthic animals, and evaluate the ecological environment quality of the water body;

[0053] The early warning module is used to monitor the changes in the benthic animal community structure in real time and issue early warning notifications in time when abnormal situations are detected;

[0054] The intelligent decision support module is used to provide targeted suggestions and measures based on the analysis results, and to evaluate and predict the effects of the measures.

[0055] Furthermore, the system also includes a user interface module, which provides an interface for users to interact with the system and is used to display images, recognition results, community structure analysis results, early warning information, and targeted suggestions and measures.

[0056] Compared with the prior art, the method and system for identifying benthic animal community structure based on image recognition have the following beneficial effects:

[0057] The present invention realizes rapid and accurate identification and classification of benthic animals by adopting image recognition technology, thereby greatly improving work efficiency. By utilizing image recognition technology and combining the morphology, texture, color and other characteristics of benthic animals for identification, the confusion and error rate in the classification and counting process are effectively reduced, and the accuracy and reliability of the results are improved. By utilizing real-time data analysis technology, abnormal changes in the community structure of benthic animals can be discovered in a timely manner, and timely warnings and responses can be provided for protecting the ecological environment of water bodies. Targeted suggestions and measures can be automatically provided to help users better improve and protect the ecological environment of water bodies and improve the scientificity and effectiveness of decision-making.

[0058] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 The figure is a flow chart of a method for identifying benthic animal community structure based on image recognition.

[0061] Figure 2 This is a schematic diagram of the structure of a benthic animal community structure recognition system based on image recognition. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.

[0063] Embodiment 1

[0064] A method for identifying benthic animal community structure based on image recognition, the method comprising:

[0065] The target water area is imaged by an underwater camera and the image is transmitted to a computer.

[0066] According to the actual conditions of the target waters (such as water depth, water quality, benthic animal species, etc.), select an appropriate underwater camera for image acquisition. For deep water areas, you may need to use an underwater camera with strong waterproof performance and good lighting capabilities. For shallow water or surface areas, a high-definition camera carried by a drone may be more appropriate.

[0067] Multiple collection points are planned in the target waters to ensure that the collected images can fully reflect the distribution of benthic animal communities. At the same time, considering the changes in benthic animal activities in different seasons and time periods, reasonable collection times and frequencies need to be formulated.

[0068] During the acquisition process, it is necessary to keep the camera stable to avoid image blur due to jitter or shaking. At the same time, attention should be paid to the sufficiency and uniformity of light to ensure that the quality of the acquired image meets the requirements of subsequent processing.

[0069] The method includes: collecting images of a target water area through an underwater camera, and transmitting the collected images to a computer. The method includes: selecting a waterproof camera, a diving shell, and a tripod to put into an underwater video recording device, ensuring that the device is in good working condition, the waterproof camera can work normally, and the diving shell is intact. Before performing underwater video recording, the underwater environment of the shooting location is understood, including water temperature, water quality, and light factors, and the impact of the environment on the device performance and shooting effect is evaluated, such as whether the water temperature is too high or too low, whether the water quality is clear, and whether the light is sufficient;

[0070] Adjust the settings of the equipment according to the actual environment, such as the camera's exposure, white balance, and focus parameters, and install a tripod to stabilize the camera to prevent image shaking caused by water flow; place the underwater camera in the target waters, control the camera to shoot, and ensure that clear and comprehensive images of benthic animals can be captured. Shoot images at different angles, distances, and times to obtain more diverse data;

[0071] The collected image data is transferred from the camera to the computer to ensure the stability and integrity of data transmission, to avoid image quality degradation and data loss due to data transmission errors, and to organize, classify and store the collected image data for subsequent processing and analysis. In actual operation, other factors may need to be considered, such as the waterproof depth of the underwater camera, battery life, and lighting conditions. In addition, in order to ensure the accuracy and efficiency of image acquisition, operators need to be trained and guided accordingly.

[0072] The acquired images were preprocessed, including denoising, contrast enhancement, and color correction steps.

[0073] Due to the complex underwater environment, various noises may exist in the collected images, such as uneven light, suspended particles, etc. The purpose of the denoising step is to eliminate these noises and improve the clarity of the image. Commonly used denoising methods include median filtering, Gaussian filtering, etc.

[0074] In order to highlight the contrast between benthic animals and the background, the image needs to be contrast enhanced, which helps to more accurately identify benthic animals in the subsequent feature extraction and recognition steps.

[0075] Due to the influence of underwater environment on light, the collected images may have color deviation. The purpose of the color correction step is to restore the true color of the image and make the characteristics of benthic animals more obvious. Commonly used color correction methods include white balance adjustment, color space conversion, etc.

[0076] The collected image is preprocessed. The method includes: identifying the noise type in the image, selecting a suitable denoising algorithm from mean filtering, median filtering, and Gaussian filtering according to the noise type for denoising, mean filtering: Where g(x,y) is the pixel value of the output image at (x,y0), f(m,n) is the pixel value of the input image at (m,n), S is the coordinate set in the filter window centered at (x,y) (for example, 3×3 or 5×5), N is the number of coordinates in S (i.e., the size of the filter), Gaussian filtering: g(x,y0=Σ (m,n)∈S f(m,n)·G(m,n;σ), where G(m,n;σ) is a two-dimensional Gaussian function defined as σ is the standard deviation of the Gaussian function, which determines the width of the filter;

[0077] Calculate the histogram and cumulative histogram of the image, count the number of pixels and cumulative number of gray levels at each gray level, calculate the adjusted gray level, and apply it to the original image to obtain the equalized image, divide the image into small areas, and perform histogram equalization on each area to better adapt to areas with different brightness changes, calculate the minimum and maximum pixel values ​​in the original image, and map the pixel value range in the original image to the new range through linear mapping to enhance the contrast of the image, contrast stretching: Among them, I(x,y) is the pixel value of the original image at point (x,y), I orig,max and I orig,min are the maximum and minimum pixel values ​​in the original image, respectively, max and I min are the maximum and minimum pixel values ​​you want to achieve after stretching, and g(x,y) is the pixel value after stretching;

[0078] Adjust the ratio of red, green and blue in the image to correct over-saturated and under-saturated colors. Adjust the hue and saturation of the image as needed to obtain more accurate color representation.

[0079] For example, convert from RGB color space to HSV or HSL color space to adjust the brightness and color of the image more intuitively.

[0080] Other preprocessing steps include image scaling: adjusting the image size as needed to facilitate subsequent processing or display, image cropping: removing unnecessary parts of the image to highlight the target area or remove irrelevant information, and grayscale conversion: if a grayscale image needs to be processed, the color image can be converted to a grayscale image. The purpose of these preprocessing steps is to improve the image quality and improve the accuracy and efficiency of subsequent image recognition and community structure analysis. The specific preprocessing steps and parameter settings can be adjusted according to actual conditions and needs.

[0081] Embodiment 2

[0082] The pre-processed images are used to identify benthic animals using image recognition technology, thereby obtaining the types and quantities of benthic animals, including: collecting positive sample images containing benthic animals and negative sample images not containing benthic animals, and selecting features that can represent benthic animals.

[0083] The edge detection image processing algorithm is used to extract the morphological, texture and color features of benthic animals. The edge detection algorithm mainly involves Gaussian filtering, gradient calculation, non-maximum suppression and double threshold detection. Gradient direction: Gradient Magnitude: G x and G y are the gradients of the image in the x and y directions, respectively.

[0084] Classifier training, using labeled benthic animal images for training, dividing positive samples and negative samples into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance. The classification model of the convolutional neural network CNN is selected. The structure of the convolutional neural network CNN includes convolution layer, pooling layer, and fully connected layer, among which the convolution layer is the key. The convolution operation: S(i,j)=(I*K)(i,j)=Σ m Σ n I(im,jn)K(m,n), where I is the input image, K is the convolution kernel, and S is the feature map after convolution.

[0085] Use the training set data to train the model, and optimize the model performance by adjusting the model parameters (such as learning rate and number of iterations). During the training process, use the validation set data to monitor the performance of the model so that timely adjustments can be made when the model is overfitting or underfitting.

[0086] The image to be identified is input into the trained classifier, and the model outputs the species and quantity of benthic animals. In order to improve the accuracy of identification, multi-classifier fusion, ensemble learning and other technologies can be used.

[0087] In the recognition stage, the trained classifier model is loaded; the image data to be recognized is prepared, and these images undergo the same preprocessing steps as the training data; the images to be recognized are input into the trained classifier model, and the model outputs the type and number of benthic animals in each image (if the model can handle multi-target recognition).

[0088] Ensuring the quality and diversity of training data is crucial to improving recognition accuracy. During the training process, you may need to try different model structures and parameter settings to find the optimal model. In practical applications, it is necessary to consider the real-time requirements of recognition to ensure that the system can complete the recognition task within a reasonable time.

[0089] Through the above steps, the benthic animals can be identified on the preprocessed images, thereby obtaining the types and quantities of the benthic animals.

[0090] Based on the identification results, the benthic animal community structure is analyzed, the number and proportion of various benthic animals are counted, and the ecological environmental quality of the water body is evaluated.

[0091] Based on the identification results, the number and proportion of various benthic animals are counted, which helps to understand the distribution of different types of benthic animals in the community.

[0092] The stability of benthic animal communities can be assessed by analyzing their species composition, quantity changes and other indicators. Stable communities usually have higher species diversity and lower species turnover rates.

[0093] The ecological environmental quality of water bodies is evaluated based on the structure and composition of benthic animal communities. As an important component of aquatic ecosystems, the community structure and composition of benthic animals can reflect information such as the degree of pollution and nutritional status of water bodies.

[0094] First, the benthic animal species and quantity data obtained by image recognition are organized into tables and databases for subsequent statistics and analysis. The number of each benthic animal is counted, including the number of individuals and the frequency of occurrence. Based on the statistical results, the proportion of various benthic animals in the community is calculated to understand the relative importance of different species in the community.

[0095] Benthic animals are used as indicator organisms to reflect the ecological environmental quality of water bodies through their species composition and quantity changes. The relative proportion of sensitive species and pollution-resistant species can indicate the degree of pollution and ecological health of water bodies. The composition and spatial distribution pattern of benthic animal communities are analyzed, and the relationship between community structure and environmental factors is understood. The ecological environmental quality of water bodies is comprehensively evaluated by combining environmental factors (such as temperature, humidity, light, soil nitrogen and phosphorus content) and biological species interaction factors.

[0096] Through the above steps, the benthic animal community structure can be analyzed comprehensively and systematically, providing a scientific basis for the ecological environment protection and management of water bodies.

[0097] The method includes: applying data analysis technology to monitor the abnormal situation of benthic animal community structure in real time, using statistical methods to set thresholds, and judging that the number and frequency of benthic animal species exceed the normal range as abnormal, and N i is defined as the number of the i-th benthic animal, N is the total number of all benthic animals, then the proportion p of the i-th benthic animal i for: By comparing the p i , to determine whether the community composition has changed significantly. When the number of benthic animal species and the frequency of occurrence indicators exceed the normal range, it is judged as abnormal.

[0098] The Shannon-Wiener diversity index was used to assess the diversity of benthic animal communities. Where S is the number of species in the community, p i is the proportion of the i-th species. When the diversity index changes significantly, it can also be judged as an abnormal situation. When the above abnormal situation is detected, an early warning notification is issued in time. The early warning information is conveyed through mobile phone text messages, emails and system pop-ups to ensure that relevant personnel can receive and respond in time.

[0099] Providing targeted suggestions and measures based on the analysis results. The method includes: analyzing the species diversity, species number, biomass, productivity and food chain length indicators of benthic animal communities; judging the ecological environmental quality of water bodies, including the degree of pollution and nutritional status of water bodies, based on the analysis results; and generating corresponding suggestions based on the above judgments, including adjusting water ecological restoration measures, strengthening water pollution control and protecting benthic animal habitats.

[0100] Based on the generated recommendations, specific measures and action plans are recommended. The measures and action plans include determining specific pollution control measures, including key areas for control, control methods and required resources, based on the pollution level of the water body and the relative proportion of sensitive and pollution-tolerant benthic animals. In addition, specific measures for protecting benthic animal habitats are recommended, including demarcating protected areas and improving habitat environments, based on the spatial distribution pattern of benthic animal communities and environmental factors (such as temperature, humidity, light, and soil nitrogen and phosphorus content);

[0101] Evaluate and predict the effects of measures and action plans taken; build models to simulate the effects of implemented measures; predict changing trends in benthic animal community structure and improvements in water ecological environment quality based on historical data; and adjust and optimize recommendations and measures in a timely manner based on evaluation results to ensure continuous improvement and protection of water ecological environment.

[0102] Through the above steps, the present invention has the functions of early warning and providing targeted suggestions and measures based on the analysis results, can timely discover abnormal changes in the benthic animal community structure, and provide effective response measures to protect the water ecological environment.

[0103] Embodiment 3

[0104] A benthic animal community structure recognition system based on image recognition, the system comprising:

[0105] An image acquisition module is used to control the underwater camera to acquire images and transmit the acquired images to a computer;

[0106] The image acquisition module integrates multiple types of underwater cameras to adapt to the needs of different water depths, light and ecological environments. It is equipped with a high-resolution camera to capture clear images of benthic animals and a stable gimbal and hovering system to ensure that the drone remains stable when collecting images.

[0107] According to the characteristics of the target waters, formulate appropriate image acquisition strategies, such as acquisition area and time interval. Support users to customize acquisition parameters to meet the needs of different research projects.

[0108] Implement efficient image transmission protocols to ensure that image data can be transmitted to the computer in real time and stably, be equipped with sufficient storage space to store large amounts of image data, and support data backup and recovery.

[0109] An image preprocessing module is used to preprocess the collected images, including denoising, contrast enhancement and color correction steps;

[0110] The image preprocessing module adopts advanced denoising algorithms, such as adaptive median filtering and bilateral filtering, to effectively remove noise from the image, uses contrast enhancement algorithms, such as histogram equalization and Retinex algorithm, to improve the contrast of the image, and applies color correction algorithms, such as white balance adjustment and color space conversion, to restore the true color of the image.

[0111] Design an automated pre-processing process to reduce manual intervention, improve processing efficiency, support batch processing, and process multiple images simultaneously.

[0112] The benthic animal recognition module uses image recognition technology to identify benthic animals in preprocessed images, thereby obtaining the types and quantities of benthic animals. It includes: a feature extraction unit that extracts the morphological, texture and color features of benthic animals; a classifier training unit that uses a large number of labeled benthic animal images for training; and a recognition unit that inputs the image to be identified into the trained classifier to obtain the types and quantities of benthic animals.

[0113] The benthic animal recognition module combines the morphology, texture, color and other characteristics of benthic animals to design an effective feature extraction method, and uses the deep learning technology of convolutional neural network (CNN) to automatically learn the characteristic representation of benthic animals.

[0114] Design a high-precision classifier to classify benthic animals and use a large amount of labeled image data to train the classifier to improve recognition accuracy.

[0115] It realizes fast image recognition and classification process, reduces processing time, supports real-time recognition, and processes acquired images instantly.

[0116] The community structure analysis module is used to analyze the benthic animal community structure based on the identification results, count the number and proportion of various benthic animals, analyze the stability and diversity of the community, and evaluate the ecological environmental quality of the water body.

[0117] Count the number and proportion of benthic animals, and apply ecological and biodiversity assessment indicators such as species richness, evenness, diversity index, etc. to analyze the stability and diversity of the community.

[0118] The early warning module monitors the changes in the benthic animal community structure in real time and issues early warning notifications when abnormal situations are detected. The intelligent decision support module provides targeted suggestions and measures based on the analysis results, and evaluates and predicts the effectiveness of the measures. Combined with the benthic animal community structure data, it provides environmental quality early warning and trend forecasting functions to help users discover ecological and environmental problems in a timely manner.

[0119] The system also includes a user interface module, which provides an interface for users to interact with the system and is used to display images, recognition results and community structure analysis results.

[0120] The user interface module designs an intuitive and easy-to-use user interface to reduce the user's learning cost and provides rich interactive functions such as image zooming, panning, rotation, etc., making it convenient for users to view images.

[0121] The collected images, identification results and community structure analysis results are displayed in real time in the user interface, and the results can be exported to multiple formats for users to save and share.

[0122] Establish a user feedback mechanism, collect users' opinions and suggestions on the system, and continuously optimize system functions and performance based on user feedback.

[0123] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the same elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for identifying benthic animal community structure based on image recognition, characterized in that: The method includes: Capture images of the target water area through an underwater camera, and transmit the captured images to a computer; Preprocessing the acquired images, including denoising, contrast enhancement, and color correction steps; Using image recognition technology, the pre-processed images are used to identify benthic animals, and the types and quantities of benthic animals are obtained; According to the identification results, the benthic animal community structure is analyzed, the number and proportion of various benthic animals are counted, and the ecological environment quality of the water body is evaluated; Monitor changes in benthic animal community structure in real time, detect abnormal situations, and issue early warning notifications in a timely manner; Provide targeted suggestions and measures based on the analysis results.

2. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: The method includes collecting images of the target water area by an underwater camera and transmitting the collected images to a computer. The method includes: First, choose the underwater video equipment including waterproof camera, diving shell, and tripod; Adjust the equipment settings according to the actual environment and install a tripod to stabilize the camera; Place underwater cameras into target waters and control the cameras to shoot images at different angles, distances, and times to obtain more diverse data; The acquired image data is transferred from the camera to the computer.

3. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: The method of preprocessing the collected image comprises: First, identify the noise type in the image, and select a denoising algorithm from mean filtering, median filtering, and Gaussian filtering according to the noise type to perform denoising processing; Enhance contrast, calculate the histogram and cumulative histogram of the image, count the number of pixels and cumulative number of gray levels at each gray level, calculate the adjusted gray level, and apply it to the original image to obtain the equalized image, divide the image into small areas, and perform histogram equalization on each area to better adapt to areas with different brightness changes, calculate the minimum and maximum pixel values ​​in the original image, and map the pixel value range in the original image to the new range through linear mapping to enhance the contrast of the image; Color correction, adjust the proportion of red, green and blue in the image to correct the color of the image and adjust the hue and saturation of the image.

4. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: The method of using image recognition technology to identify benthic animals on the pre-processed image includes: Collect positive sample images containing benthic animals and negative sample images not containing benthic animals, and select features that can represent benthic animals; The edge detection image processing algorithm is used to extract the morphological, texture and color features of benthic animals. The edge detection algorithm mainly involves Gaussian filtering, gradient calculation, non-maximum suppression and double threshold detection. Gradient direction: Gradient Magnitude: G x and G y are the gradients of the image in the x and y directions respectively; Classifier training, using labeled benthic animal images for training; The image to be identified is input into the trained classifier, and the model outputs the species and quantity of benthic animals.

5. The method for identifying benthic animal community structure based on image recognition according to claim 4, characterized in that: The specific steps of using the image recognition technology to train the classifier when identifying benthic animals on the pre-processed image include: Divide the positive and negative samples into training sets, validation sets, and test sets; The classification model using convolutional neural network CNN includes convolutional layer, pooling layer and fully connected layer. The calculation formula of convolutional layer is: S(i,j)=(I*K)(i,j)=Σ m Σ n I(im,jn)K(m,n), where I is the input image, K is the convolution kernel, and S is the feature map after convolution; Use the training set data to train the model and adjust the model parameters to optimize the model performance. During the training process, use the validation set data to monitor the performance of the model so that timely adjustments can be made when the model is overfitting or underfitting.

6. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: According to the identification results, the benthic animal community structure is analyzed, the number and proportion of various benthic animals are counted, and the ecological environment quality of the water body is evaluated. The method includes: First, the benthic animal species and quantity data obtained by image recognition were organized into tables and database forms; Count the number of each benthic animal, including the number of individuals and the frequency of occurrence; Based on the statistical results, calculate the proportion of various benthic animals in the community and understand the relative importance of different species in the community; Analyze the stability and diversity of the community, and use the complexity and diversity community structure parameters of the biological community to reflect the stability of the biological community; Assess the diversity, number of species, biomass, productivity and food chain length of benthic animal communities to understand the complexity and stability level of the communities.

7. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: Real-time monitoring of changes in benthic animal community structure, detection of abnormal conditions, and timely issuance of early warning notifications, the method includes: Apply data analysis technology to monitor abnormalities in benthic animal community structure in real time, and use statistical methods to set thresholds. If the number of benthic animal species and the frequency of occurrence exceed the normal range, it will be judged as abnormal; N i is defined as the number of the i-th benthic animal, N is the total number of all benthic animals, then the proportion p of the i-th benthic animal i for: By comparing the p i , to determine whether the community composition has changed significantly. When the number of benthic animal species and the frequency of occurrence exceed the normal range, it is considered abnormal; The Shannon-Wiener diversity index was used to assess the diversity of benthic animal communities. Where S is the number of species in the community, p i is the proportion of the i-th species. When the diversity index changes significantly, it can also be judged as an abnormal situation; When the above abnormal situations are detected, early warning notifications will be issued in a timely manner, and the early warning information will be conveyed through methods including mobile phone text messages, emails and system pop-up windows.

8. The method for identifying benthic animal community structure based on image recognition according to claim 1, characterized in that: Providing targeted recommendations and measures based on the analysis results. The approach includes: Analyze the diversity, number of species, biomass, productivity and food chain length indicators of benthic fauna; Based on the analysis results, judge the ecological environment quality of the water body, including the pollution level and nutrient status of the water body; Based on the above judgments, corresponding suggestions are generated, including adjusting water ecological restoration measures, strengthening water pollution control and protecting benthic animal habitats; Based on the generated recommendations, specific measures and action plans are recommended. The measures and action plans include determining specific pollution control measures based on the pollution level of the water body and the relative proportion of sensitive and pollution-tolerant species of benthic animals, including key areas for control, control methods and required resources, and recommending specific measures to protect benthic animal habitats, including demarcating protected areas and improving habitat environments, in combination with the spatial distribution pattern of benthic animal communities and environmental factors; Evaluate and predict the effects of measures and action plans taken, simulate the effects of measures after implementation by building models, and predict the changing trends of benthic animal community structure and the improvement of water ecological environment quality based on historical data; Based on the evaluation results, timely adjust and optimize suggestions and measures.

9. A benthic animal community structure recognition system based on image recognition, characterized in that: The system includes: An image acquisition module is used to control the underwater camera to acquire images and transmit the acquired images to a computer; An image preprocessing module is used to preprocess the collected images, including denoising, contrast enhancement and color correction steps; The benthic animal recognition module uses image recognition technology to identify benthic animals in the preprocessed images and obtain the types and quantities of benthic animals; The community structure analysis module is used to analyze the benthic animal community structure based on the identification results, count the number and proportion of various benthic animals, and evaluate the ecological environment quality of the water body; The early warning module is used to monitor the changes in the benthic animal community structure in real time and issue early warning notifications in time when abnormal situations are detected; The intelligent decision support module is used to provide targeted suggestions and measures based on the analysis results, and to evaluate and predict the effects of the measures.

10. The system for identifying benthic animal community structure based on image recognition according to claim 9, characterized in that: The system also includes a user interface module, which provides an interface for users to interact with the system and is used to display images, recognition results, community structure analysis results, early warning information, and targeted suggestions and measures.

Images