Image recognition-based endangered animal recognition feedback system, method and related device

Abstract

The invention discloses an endangered animal recognition feedback system and method based on image recognition and a related device, and relates to the technical field of endangered animal recognized.The system comprises a man-machine interaction module, an image recognition module and a knowledge base module, and the man-machine interaction module is used for receiving a to-be-detected image, input by a user, of a to-be-detected endangered animal; the image recognition module is used for performing image recognition on the to-be-detected image and determining species of the to-be-detected endangered animal, and the man-machine interaction module is further used for calling detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal. And displaying the species of the to-be-detected endangered animal and detailed information of the species of the to-be-detected endangered animal, wherein the detailed information comprises pictures, names, protection levels, morphological characteristics, common transaction categories, common illegal utilization forms, values and distribution places. The method can efficiently and accurately complete the recognition and feedback of the endangered animal based on image recognition, and is timely and reliable.

Description

Technical Field

[0001] This application relates to the technical field of endangered animal identification, and particularly to an endangered animal identification and feedback system, method and related device based on image recognition. Background Art

[0002] Criminal acts involving endangered animals increasingly threaten global biodiversity, disrupt the ecological balance, and undermine the achievements of ecological civilization construction. Illegal hunting, transportation, and trading of endangered animals exacerbate the risk of species extinction, seriously affecting global ecological security. Accurately identifying the species attributes of endangered animal specimens has become an important link. For common endangered animals, morphological and DNA identification technologies have been widely used in domestic and foreign forensic appraisals. However, at present, the image recognition of endangered animals mainly relies on manual operation, which is not only inefficient but also difficult, and it is difficult to meet the needs of actual protection work. At the same time, the police officers responsible for on-site handling have limited experience and may not be able to quickly and accurately distinguish species. Therefore, it is particularly important to build a reliable and timely species identification and feedback system. Summary of the Invention

[0003] The purpose of this application is to provide an endangered animal identification and feedback system, method and related device based on image recognition, which can efficiently and accurately complete the identification and feedback of endangered animals based on image recognition, and is timely and reliable.

[0004] To achieve the above purpose, this application provides the following solutions:

[0005] In the first aspect, this application provides an endangered animal identification and feedback system based on image recognition. The endangered animal identification and feedback system based on image recognition includes: a human-computer interaction module, an image recognition module, and a knowledge base module. The human-computer interaction module and the image recognition module are both communicatively connected to the knowledge base module. The knowledge base module stores detailed information of each species among multiple species, and the detailed information includes pictures, names, protection levels, morphological characteristics, common trading categories, common illegal utilization forms, values, and distribution areas.

[0006] The human-computer interaction module is used to receive the to-be-detected image of the endangered animal input by the user.

[0007] The image recognition module is communicatively connected to the human-computer interaction module. The image recognition module is used to perform image recognition on the to-be-detected image to determine the species of the to-be-detected endangered animal.

[0008] The human-computer interaction module is further used to retrieve the detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal, and display the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal.

[0009] Optionally, the image recognition module includes a target detection model and a species classification model connected in sequence;

[0010] The target detection model is used to identify the endangered animals to be detected in the image to be detected, and crop the area where the endangered animals to be detected are located to obtain an animal recognition image, where the animal recognition image is a partial image of the area where the endangered animals to be detected are located in the image to be detected;

[0011] The species classification model is used to classify the animal recognition image and determine the species of the endangered animals to be detected.

[0012] Optionally, the target detection model uses the YOLOv8 model, and the species classification model uses the ResNet model.

[0013] Optionally, the human-computer interaction module is also used to update and query the content stored in the knowledge base module based on user input.

[0014] Optionally, the knowledge base module is located in the internal network.

[0015] Optionally, the endangered animal recognition feedback system based on image recognition further includes: a background management module, and the background management module is communicatively connected to the human-computer interaction module, the image recognition module, and the knowledge base module respectively;

[0016] The human-computer interaction module is also used to receive a new data set input by the user; the data set includes a plurality of samples and labels corresponding to each sample, the sample is an image of an endangered animal, and the label is the species of the endangered animal;

[0017] The background management module is used to retrieve the historical data set from the knowledge base module, and fine-tune the target detection model and the species classification model in the image recognition module by using the new data set and the historical data set;

[0018] The knowledge base module is used to store the new data set.

[0019] Optionally, fine-tuning the target detection model and the species classification model in the image recognition module by using the new data set and the historical data set specifically includes:

[0020] Preprocess the sample to obtain a preprocessed sample; the preprocessing includes resolution enhancement, angle adjustment, brightness adjustment, background blurring, denoising, and normalization;

[0021] Using the preprocessed sample and the label as inputs, and fine-tuning the target detection model and the species classification model in the image recognition module by using the method of incremental learning.

[0022] In a second aspect, the present application provides an endangered animal recognition feedback method based on image recognition, which is applied to the above-mentioned endangered animal recognition feedback system based on image recognition. The endangered animal recognition feedback method based on image recognition includes:

[0023] Obtain a to-be-detected image of the to-be-detected endangered animal input by the user;

[0024] Perform image recognition on the to-be-detected image to determine the species of the to-be-detected endangered animal;

[0025] Retrieve the detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal, and display the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal.

[0026] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor executes the computer program to implement the above-mentioned endangered animal recognition feedback method based on image recognition.

[0027] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the above-mentioned endangered animal recognition feedback method based on image recognition.

[0028] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0029] The present application provides an endangered animal recognition feedback system, method and related device based on image recognition, including: a human-computer interaction module, an image recognition module, and a knowledge base module. The human-computer interaction module and the image recognition module are both communicatively connected to the knowledge base module. The knowledge base module stores the detailed information of each species among multiple species, and the detailed information includes pictures, names, protection levels, morphological characteristics, common trading categories, common illegal utilization forms, values, and distribution areas. The human-computer interaction module is used to receive the to-be-detected image of the to-be-detected endangered animal input by the user, the image recognition module is used to perform image recognition on the to-be-detected image to determine the species of the to-be-detected endangered animal, and the human-computer interaction module is also used to retrieve the detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal, and display the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal. The present application can determine the species of the to-be-detected endangered animal based on image recognition, and feedback the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal to the user, efficiently and accurately complete the recognition and feedback of endangered animals. Compared with the method of manually performing image recognition, it has high efficiency, low difficulty, high accuracy, and is timely and reliable. Brief Description of the Drawings

[0030] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0031] Figure 1 It is a schematic structural diagram of an endangered animal recognition and feedback system based on image recognition provided in Embodiment 1 of the present application.

[0032] Figure 2 It is a schematic architecture diagram of an endangered animal recognition and feedback system based on image recognition provided in Embodiment 1 of the present application.

[0033] Figure 3 It is a display schematic diagram of the human-computer interaction module provided in Embodiment 1 of the present application.

[0034] Figure 4 It is a schematic diagram of the data set provided in Embodiment 1 of the present application.

[0035] Figure 5 It is a schematic diagram of the network structure of the YOLOv8 model provided in Embodiment 1 of the present application.

[0036] Figure 6 It is a schematic diagram of the network structure of the ResNet model provided in Embodiment 1 of the present application.

[0037] Figure 7 It is a schematic diagram of the recognition result provided in Embodiment 1 of the present application; among them, Figure 7 (a) in it is the species, Figure 7 (b) in it is the detailed information of the species.

[0038] Figure 8 It is a schematic diagram of the knowledge base module provided in Embodiment 1 of the present application.

[0039] Figure 9 It is a schematic diagram of incremental learning provided in Embodiment 1 of the present application.

[0040] Figure 10 It is a schematic flowchart of a method for recognizing and providing feedback on endangered animals based on image recognition provided in Embodiment 2 of the present application.

[0041] Figure 11 It is a schematic structural diagram of a computer device provided in Embodiment 3 of the present application. Detailed Embodiments

[0042] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0043] Embodiment 1

[0044] This embodiment is dedicated to developing innovative image recognition technology, constructing a comprehensive and intelligent endangered animal recognition and feedback system, significantly improving the recognition efficiency and accuracy of endangered animal species by using deep learning algorithms, training with a large amount of image data to make up for the deficiencies of existing manual recognition, and ensuring the rapid and accurate recognition of endangered animals and their related products in complex environments.

[0045] As Figure 1 shown, this embodiment provides an endangered animal recognition and feedback system based on image recognition. The endangered animal recognition and feedback system based on image recognition includes: a human-computer interaction module, an image recognition module, and a knowledge base module. The human-computer interaction module and the image recognition module are both communicatively connected to the knowledge base module. The knowledge base module stores detailed information about each species among multiple species, and the detailed information includes pictures, names, protection levels, morphological characteristics, common trading categories, common illegal utilization forms, values, and distribution areas.

[0046] The human-computer interaction module is used to receive the to-be-detected image of the endangered animal input by the user.

[0047] The image recognition module is communicatively connected to the human-computer interaction module, and the image recognition module is used to perform image recognition on the to-be-detected image to determine the species of the to-be-detected endangered animal.

[0048] The human-computer interaction module is further used to retrieve the detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal, and display the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal.

[0049] Currently, there are obvious deficiencies in the construction of the intelligent policing system for combating crimes involving endangered animals. At present, there is no service system with good data security, high research and judgment accuracy, and strong pertinence, making it difficult to effectively prevent and combat wildlife crimes. Therefore, it is urgent to develop a highly professional and secure identification and feedback system to provide a fast and accurate application system for front-line law enforcement officers. In terms of image recognition, this embodiment will adopt advanced image recognition technologies such as YOLOv8 and ResNet (Residual Network), which will effectively improve the detection and classification capabilities of endangered animal species in images and ensure high accuracy under changing environmental conditions. At the same time, in order to effectively manage and process data, this embodiment has established a highly secure and professional knowledge base module, which centrally stores monitoring data, identification results, and case information related to endangered animals. This knowledge base module has functions of rapid retrieval and efficient analysis, helping law enforcement officers quickly obtain the required information during case handling. Through the combination and innovation of the above technical means, it is possible to effectively curb crimes involving endangered animals, protect global biodiversity, and promote the sustainable development of ecological civilization construction.

[0050] This embodiment intends to construct an intelligent identification and feedback system for difficult-to-detect samples of endangered animals based on deep learning technology. By integrating the existing image data of physical evidence of endangered animals, this embodiment decides to use the image data, annotation data of difficult-to-detect samples, and public species classification as data sources, as Figure 2 shown below. Next, a detailed introduction to the endangered animal identification and feedback system based on image recognition used in this embodiment will be given:

[0051] (1) Human-computer interaction module

[0052] As Figure 3 shown, the human-computer interaction module of this embodiment is used to receive the image to be detected of the endangered animal input by the user, retrieve the detailed information of the species of the endangered animal to be detected from the knowledge base module based on the species of the endangered animal to be detected, and display the species of the endangered animal to be detected and the detailed information of the species of the endangered animal to be detected.

[0053] To ensure the long-term sustainability of the image recognition model (that is, the model that performs image recognition on the image to be detected to determine the species of the endangered animal to be detected, that is, the target detection model and species classification model connected in sequence in this embodiment), this embodiment has specially designed a convenient user interaction update function, aiming to enable users to easily participate in the optimization process of the image recognition model through a friendly front-end interface. Users only need to perform simple operations to upload a new data set to fine-tune the image recognition model, which directly affects the performance and recognition ability of the image recognition model.

[0054] The front - end interface is designed intuitively, providing a good user experience. Users can quickly select the image files to be uploaded and attach relevant description information, that is, upload a new data set, which facilitates subsequent processing by the system. This convenience reduces the technical threshold for users, enabling non - professional users to effectively participate in data collection and model optimization.

[0055] The newly uploaded data set will undergo pre - processing, including format conversion, size adjustment, and standardization, to ensure its compliance with the requirements of model training. Subsequently, the system will use the YOLO algorithm for real - time object detection, accurately identifying and cropping the relevant species in the image. This process is not only efficient but also capable of handling complex backgrounds and multi - species scenarios, thus ensuring the quality of the generated data set. The generated high - quality data set will be combined with the existing data set, and a fine - tuning strategy will be adopted to retrain the image recognition model and update the model's weights. This method not only improves the recognition accuracy of the image recognition model but also expands the types of species it can recognize, ensuring that the image recognition model can maintain good adaptability in a changing environment.

[0056] The convenient front - end interface and interactive design not only enrich the data sources but also enhance users' sense of participation, forming a dynamically optimized system. In this way, users can more directly influence the performance of the image recognition model, ensuring its efficiency and accuracy in dealing with the ever - changing wildlife protection needs. This collaborative mechanism not only improves the intelligence level of image recognition technology but also strengthens the close connection between humans and technology, providing solid support for protecting endangered species.

[0057] (2) Image Recognition Module

[0058] In this embodiment, advanced image recognition technologies such as YOLOv8 and ResNet are adopted. These image recognition technologies will effectively improve the detection and classification capabilities of endangered animal species in images, ensuring high accuracy under changing environmental conditions.

[0059] At this time, in this embodiment, the image recognition module includes a target detection model and a species classification model connected in sequence. The target detection model is used to identify the endangered animals to be detected in the image to be detected and crop the area where the endangered animals to be detected are located, obtaining an animal recognition image. The animal recognition image is a partial image of the area where the endangered animals to be detected are located in the image to be detected. The species classification model is used to classify the animal recognition image to determine the species of the endangered animals to be detected.

[0060] Among them, the target detection model adopts the YOLOv8 model, and the species classification model adopts the ResNet model.

[0061] In this embodiment, the YOLOv8 model and the ResNet model need to be pre-trained to obtain a target detection model and a species classification model. Specifically, the training data mainly comes from the datasets provided by the criminal investigation, food and drug, and environmental protection departments, as well as the data collected by web crawlers, that is, images of difficult and endangered animal specimens are collected, including deeply processed animal products, such as Figure 4 As shown, the images are processed, including operations such as resolution enhancement, angle adjustment, brightness adjustment, background blurring, denoising, and normalization, to ensure that the input image data meets the model requirements and enhance the generalization ability of the model.

[0062] The target detection model is responsible for detecting the relevant parts of endangered animals from the images. Specifically, the YOLOv8 model is used to achieve this goal. The YOLOv8 model is an efficient target detection model, known for its fast detection speed and high accuracy, and is particularly suitable for real-time application scenarios. By inputting the images into the YOLOv8 model, it can quickly identify and accurately locate the areas where animals or animal products are located, providing accurate basic data for the species classification model.

[0063] The species classification model aims to accurately classify the species of the animal regions extracted by the target detection model. It is based on advanced convolutional neural network variant models such as the ResNet model. By training and optimizing these networks, it can efficiently identify the species to which the animals in the images belong.

[0064] Therefore, in the image recognition module, the YOLOv8 model is used for preliminary localization and detection of the specimen area. Its fast processing speed makes real-time applications possible. The characteristics of the YOLOv8 model enable it to maintain high accuracy when processing complex backgrounds, thus providing a clear image input for subsequent classification. The ResNet model further improves the classification accuracy through its unique residual learning framework, capturing the feature information of the images from multiple dimensions and extracting more detailed features, which is crucial for the accurate identification of difficult specimens. Combining the efficient detection ability of the YOLOv8 model and the deep feature extraction of the ResNet model, the image recognition module can achieve efficient and accurate species classification of various specimens, providing strong technical support for the protection of endangered animals.

[0065] Such as Figure 5As shown, in this embodiment, the YOLOv8 model is used for object detection. The main goal is to train a model specifically for locating endangered animals. The design of the YOLOv8 model aims to quickly identify animals of interest in images, such as birds, bears, tigers, etc., without disturbing other irrelevant parts of the image. By constructing a diverse training set, including animal images in different scenarios and backgrounds, the generalization ability of the model is enhanced. The output of this model will provide high-quality animal location information for subsequent processing to ensure that animal parts can be accurately extracted from complex environments. The entire object detection process includes steps such as data preprocessing, model training, evaluation, and optimization to ensure its accuracy and real-time performance, ultimately achieving an efficient animal detection effect.

[0066] After completing object detection, deep learning techniques are required to classify the detected animals. In this embodiment, the ResNet model is selected, as Figure 6 shown. This is a proven variant of the convolutional neural network model, renowned for its powerful feature extraction ability. The ResNet model effectively solves the problem of vanishing gradients in the training of deep networks by introducing a residual learning framework, enabling the construction of deeper network structures and improving the accuracy of classification. Receiving the animal images output by the YOLOv8 model, the multi-layer convolution and pooling operations of the ResNet model are used to extract image features, which are then fed into the fully connected layer (i.e., Figure 6 the classification layer in Figure 7 ) for final species identification. The entire process includes the construction, training, tuning, and verification of the ResNet model to ensure that the ResNet model can efficiently and accurately identify the specific species of difficult samples. To improve the accuracy of the ResNet model in real scenarios, this embodiment trains samples under various conditions, including different lighting, angles, backgrounds, etc. Such training will make the model more robust and able to adapt to various environmental changes, ultimately achieving efficient classification and identification of endangered animal samples. The identification results are as

[0067] (III) Knowledge Base Module

[0068] To effectively manage and process data, in this embodiment, a highly secure and professional database (i.e., the knowledge base module) is established to centrally store image data, identification results, and detailed information related to endangered animals. This knowledge base module has functions of fast retrieval and efficient analysis, helping users quickly obtain the required information during the processing, thus improving efficiency.

[0069] As Figure 8As shown, the content in the knowledge base module is personally compiled by multiple experienced public security experts who have a profound understanding of endangered wild animals and rich practical experience in public security. These experts ensure that the information in the knowledge base module is not only accurate, but also highly practical and operable, and endangered species already in the library can be easily queried. The design and content selection of the knowledge base module closely revolve around the actual combat needs of public security, including but not limited to aspects such as the names, scientific names, taxonomic status, protection levels, morphological characteristics, common trading categories, common illegal utilization forms, values, distribution areas, etc. of endangered species, aiming to provide the most direct and effective support for police officers.

[0070] The knowledge base module of this embodiment has the ability to be updated in real time. Through continuous update and maintenance, it can ensure that the information in the knowledge base module always remains up-to-date to adapt to the changing needs of wild animal protection. The design of the knowledge base module is user-friendly, and police officers can easily access the required information through the front-end interface. Whether in the office or at the scene, they can quickly obtain key knowledge to support decision-making and actions.

[0071] Considering the particularity of public security work, the knowledge base module adopts high-standard security measures, which are identified by police officers' dedicated equipment to ensure the security and confidentiality of all information, preventing unauthorized access and data leakage.

[0072] At this time, in this embodiment, the human-computer interaction module is also used to update and query the content stored in the knowledge base module based on user input.

[0073] The knowledge base module is located in the internal network, specifically in the public security internal network, to ensure the security and confidentiality of all information.

[0074] (IV) Back-end management module

[0075] The back-end management module is a comprehensive platform designed specifically for the identification and protection of endangered species, aiming to effectively manage and analyze data. One of the core functions of the back-end management module is to support users in uploading new data sets in various formats. Users can conveniently upload images and related information to the platform, and the platform will automatically process these data to ensure their unified format for subsequent update of model weights.

[0076] In terms of model update, the back-end management module adopts fine-tuning technology, enabling the existing image recognition model to be retrained based on the newly uploaded data set. Through fine-tuning, the image recognition model can quickly adapt to the new species characteristics, improving the accuracy and flexibility of recognition. This method not only saves training time but also maximally utilizes the knowledge of the existing model.

[0077] During the class-incremental learning process, if the model is directly fine-tuned, the model will suffer from catastrophic forgetting due to the lack of explicit constraints on historical knowledge. As Figure 9 shown, the image recognition model uses the CIL-QUD (Class-Incremental Learning-Queried Unlabeled Data) scheme for fine-tuning. Based on the model, CIL-QUD introduces two mechanisms: the unlabeled query data (abbreviated as QUD) and the auxiliary classifier balanced training. The unlabeled query data mechanism enables the model to extract visually similar replay samples from memory during fine-tuning by querying anchors (a small number of stored instances for each category). The auxiliary classifier balanced training mechanism solves the problem of unbalanced training samples for replay. Its specific role is to balance the training samples of new tasks and old tasks, so that each batch is evenly distributed. During the training phase, the main classifier and the auxiliary classifier work together. Through the regularization term of knowledge distillation, the prediction results of the model on the old task query data are made as close as possible to the old model. By minimizing the expected loss of dataset B CB and B RS and adjusting the regularization coefficient λ, the model has better class-incremental performance.

[0078] The image recognition model of this embodiment can not only excellently complete the recognition and classification work, but also has the ability of real-time update and lightweight advantages, and can adapt to the changing needs of wildlife protection. Through fine-tuning, based on the existing pre-trained model, the model can use the new dataset uploaded by users for retraining, so that the system can quickly adapt to specific application scenarios and newly emerging species characteristics. This method significantly improves the recognition accuracy and response speed. Especially in the case of the constantly changing ecological environment and species distribution, it ensures that police officers obtain the latest information and support.

[0079] In summary, the background management system not only improves the efficiency of endangered species recognition, but also provides solid data support for the work of endangered species recognition and protection, helping to carry out public security work.

[0080] At this time, the endangered animal recognition feedback system based on image recognition of this embodiment further includes: a background management module, and the background management module is respectively communicatively connected to the human-computer interaction module, the image recognition module, and the knowledge base module.

[0081] The human-computer interaction module is also used to receive a new dataset input by the user. The dataset includes multiple samples and the label corresponding to each sample. The sample is an image of an endangered animal, and the label is the species of the endangered animal.

[0082] The background management module is used to retrieve historical data sets from the knowledge base module (the historical data sets are the data sets that have been used to train the image recognition model), and use the new data set and the historical data set to fine-tune the object detection model and the species classification model in the image recognition module.

[0083] The knowledge base module is used to store the new data set. At this time, the new data set becomes the historical data set.

[0084] Among them, using the new data set and the historical data set to fine-tune the object detection model and the species classification model in the image recognition module specifically includes: preprocessing the samples to obtain preprocessed samples. The preprocessing includes resolution enhancement, angle adjustment, brightness adjustment, background blurring, denoising, and normalization; using the preprocessed samples and labels as inputs, and using the method of incremental learning (specifically class incremental learning) to fine-tune the object detection model and the species classification model in the image recognition module.

[0085] The endangered animal recognition feedback system integrated by multiple modules in this embodiment can provide strong support for the police and help solve the recognition problems in endangered species crimes. This system integrates advanced technical means, can accurately identify the species attributes of the animals involved in the case, and evaluate the scientific value of the physical evidence, providing a reliable basis for case filing, investigation, and sentencing. This system not only improves the accuracy and efficiency of law enforcement, but also provides all-round technical guarantee for cracking down on endangered species crimes.

[0086] This embodiment aims to construct an intelligent recognition system for difficult-to-detect materials of endangered animals based on deep learning technology to effectively crack down on crimes involving endangered animals. It adopts a dual technical route of using the YOLOv8 model for object detection and the ResNet model for species classification to ensure the accurate identification and classification of endangered animals in complex backgrounds. In the object detection stage, using the YOLOv8 model, an efficient animal detection model is successfully constructed. Through training with a large number of animal images, it can quickly identify and locate the target animals, avoiding other irrelevant parts in the interfering images, and providing high-quality input data for subsequent species classification. In the species classification stage, the ResNet model is used for classification recognition. By extracting multi-layer features from the detected animal images, the accurate identification of difficult-to-detect materials is realized. The work in this stage not only improves the recognition accuracy of the model, but also enhances its adaptability and accuracy in real scenarios. Through advanced deep learning technology, strong technical support is provided for the recognition and classification of endangered animals. The construction of this system will help improve the technical system for cracking down on crimes involving endangered animals and provide a scientific basis for public security and judicial organs in case investigation, determination, and quantification.

[0087] Embodiment 2

[0088] This embodiment provides a method for identifying and feedback of endangered animals based on image recognition, as follows Figure 10 shown, and is applied to the system for identifying and feedback of endangered animals based on image recognition described in Embodiment 1. The method for identifying and feedback of endangered animals based on image recognition includes:

[0089] S1: Obtain a to-be-detected image of the to-be-detected endangered animal input by the user.

[0090] S2: Perform image recognition on the to-be-detected image to determine the species of the to-be-detected endangered animal.

[0091] S3: Retrieve the detailed information of the species of the to-be-detected endangered animal from the knowledge base module based on the species of the to-be-detected endangered animal, and display the species of the to-be-detected endangered animal and the detailed information of the species of the to-be-detected endangered animal.

[0092] Embodiment 3

[0093] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 11 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for identifying and feedback of endangered animals based on image recognition.

[0094] Those skilled in the art can understand that Figure 11 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0095] In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the endangered animal recognition feedback method based on image recognition in Embodiment 2 is implemented.

[0096] Embodiment 4

[0097] In an exemplary embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the endangered animal recognition feedback method based on image recognition in Embodiment 2 is implemented.

[0098] Embodiment 5

[0099] In an exemplary embodiment, a computer program product is provided, which includes a computer program. When the computer program is executed by a processor, the endangered animal recognition feedback method based on image recognition in Embodiment 2 is implemented.

[0100] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0101] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0102] Specific examples are used in this article to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, based on the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. An endangered animal identification and feedback system based on image recognition, characterized in that: The endangered animal identification feedback system based on image recognition includes: a human-computer interaction module, an image recognition module and a knowledge base module, wherein the human-computer interaction module and the image recognition module are both in communication connection with the knowledge base module; the knowledge base module stores detailed information of each of a plurality of species, wherein the detailed information includes pictures, names, protection levels, morphological characteristics, common transaction categories, common forms of illegal use, values ​​and distribution areas; The human-computer interaction module is used to receive an image of an endangered animal to be detected input by a user; The image recognition module is in communication with the human-computer interaction module; the image recognition module is used to perform image recognition on the image to be detected to determine the species of the endangered animal to be detected; The human-computer interaction module is also used to retrieve detailed information of the species of the endangered animal to be detected from the knowledge base module based on the species of the endangered animal to be detected, and display the species of the endangered animal to be detected and detailed information of the species of the endangered animal to be detected.

2. The endangered animal identification and feedback system based on image recognition according to claim 1, characterized in that: The image recognition module includes a target detection model and a species classification model connected in sequence; The target detection model is used to identify the endangered animal to be detected in the image to be detected, and to crop the area where the endangered animal to be detected is located to obtain an animal recognition image, wherein the animal recognition image is a partial image of the area where the endangered animal to be detected is located in the image to be detected; The species classification model is used to classify the animal identification image and determine the species of the endangered animal to be detected.

3. The endangered animal identification and feedback system based on image recognition according to claim 2, characterized in that: The target detection model adopts the YOLOv8 model, and the species classification model adopts the ResNet model.

4. The endangered animal identification and feedback system based on image recognition according to claim 1, characterized in that: The human-computer interaction module is also used to update and query the content stored in the knowledge base module based on user input.

5. The endangered animal identification and feedback system based on image recognition according to claim 1, characterized in that: The knowledge base module is located in the intranet.

6. The endangered animal identification and feedback system based on image recognition according to claim 2, characterized in that: The endangered animal identification and feedback system based on image recognition further includes: a background management module, the background management module is respectively connected to the human-computer interaction module, the image recognition module and the knowledge base module in communication; The human-computer interaction module is also used to receive a new data set input by a user; the data set includes a plurality of samples and a label corresponding to each of the samples, the samples are images of endangered animals, and the labels are species of endangered animals; The background management module is used to retrieve historical data sets from the knowledge base module, and use the new data sets and historical data sets to fine-tune the target detection model and species classification model in the image recognition module; The knowledge base module is used to store new data sets.

7. The endangered animal identification and feedback system based on image recognition according to claim 6, characterized in that: The object detection model and species classification model in the image recognition module are fine-tuned using new and historical datasets, including: Preprocessing the sample to obtain a preprocessed sample; the preprocessing includes resolution enhancement, angle adjustment, brightness adjustment, background blur, denoising and standardization; Taking the preprocessed samples and the labels as input, the target detection model and the species classification model in the image recognition module are fine-tuned using an incremental learning method.

8. An endangered animal identification and feedback method based on image recognition, applied to the endangered animal identification and feedback system based on image recognition as claimed in any one of claims 1 to 7, characterized in that: The endangered animal identification feedback method based on image recognition includes: Obtaining an image of an endangered animal to be detected input by a user; Performing image recognition on the image to be detected to determine the species of the endangered animal to be detected; Based on the species of the endangered animal to be detected, detailed information of the species of the endangered animal to be detected is retrieved from a knowledge base module, and the species of the endangered animal to be detected and detailed information of the species of the endangered animal to be detected are displayed.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the endangered animal identification and feedback method based on image recognition as described in claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the endangered animal identification and feedback method based on image recognition described in claim 8 is implemented.

Images