Animal and plant identification system based on AI image recognition

Abstract

The application discloses an animal and plant identification system based on AI image recognition and belongs to the technical field of animal and plant identification. The system collects images of animals and plants through a sensing end, obtains animal and plant collection data, and sends the animal and plant collection data to a cloud platform. The cloud platform analyzes and identifies animals and plants according to user requirements, presets an animal and plant identification mode, obtains animal and plant collection data sent by the sensing end, analyzes the animal and plant collection data through the animal and plant identification mode, obtains basic information or deep identification data of animals and plants, associates and matches each deep identification data with a deep correlation data set of each animal and plant stored in a database, obtains an association matching result, stores and manages data according to the association matching result, identifies a difference correlation analysis set stored in the database in real time, and generates a differential analysis report according to feature differences of deep identification data of each different time point in the difference correlation analysis set.

Description

Technical Field

[0001] This invention belongs to the field of plant and animal identification technology, specifically a plant and animal identification system based on AI image recognition. Background Technology

[0002] With the rapid development of artificial intelligence image recognition technology, image recognition-based plant and animal identification systems are gradually being applied in various fields, replacing traditional manual identification methods to a certain extent and improving identification efficiency.

[0003] For example, Chinese patent application No. 201510270381.X discloses a method and system for identifying plants and animals on a mobile terminal. The method involves acquiring photos of plants and animals to be identified taken on the mobile terminal, scanning the features of the photos, and comparing the scanned features with plant and animal photo samples in a plant and animal database. If a match is found, the matching plant and animal photo samples are retrieved from the database and displayed. This allows for intelligent identification of various plants and animals on the mobile terminal.

[0004] The aforementioned patent application identifies plants and animals by analyzing acquired images; however, existing plant and animal identification systems, such as those mentioned above and others based on image recognition, generally suffer from a core deficiency: superficial identification levels and a lack of coherent correlation analysis, making it difficult to meet the deeper needs of users in practical applications. The core function of existing systems is limited to superficial identification of a single input image, that is, it can only identify basic species information of plants and animals in the image, and then simply generate general related information such as family, genus, and basic habits, but it cannot perform deeper identification of plants and animals. For example, it cannot accurately identify deeper features such as the growth stage, health status, and morphological changes of plants and animals, making it difficult to meet the identification needs of deep plant and animal information in scientific research, agricultural production, and other scenarios.

[0005] Meanwhile, existing systems do not fully consider the need for correlation and identification across multiple time intervals in practical applications. In real-world scenarios, users often need to take multiple photos of the same plant or animal at different intervals for identification. Examples include taking photos of crops at different growth stages in agricultural settings, capturing the activity status of wild animals at different times in ecological conservation settings, and tracking plant growth processes in scientific research settings. However, existing systems can only identify the species of each image individually and output the species information of the plant or animal from multiple photos. They cannot correlate and match multiple identification records, cannot determine whether multiple photos correspond to the same plant or animal, and therefore cannot compare and identify consecutive images of the same plant or animal. They also cannot perform differentiated analysis of images at different time points, making it difficult to meet users' deeper needs for tracking the growth process and state changes of plants and animals. For example, they cannot analyze the growth process of plant leaves, flowering, and fruiting through multiple photos taken at different intervals, nor can they track changes in animal body size and behavior. This limits the application scenarios of the system and significantly reduces its practicality.

[0006] In order to solve the above problems, this invention provides an animal and plant identification system based on AI image recognition. Summary of the Invention

[0007] To address the problems of the above solutions, this invention provides an animal and plant identification system based on AI image recognition.

[0008] The objective of this invention can be achieved through the following technical solutions: An AI-based image recognition-based plant and animal identification system, including a cloud platform and a sensing terminal; Furthermore, the cloud platform is communicatively connected to each sensing terminal.

[0009] The sensing end is used to collect images of the plants and animals that need to be identified, obtain plant and animal collection data, preprocess the plant and animal collection data, and send the preprocessed plant and animal collection data to the cloud platform.

[0010] The cloud platform includes an authentication module, an authentication analysis module, a database, and a correlation analysis module; The identification module is used to identify and analyze plants and animals according to user needs, and preset plant and animal identification methods, which include basic identification methods and in-depth identification methods. The application determines the method for identifying plants and animals based on user needs, and obtains the plant and animal collection data sent by the sensing terminal; When the application uses the basic identification method for plant and animal identification, it identifies the collected plant and animal data based on the basic identification method, obtains basic plant and animal information, and displays the basic plant and animal information to the user. When the plant and animal identification method used is the deep identification method, the plant and animal data collected is subjected to deep identification based on the deep identification method to obtain the deep identification data of each plant and animal that needs to be identified, and the deep identification data is sent to the identification analysis module.

[0011] The identification analysis module is used to perform in-depth identification analysis, identify the in-depth identification data of each plant and animal, and match each in-depth identification data with the in-depth association data set of each plant and animal stored in the database to obtain the association matching result. When the association matching result is a successful match, the deep association data is added to the corresponding deep association data set in the database; When the association matching result is a failure, a new deep association data set is generated based on the deep association data, and the deep association data set is stored in the database.

[0012] Furthermore, the user needs are calibrated, including: Identify the various deep identification data sets stored in the database, obtain the collection time corresponding to each deep identification data set in the deep identification data set, generate the corresponding identification timeline of plants and animals based on each collection time, display the identification timeline to the user, and calibrate according to the user's needs based on the identification timeline.

[0013] Furthermore, each deep identification data set stored in the database is labeled with a screening tag for the corresponding animal and plant. When matching each deep identification data set with the deep association data set of each animal and plant stored in the database, an initial screening is performed using each deep identification data set and the corresponding screening tags for the animal and plant. The deep identification data set of the animal and plant that does not include all the screening tags is considered a failed match.

[0014] Furthermore, the cloud platform includes an identification and replacement module, but does not include an identification module or an identification analysis module; The identification and replacement module is used to identify and analyze plants and animals. It identifies the collected data of plants and animals through a preset deep identification method to obtain deep identification data of each plant and animal. The deep identification data is matched with the deep association data set of plants and animals stored in the database and the basic information of plants and animals; When a match is successful, the deep association data is added to the corresponding deep association data set in the database; When a match fails, basic information about plants and animals is extracted from the deep association data, displayed to the user, and stored in the database.

[0015] Furthermore, the cloud platform includes an identification and replacement module, but does not include an identification module or an identification analysis module; The identification and replacement module is used for identification and analysis of plants and animals, and presets basic identification methods and deep identification methods; The basic identification method is used to analyze the collected data of animals and plants to obtain basic information about them. The basic information about animals and plants is then correlated with the deep correlation data set stored in the database and the basic information about animals and plants to make judgments. When a correlation probability is determined, deep identification data is obtained through deep identification. The deep identification data is then matched with the deep correlation data set and basic information of plants and animals stored in the database. If the match is successful, the deep correlation data is added to the corresponding deep correlation data set in the database. If the match fails, the basic information of plants and animals is displayed to the user and stored in the database. When it is determined that there is no correlation probability, the basic information of the plants and animals will be displayed to the user and stored in the database.

[0016] The database is used for data storage.

[0017] The correlation analysis module is used to perform correlation analysis, identify the update records of each deep identification data set stored in the database in real time, mark the deep identification data set with updates as the difference correlation analysis set; generate a difference analysis report based on the feature differences of the deep identification data at different time points in the difference management analysis set, and send the difference analysis report to the user.

[0018] Furthermore, based on the characteristic differences of the in-depth identification data at different time points in the difference management analysis set, a difference analysis report is generated, including: By comparing the set of difference management analyses with the reference analysis report, dynamic difference characteristics can be obtained; Modular analysis units pre-defined by the platform for users to perform correlation analysis on various plants and animals; The system obtains users' association analysis requirements for plants and animals, matches corresponding modular analysis units according to these requirements, and analyzes the reference analysis report and dynamic difference characteristics through the modular analysis units to obtain the association analysis results for the association analysis requirements. The results of each correlation analysis are summarized to generate a differential analysis report.

[0019] Furthermore, the modular analysis unit includes a plant and animal growth analysis unit, a plant and animal disease identification unit, a plant and animal disease development assessment unit, and a plant and animal maintenance analysis unit.

[0020] Compared with the prior art, the beneficial effects of the present invention are: This system overcomes the limitations of existing shallow identification methods, combining shallow and deep identification of plants and animals. It not only rapidly identifies plant and animal species but also accurately identifies their growth stages, health status, morphological details, and other deep characteristics, outputting more comprehensive identification information to meet the advanced needs of scientific research, agriculture, and other scenarios. It also addresses the lack of coherent and correlated identification in existing systems by using a correlation analysis layer to match identification records across multiple time intervals. This allows for precise determination of whether multiple images correspond to the same plant or animal, enabling continuous tracking of the same species and filling the gap in existing technologies that cannot achieve coherent identification and growth process tracking. Furthermore, it can perform differential analysis on images of the same plant or animal across multiple time intervals, clearly presenting growth and status changes, providing users with intuitive change trajectories and analysis reports. This eliminates the need for manual comparison, significantly improving work efficiency, and is particularly suitable for scenarios such as plant growth process tracking and wildlife status monitoring. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a block diagram illustrating the principle of the present invention; Figure 2 This is a schematic diagram of a cloud platform according to another embodiment. Detailed Implementation

[0023] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0024] like Figures 1 to 2 As shown, the plant and animal identification system based on AI image recognition includes a cloud platform and a sensing terminal; The cloud platform communicates with each sensing terminal.

[0025] The sensing end is used to collect images of the plants and animals that need to be identified, obtain corresponding plant and animal data, preprocess the collected plant and animal data, and send the preprocessed plant and animal data to the cloud platform.

[0026] In one embodiment, the sensing device can be installed in a mobile device such as a mobile phone or tablet, and data can be collected based on the corresponding hardware and software conditions of the mobile device; data can also be collected through multiple devices such as field cameras, drones, infrared cameras, and underwater cameras.

[0027] In one embodiment, the collected animal and plant data is preprocessed using existing image processing methods, such as enhancing, cropping, normalizing, and amplifying the collected images to improve image quality, provide support for subsequent identification, and reduce the amount of data transmitted, thereby improving identification efficiency.

[0028] The cloud platform includes an authentication module, an authentication analysis module, a database, and a correlation analysis module; The identification module is used to identify and analyze plants and animals according to user needs, and preset plant and animal identification methods, including basic identification methods and in-depth identification methods. Basic identification methods refer to identifying basic identification information such as species name, family, and uses of animals and plants according to existing animal and plant identification needs. For example, a fusion architecture of the YOLO-World object detection model and the EfficientNet image classification model is used to quickly identify the species of animals and plants in the image and output basic identification results such as species name and family.

[0029] Deep identification methods are used to further extract deeper features of plants and animals based on basic identification, identifying their growth stages, health status, and morphological details (such as leaf size, color, animal body shape, and fur color). For plants and other plants with fixed locations, features of adjacent plants and animals can also be included to facilitate collaborative verification analysis. The deep identification results are output. For example, the MiniCPM-V multimodal model and feature refinement extraction model are integrated, and a transfer learning strategy is adopted. Based on the ImageNet pre-trained model, it is fine-tuned with a dataset of plants and animals labeled with deep features. Multi-scale fusion and adaptive anchor box optimization techniques are introduced to improve identification accuracy. At the same time, iterative optimization is carried out through indicators such as accuracy and recall.

[0030] Acquire plant and animal data sent by the sensing terminal; The application's plant and animal identification methods are determined based on user needs, meaning that users select the appropriate plant and animal identification methods according to their needs, thus forming corresponding user requirements. When the application uses the basic identification method for plant and animal identification, it identifies the collected plant and animal data based on the basic identification method, obtains the corresponding basic plant and animal information, and displays the basic plant and animal information to the user. When the plant and animal identification method used is the deep identification method, the plant and animal data collected is subjected to deep identification based on the deep identification method to obtain the deep identification data of each plant and animal that needs to be identified, and the deep identification data is sent to the identification analysis module.

[0031] The plants and animals to be identified in depth are determined according to user needs. For example, after the images are collected, the user marks the plants and animals that need to be identified in depth. For multi-target plant and animal identification, it can be pre-set through text, images, etc. For example, to identify plants in a certain farmland area, the depth identification requirement is formed by the area, that is, all plants in the farmland area, or it can be further limited to corn, wheat, etc.

[0032] In one embodiment, a deeper identification method can also be set up based on other methods to achieve deeper identification of plant and animal data, such as identifying more plant and animal features based on other existing AI image recognition technologies.

[0033] The identification analysis module is used to perform in-depth identification analysis, identify the in-depth identification data of each plant and animal, and match each in-depth identification data with the in-depth association data set of each plant and animal stored in the database to obtain the association matching result. When the association matching result is a successful match, the deep association data is added to the corresponding deep association data set in the database.

[0034] When the association matching result is a failure, a new deep association data set is generated based on the deep association data. At this time, the deep association data set contains only one deep association data; the deep association data set is then stored in the database.

[0035] In one embodiment, the deep identification data is matched with the deep association data sets of various plants and animals stored in the database, including: The system extracts feature information (shallow + deep), shooting location, shooting time, etc. from the current identification image and compares and matches it with the deep association data set in the database. For scenes with multiple plants and animals (such as multiple wheat plants in a wheat field or groups of animals), a triple matching strategy of "individual feature anchoring + group contour association + spatiotemporal constraints" is adopted, taking into account the actual situation of shooting angle and position changes. This strategy determines the matching relationship between each plant and animal in the current identification image and the corresponding individual in the deep association data set, realizing the coherent association of multiple time intervals and multiple plants and animals. The association matching results are obtained, including successful matching and failed matching.

[0036] Among them, individual feature anchoring refers to extracting the unique deep features of each animal and plant (such as the leaf texture and tillering number of wheat, and the fur color texture and body shape outline details of animals) as anchor features. These features are not affected by slight changes in shooting angle and position, ensuring the uniqueness of individual identification. Group contour association refers to extracting the overall contour distribution features of the group for group animal and plant scenes, and combining them with the relative positional relationship of individuals in the group to assist in matching (such as the row and column distribution of wheat in a wheat field, and the relative spacing of individuals in a group of animals), making up for the shortcomings of single feature matching in scenarios where the group position changes. Spatiotemporal constraints refer to combining the shooting time interval (consistent with the growth / activity patterns of animals and plants) and the consistency of shooting location (same area) to filter unreasonable matching results, further improving the accuracy of multi-animal and plant association matching.

[0037] In one embodiment, each deep identification data is matched with a set of deep association data of various plants and animals stored in the database. Alternatively, an intelligent matching model can be established based on machine learning, deep learning algorithms, etc., to perform association matching.

[0038] In one embodiment, the deep identification data is matched with the deep association data sets of various plants and animals stored in the database, including: Based on the various deep identification datasets stored in the database, corresponding filtering tags are assigned to the relevant plants and animals. This involves extracting unique feature tags from the deep identification datasets to improve subsequent association matching efficiency. Examples of tags include plant / animal species tags, regional location tags, and growth stage tags (based on the time of evaluation). For instance, if deep identification data corresponds to wheat, region A, and growth stage B, these features can significantly reduce the need for association matching analysis with other unrelated deep identification datasets, thus improving matching efficiency. Furthermore, tags such as animal coat color classification can be added for further tag segmentation. Specifically, based on the platform's requirements and a large amount of historical plant and animal data, various filtering tag types are preset. Subsequent updates and feature recognition are performed on the stored deep identification data to obtain the corresponding filtering tags. Based on the deep identification data, the filter tags corresponding to each animal and plant in the database are initially matched. Based on the initial matching results (whether the deep identification data has all the filter tags), the deep identification data set stored in the database is initially filtered, that is, the deep identification data set that failed the initial matching is removed. The deep identification data is matched with the corresponding set of remaining deep identification data to obtain the corresponding matching results. For example, the similarity between plants and animals is greater than the preset value, and the characteristics of adjacent plants and animals also meet the matching requirements.

[0039] In one embodiment, the matching results of multiple plants and animals can be optimized. For example, the Hungarian algorithm can be used to solve the one-to-one matching problem of multiple targets, avoiding the situation where a historical individual matches multiple current individuals or the current individual cannot match the corresponding historical individual. At the same time, feature alignment algorithms can be used to correct feature deviations caused by shooting angle shifts and position movements, thereby improving matching accuracy and adapting to the association recognition needs of multi-target scenes such as wheat fields with multiple wheat plants and groups of animals.

[0040] In one embodiment, during the above-mentioned association matching process, changes are inferred from the interval time and plant and animal information, and then matching is performed. At the same time, the matching results can be calibrated based on the inferred changes.

[0041] In one embodiment, relying on the user to select a plant and animal identification method can easily lead to errors, such as forgetting to adjust the previous identification method, resulting in the application of the wrong method. At the same time, user-initiated adjustments also offer room for intelligent optimization. Based on this, this embodiment can be analyzed in the following way: Replace the identification module and the identification analysis module with the identification replacement module; Identification and Replacement Module: Only preset deep identification methods are used to identify the collected animal and plant data and obtain deep identification data for each animal and plant; the deep identification data is then matched with the deep association data set of animals and plants and the basic information of animals and plants stored in the database. When a match is successful, it is determined to be a deep identification requirement, and further analysis is performed. When a match fails, it is considered a basic identification requirement. Feature extraction is performed on the deep identification data according to the data range corresponding to the basic information of plants and animals to obtain the basic information of plants and animals.

[0042] In this embodiment, the database stores a set of deeply related data and basic information on plants and animals, and the basic information on plants and animals is regarded as the first collection and identification of possible associations.

[0043] This embodiment enables users to intelligently identify plants and animals and perform automatic correlation analysis, avoiding identification problems caused by relying on user needs.

[0044] In one embodiment, the basic identification method and the deep identification method can be preset simultaneously according to the above embodiments; Replace the identification module and the identification analysis module with the identification replacement module; Identification and Replacement Module: Analyzes the collected animal and plant data through basic identification methods to obtain basic information about the animals and plants. It then performs association judgments by comparing the basic information about the animals and plants with the deep association data set stored in the database and the basic information about the animals and plants. For example, if the similarity is greater than a preset value, the requirement can be lower than the association matching requirement to reduce the amount of data analysis. When a correlation probability is determined, analysis is performed based on a deep identification method to obtain deep identification data. The deep identification method is then matched with the deep correlation data set stored in the database and the basic information of plants and animals. When a match is successful, it is determined as a deep identification requirement and further analysis is performed. When a match fails, it is considered a basic identification requirement and the basic information of plants and animals is displayed.

[0045] When it is determined that there is no correlation probability, it is considered a basic identification requirement, and basic information about the plants and animals is displayed.

[0046] This embodiment also enables users to intelligently identify plants and animals and perform automatic correlation analysis, avoiding identification problems caused by relying on user needs. However, by performing pre-analysis based on basic identification methods, it is easier to improve the efficiency of the entire identification process.

[0047] In one embodiment, when a user selects a method for identifying plants and animals, errors can easily occur, such as forgetting to adjust the previous identification method, leading to the application of the wrong method. Verification can be performed using the following methods: It identifies each deep identification data set stored in the database, obtains the collection time corresponding to each deep identification data set, generates the corresponding identification timeline of plants and animals based on each collection time, and displays the identification timeline to the user; it is used to calibrate according to user needs based on the identification timeline, and when the mode selection is incorrect, it can be recalled through the history and re-identified and analyzed.

[0048] The database is used for data storage.

[0049] The correlation analysis module is used to perform correlation analysis, identify the update records of each deep identification data set stored in the database in real time, and mark the deep identification data set with updates as the differential correlation analysis set; generate a differential analysis report based on the characteristic differences of the deep identification data at different time points in the differential analysis set, clarify the growth changes and state changes of plants and animals (such as changes in plant leaf growth size, flowering and fruiting process, changes in animal body size, etc.); and send the differential analysis report to the user.

[0050] Specifically, intelligent generation is based on existing report generation technology. By identifying corresponding feature differences through in-depth identification data at different time points, in-depth analysis of plant and animal growth and diseases is conducted based on feature differences and time duration. For example, the analysis can be based on existing analytical functions to assess whether the wheat growth progress is up to standard, whether there are pests and diseases, and the effectiveness of pest and disease control.

[0051] In one embodiment, a differential analysis report is generated based on the characteristic differences of deep identification data at different time points in the differential management analysis set, including: Mark the difference analysis report corresponding to the last update of the difference management analysis set as the reference analysis report; By comparing the set of difference management analyses with the reference analysis report, dynamic difference characteristics can be obtained; The platform pre-sets modular analysis units for users to perform correlation analysis on various plants and animals. For example, to evaluate whether the growth progress of a certain plant is qualified over a period of time, it can be set up based on the isolated forest algorithm, machine learning, etc. Specifically, the platform sets up each modular analysis unit based on the existing technologies for analyzing various analysis needs. One modular analysis unit can correspond to multiple analysis needs. Obtain user requirements for association analysis of plants and animals, match corresponding modular analysis units according to the requirements, and analyze the reference analysis report and dynamic difference characteristics through the matched modular analysis units to obtain the association analysis results for the corresponding requirements. A differential analysis report is generated by summarizing the results of each correlation analysis.

[0052] In one embodiment, the modular analysis unit includes a plant and animal growth analysis unit, a plant and animal disease identification unit, a plant and animal disease development assessment unit, and a plant and animal maintenance analysis unit.

[0053] Based on the correlation analysis requirements corresponding to each modular analysis unit, staff can configure it on the cloud platform, or they can call other systems with corresponding analysis functions to perform the analysis.

[0054] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the AI ​​image recognition-based plant and animal identification system as described in the above embodiments.

[0055] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0056] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An AI-based image recognition system for identifying plants and animals, characterized in that, This includes cloud platforms and sensing terminals; The sensing end is used to collect images of the plants and animals that need to be identified, obtain plant and animal collection data, preprocess the plant and animal collection data, and send the preprocessed plant and animal collection data to the cloud platform. The cloud platform includes an authentication module, an authentication analysis module, a database, and a correlation analysis module; The identification module is used to identify and analyze plants and animals according to user needs, and preset plant and animal identification methods, including basic identification methods and in-depth identification methods. Acquire plant and animal data sent by the sensing terminal; The data collected from plants and animals is analyzed using plant and animal identification methods to obtain basic information or in-depth identification data; the basic information is displayed to the user; and the in-depth identification data is sent to the identification and analysis module. The identification analysis module is used to perform in-depth identification analysis, identify the in-depth identification data of each plant and animal, and match each in-depth identification data with the in-depth association data set of each plant and animal stored in the database to obtain the association matching result. When the association matching result is a successful match, the deep association data is added to the corresponding deep association data set in the database; When the association matching result is a failure, a new deep association data set is generated based on the deep association data, and the deep association data set is stored in the database; The database is used for data storage; The correlation analysis module is used to perform correlation analysis, identify the update records of each deep identification data set stored in the database in real time, mark the deep identification data set with updates as the differential correlation analysis set, generate a differential analysis report based on the feature differences of the deep identification data at different time points in the differential correlation analysis set, and send the differential analysis report to the user.

2. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, The cloud platform communicates with each sensing terminal.

3. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, Analyzing the collected data on plants and animals through identification methods, including: The method for identifying plants and animals in the application is determined based on user needs; When the applied plant and animal identification method is the basic identification method, the collected plant and animal data are identified based on the basic identification method to obtain basic plant and animal information; When the plant and animal identification method used is the deep identification method, the plant and animal data collected are subjected to deep identification based on the deep identification method to obtain deep identification data for each plant and animal that needs to be identified.

4. The plant and animal identification system based on AI image recognition according to claim 3, characterized in that, Calibrate based on user needs, including: Identify the various deep identification data sets stored in the database, obtain the collection time corresponding to each deep identification data set in the deep identification data set, generate the corresponding identification timeline of plants and animals based on each collection time, display the identification timeline to the user, and calibrate according to the user's needs based on the identification timeline.

5. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, Each deep identification dataset stored in the database is labeled with a screening tag for the corresponding plant and animal. When matching each deep identification dataset with the deep association dataset of each plant and animal stored in the database, an initial screening is performed using each deep identification dataset and the corresponding screening tags for the plant and animal. Deep identification datasets of plants and animals that do not include all screening tags are considered as failed matches.

6. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, The cloud platform includes an identification and replacement module, but does not include an identification module or an identification analysis module; The identification and replacement module is used to identify and analyze plants and animals. It identifies the collected data of plants and animals through a preset deep identification method to obtain deep identification data of each plant and animal. The deep identification data is matched with the deep association data set of plants and animals stored in the database and the basic information of plants and animals; When a match is successful, the deep association data is added to the corresponding deep association data set in the database; When a match fails, basic information about plants and animals is extracted from the deep association data and displayed to the user. Basic information about plants and animals is stored in a database.

7. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, The cloud platform includes an identification and replacement module, but does not include an identification module or an identification analysis module; The identification and replacement module is used for identification and analysis of plants and animals, and presets basic identification methods and deep identification methods; By analyzing the collected data of plants and animals through basic identification methods, basic information about plants and animals can be obtained. The association between basic information on plants and animals and the deep-linked data set stored in the database and the basic information on plants and animals is determined. When a correlation probability is determined, deep identification data is obtained through deep identification. The deep identification data is then matched with the deep correlation data set and basic information of plants and animals stored in the database. When a match is successful, the deep correlation data is added to the corresponding deep correlation data set in the database. When a match fails, basic information about the plants and animals is displayed to the user and stored in the database. When it is determined that there is no correlation probability, the basic information of the plants and animals will be displayed to the user and stored in the database.

8. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, Based on the characteristic differences of the in-depth identification data at different time points in the difference management analysis set, a difference analysis report is generated, including: By comparing the set of difference management analyses with the reference analysis report, dynamic difference characteristics can be obtained; Modular analysis units pre-defined by the platform for users to perform correlation analysis on various plants and animals; The system obtains users' association analysis requirements for plants and animals, matches corresponding modular analysis units according to these requirements, and analyzes the reference analysis report and dynamic difference characteristics through the modular analysis units to obtain the association analysis results for the association analysis requirements. A differential analysis report is generated based on the results of each correlation analysis.

9. The plant and animal identification system based on AI image recognition according to claim 1, characterized in that, The modular analysis unit includes a plant and animal growth analysis unit, a plant and animal disease identification unit, a plant and animal disease development assessment unit, and a plant and animal maintenance analysis unit.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the plant and animal identification system based on AI image recognition as described in any one of claims 1 to 9.

Images