Frog identification method and system based on image processing

Abstract

The present invention provides a frog identification method and system based on image processing, including obtaining a color image of a frog to be identified and constructing a 3D color standard model; extracting features to obtain an identification feature group; binding the identification feature group to an identification code; obtaining a color image of the frog to be identified and constructing a 3D color entity model; comparing and analyzing the 3D color entity model with all 3D color standard models in turn; if the result of the comparison and analysis exceeds a preset matching threshold, the 3D color entity model is determined to be the same as the 3D color standard model, and the identification code of the 3D color standard model is assigned to the frog corresponding to the 3D color entity model. Since the surface patterns of frogs change little during their growth, and different frogs have different surface patterns, different frogs can be distinguished by analyzing the surface images of frogs, and the frogs can be identified without harming them, thereby solving the problem that the existing frog identification methods cause great harm to frogs.

Description

Technical Field

[0001] The present invention relates to the field of biological identification technology, and in particular to a frog identification method and system based on image processing. Background Art

[0002] Frogs are widely used in various biological experiments. When conducting frog experiments, batches of frogs are usually used as samples to obtain group laws. Specifically, frogs can be randomly divided into multiple groups, and different group experiments can be conducted to obtain the reactions of frogs after different experiments, and then the experimental results can be obtained. Therefore, in order to facilitate experimental records, frogs are usually marked and distinguished.

[0003] Commonly used frog identification methods include fluorescent pigment marking, injection chips, and toe clipping. However, existing identification methods can lead to deviations in research results in some studies and may also cause harm or death to frogs. Summary of the invention

[0004] The purpose of the present invention is to provide a frog identification method and system based on image processing, so as to at least solve the problem that the existing frog identification method causes great harm to frogs, and can be used in animal life history characteristic data collection experiments.

[0005] In order to solve the above technical problems, the present invention provides a frog identification method based on image processing, and the frog identification method comprises:

[0006] Obtain a color image of the frog to be marked, preferably the black-spotted sidefold frog, commonly known as the black and white frog, which has obvious lines and colors on its surface, and construct a 3D color standard model based on the obtained color image;

[0007] Extracting features from the 3D color standard model to obtain a set of identification features, wherein the parts from which the features are extracted include the head, the back, four fingers / toes, and four limbs;

[0008] Binding the identification feature group to an identification code, wherein the identification codes corresponding to different identification feature groups are different and unique;

[0009] Binding all the 3D color standard models with corresponding identification feature groups, and summarizing them to obtain a frog identification library;

[0010] Acquire a color image of the frog to be identified, and construct a 3D color entity model based on the acquired color image;

[0011] Comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in turn, wherein the basis for the comparative analysis is the identification feature group corresponding to the current 3D color standard model;

[0012] If the result of the comparative analysis exceeds the preset matching threshold, it is determined that the frog corresponding to the 3D color entity model and the frog corresponding to the 3D color standard model are the same frog, and the identification code corresponding to the 3D color standard model is assigned to the frog corresponding to the 3D color entity model.

[0013] Optionally, in the frog identification method based on image processing, the method of extracting features from the 3D color standard model to obtain an identification feature group includes:

[0014] Performing dot matrix pixel processing on the surface of the 3D color standard model;

[0015] Select a×b pixel blocks with the largest RGB color difference values ​​from the head, back, four fingers / toes and four limbs as the first identification features, where the sizes of pixel blocks in different areas are not exactly the same;

[0016] Select the pixel blocks of c×d pixels with the largest brightness difference from the head, back, four fingers / toes and four limbs as the second identification features, where the sizes of pixel blocks in different areas are not exactly the same;

[0017] All the first identification features are aggregated to obtain a first identification feature group, all the second identification features are aggregated to obtain a second identification feature group, and the first identification feature group and the second identification feature group are aggregated to obtain an identification feature group.

[0018] Optionally, in the frog identification method based on image processing, the method of comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in sequence includes:

[0019] Scaling the 3D color entity model until the size of the 3D color entity model is consistent with the size of the 3D color standard model currently being compared;

[0020] Set all pixel block coordinates in the identification feature group corresponding to the current 3D color standard model in the 3D color entity model to obtain a feature selection box on the 3D color entity model;

[0021] Compare and analyze the identification feature and the feature selection box of the same coordinates. If the degree of overlap is greater than the preset overlap threshold, the identification feature is determined to be the same as the feature selection box.

[0022] If the matching degree of all identification features corresponding to the current 3D color standard model and the feature selection box of the 3D color entity model is greater than a preset matching threshold, the 3D color entity model is determined to be the same as the current 3D color standard model, otherwise the next 3D color standard model is replaced and the above steps are repeated.

[0023] Optionally, in the frog identification method based on image processing, the method of comparing and analyzing the identification features and feature selection boxes of the same coordinates includes:

[0024] Compare and analyze the first identification feature and the feature selection box with the same coordinates. If the RGB pixel arrangement overlap is greater than 75%, the first identification feature is determined to be the same as the feature selection box.

[0025] The second identification feature and the feature selection box with the same coordinates are compared and analyzed. If the brightness value distribution overlap is greater than 70%, it is determined that the second identification feature is the same as the feature selection box.

[0026] Optionally, in the frog identification method based on image processing, the method for determining whether the 3D color entity model is the same as the current 3D color standard model includes:

[0027] If it is determined that the number of features that are identical to the first identification feature and the feature selection box accounts for more than 80% of the number of all first identification features, then it is determined that the 3D color solid model is identical to the current 3D color standard model;

[0028] If there are multiple 3D color standard models corresponding to a proportion greater than 80%, the 3D color standard model corresponding to the one with the largest proportion is selected to be the same as the 3D color entity model.

[0029] Optionally, in the frog identification method based on image processing, the method for determining whether the 3D color entity model is the same as the current 3D color standard model further comprises:

[0030] If the number of features determined to be identical to the feature selection box for the first identification feature accounts for more than 50% and less than 80% of the number of all first identification features, then the ratio of the number of features determined to be identical to the feature selection box for the second identification feature accounts for the number of all second identification features;

[0031] If the number of features that are determined to be the same as the feature selection box in the second identification features accounts for more than 85% of all the second identification features, then it is determined that the 3D color entity model is the same as the current 3D color standard model.

[0032] In order to solve the above technical problems, the present invention also provides a frog identification system based on image processing, which is used to perform the frog identification method based on image processing as described in any one of the above items, and the frog identification system includes an image acquisition module, a 3D modeling module, a feature processing module and an identification binding module; the image acquisition module is used to acquire the color image of the frog; the 3D modeling module is used to construct a 3D color standard model and a 3D color entity model according to the color image; the feature processing module is used to extract features from the 3D color standard model and perform feature comparison analysis on the 3D color entity model; the identification binding module is used to generate an identification code for the 3D color standard model, and match the identification code to the 3D color entity model according to the result of the feature comparison analysis.

[0033] Optionally, in the frog identification system based on image processing, the 3D modeling module includes an image analysis unit, a posture correction unit and a standard model generation unit; the image analysis unit is used to analyze and process the color image of the frog to distinguish the structural characteristics of different regions of the frog; the posture correction unit is used to reconstruct the body posture of the frog according to the acquired color image, so that all constructed 3D color standard models and 3D color solid models have the same standard posture; the standard model generation unit is used to construct the image features of different regions of the frog in the color image on the standard posture according to the analysis results of the image analysis unit to generate a 3D color standard model and a 3D color solid model.

[0034] Optionally, in the frog identification system based on image processing, the feature processing module includes a standard model feature processing unit and a physical model feature processing unit; the standard model feature processing unit is used to extract features from the 3D color standard model to obtain an identification feature group; the physical model feature processing unit is used to compare and analyze the 3D color physical model with all the 3D color standard models in turn according to the identification feature group to determine whether the 3D color physical model is the same as the 3D color standard model.

[0035] Optionally, in the frog identification system based on image processing, the identification binding module includes an identification generation unit and an identification matching unit; the identification generation unit is used to generate a unique identification code for each 3D color standard model; the identification matching unit is used to match the identification code to the 3D color entity model according to the result of feature comparison analysis.

[0036] The frog identification method and system based on image processing provided by the present invention include: obtaining a color image of a frog to be identified, and constructing a 3D color standard model according to the obtained color image; extracting features from the 3D color standard model to obtain an identification feature group; binding the identification feature group with an identification code; binding all the 3D color standard models with the corresponding identification feature group, and summarizing them to obtain a frog identification library; obtaining a color image of a frog to be identified, and constructing a 3D color entity model according to the obtained color image; comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in turn, wherein the basis for the comparison and analysis is the identification feature group corresponding to the current 3D color standard model; if the result of the comparison and analysis exceeds a preset matching threshold, it is determined that the frog corresponding to the 3D color entity model and the frog corresponding to the 3D color standard model are the same, and the identification code corresponding to the 3D color standard model is assigned to the frog corresponding to the 3D color entity model. Since the surface patterns of frogs change little during their growth, and different frogs have different surface patterns, different frogs can be distinguished by analyzing their surface images, and the frogs can be identified without harming them, thus solving the problem that existing frog identification methods cause great harm to frogs. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flowchart of a frog identification method based on image processing provided in this embodiment;

[0038] Figure 2 A schematic diagram of a part for feature extraction provided in this embodiment;

[0039] Figure 3 This is a schematic diagram of the structure of the frog identification system based on image processing provided in this embodiment. DETAILED DESCRIPTION

[0040] The frog identification method and system based on image processing proposed by the present invention are further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the accompanying drawings are all in a very simplified form and use non-precise proportions, which are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention. In addition, the structure shown in the accompanying drawings is often a part of the actual structure. In particular, the emphasis of each accompanying drawing is different, and sometimes different proportions are used.

[0041] It should be noted that the terms "first", "second", etc. in the specification, claims and drawings of the present invention are used to distinguish similar objects in order to describe the embodiments of the present invention, rather than to describe a specific order or sequence. It should be understood that the structures used in this way can be interchanged under appropriate circumstances. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0042] This embodiment provides a frog identification method based on image processing. Figure 1 As shown, the frog identification method comprises:

[0043] Obtain a color image of the frog to be marked, preferably the black-spotted sidefold frog, commonly known as the black and white frog, which has obvious lines and colors on its surface, and construct a 3D color standard model based on the obtained color image;

[0044] Extracting features from the 3D color standard model to obtain a set of identification features, wherein the parts from which the features are extracted include the head, the back, four fingers / toes, and four limbs;

[0045] Binding the identification feature group to an identification code, wherein the identification codes corresponding to different identification feature groups are different and unique;

[0046] Binding all the 3D color standard models with corresponding identification feature groups, and summarizing them to obtain a frog identification library;

[0047] Acquire a color image of the frog to be identified, and construct a 3D color entity model based on the acquired color image;

[0048] Comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in turn, wherein the basis for the comparative analysis is the identification feature group corresponding to the current 3D color standard model;

[0049] If the result of the comparative analysis exceeds the preset matching threshold, it is determined that the frog corresponding to the 3D color entity model and the frog corresponding to the 3D color standard model are the same frog, and the identification code corresponding to the 3D color standard model is assigned to the frog corresponding to the 3D color entity model.

[0050] Since most of the frogs used in the experiment are adult frogs, the surface patterns, convex hulls, etc. of adult frogs change little during their growth, and different frogs have different surface patterns, convex hulls, etc.; in addition, the patterns of frogs are concentrated on the head, back, and limbs, and the patterns on the fingers / toes (or webbed feet) of different frogs are also different. In view of this, the frog identification method based on image processing provided in this embodiment analyzes the surface image of the frog, and extracts the patterns, convex hulls, etc. on the surface of the frog's head, back, limbs, and fingers / toes as features, so as to achieve the distinction and identification of different frogs, and further, the frog can be identified without harming the frog, thereby solving the problem that the existing frog identification method causes great harm to frogs.

[0051] It should be noted that the fingers / toes described in the present invention specifically include the fingers of the two forelimbs and the toes of the two hind limbs of frogs. The naming and definition of fingers and toes are well known to those skilled in the art and will not be repeated here. However, it should be understood that the four fingers / toes mentioned in the present invention are specifically the fingers or toes at the ends of the limbs of frogs. Of course, due to different types of frogs, the fingers or toes may have webs, and the webs (including the lines of the webs) can also be extracted as the features of the fingers / toes. The feature extraction of other ends of the limbs without violating the main purpose of the present invention should also fall within the protection scope of the present invention.

[0052] Considering that in the process of photographing a frog to obtain a color image, the frog's body posture may be in various situations, which may cause some feature areas to be blocked in a single shooting direction. For example, when shooting from top to bottom, when the frog is in a prone position, the feature image of its fingers / toes may not be obtained.

[0053] To solve this problem, in the present embodiment, the frog can be placed on a transparent partition, and the frog can be photographed simultaneously by an upper camera and a lower camera. Preferably, the upper camera and the lower camera include a plurality of shooting lenses with different shooting angles to obtain the omnidirectional image of the frog. More preferably, each shooting lens can be rotated at an angle, adjusted in focus, and adjusted in white balance according to actual conditions, and a clear omnidirectional image of the frog can be obtained by automatic adjustment, so that the accuracy of feature extraction can be effectively improved, and the accuracy of identification can be guaranteed.

[0054] Furthermore, in this embodiment, the method of extracting features from the 3D color standard model to obtain the identification feature group includes:

[0055] First, the surface of the 3D color standard model is subjected to dot matrix pixel processing, and then the surface of the 3D color standard model forms a pixel array of countless RGB color blocks. Specifically, the dot matrix pixel processing method is well known to those skilled in the art of image processing, for example, after converting the original color image 2D into dot matrix pixels, the dot matrix pixels are attached to the surface of the 3D color standard model in proportion.

[0056] Next, refer to Figure 2 As shown in the dotted box, a×b pixel blocks with the largest RGB color difference values ​​are selected from the head, back, 4 fingers / toes and 4 limbs as the first identification features, where the sizes of pixel blocks in different areas are not exactly the same. Since the pixel block is selected from the area with the largest color difference value, in order to reflect the color difference, the values ​​of a and b must be ≥2.

[0057] Since the frog's pattern is formed by the color difference between different colors, selecting the parts with the largest color difference values ​​on the head, back, four fingers / toes and four limbs can reflect the frog's pattern characteristics, that is, the first identification feature reflects the frog's pattern and pattern color.

[0058] Then, similar to the selection of the first identification feature, pixel blocks of c×d pixels with the largest brightness difference are selected from the head, back, 4 fingers / toes and 4 limbs as the second identification feature, where the sizes of pixel blocks in different areas are not exactly the same. Similarly, since the pixel block selects the area with the largest brightness difference, in order to reflect the brightness difference, the values ​​of c and d must be ≥2.

[0059] Since the convex hull or concave hull on the frog's body will form a difference in brightness during the image formation process, the convex hull or concave hull features on the frog's body can be clearly seen through the brightness difference. Therefore, selecting the area with the largest brightness difference can reflect the convex hull characteristics of the frog and reduce the misjudgment caused by the influence of the brightness of the shooting environment in the subsequent comparison process.

[0060] Finally, all the first identification features are aggregated to obtain a first identification feature group, all the second identification features are aggregated to obtain a second identification feature group, and the first identification feature group and the second identification feature group are aggregated to obtain an identification feature group.

[0061] It should be noted that a frog forms a 3D color standard model, which is used as a subsequent comparison sample, in which the identification feature groups can also be associated. A 3D color standard model is associated with 10 (1 head, 1 back, 4 fingers / toes and 4 limbs) first identification features and 10 second identification features.

[0062] It should also be noted that in the actual identification feature selection process, the pixel blocks selected for the first identification feature and the second identification feature of the same part can be the same or different; and the sizes of the pixel blocks of the first identification features of different parts can be the same or different, and the sizes of the pixel blocks of the second identification features of different parts can be the same or different.

[0063] In the specific application process, the selection of pixel blocks can be divided manually or automatically through a computer preset image processing selection program. The software intelligent selection method of pixel blocks without violating the main purpose of the present invention should also belong to the protection scope of the present invention. Using computer software to automatically judge and select the first identification feature and the second identification feature can save identification time and improve identification efficiency when the frog sample size is large.

[0064] Further, in this embodiment, the method of comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in sequence includes:

[0065] Step 1, scaling the 3D color entity model until the size of the 3D color entity model is consistent with the size of the currently compared 3D color standard model. Since the sizes of the frogs photographed are different when acquiring color images (depending on the growth of the frogs, the shooting distance, etc.), the sizes of the obtained 3D color entity model and the 3D color standard model will be different. In order to improve the comparison accuracy of the first identification feature and the second identification feature, the 3D color entity model is first scaled to make the size of the 3D color entity model consistent with that of the 3D color standard model. Preferably, the body shape of the 3D color entity model is consistent with that of the 3D color standard model, so that after the size of the model is adjusted to be consistent, the two can completely overlap in three dimensions.

[0066] Step 2: Set the coordinates of all pixel blocks in the identification feature group corresponding to the current 3D color standard model in the 3D color entity model to obtain a feature selection box on the 3D color entity model. Specifically, a three-dimensional coordinate system is established with the basic point in the 3D color standard model as the origin, and the three-dimensional coordinate points of all first identification features and all second identification features are obtained; when the 3D color entity model and the 3D color standard model can be completely overlapped, a three-dimensional coordinate system can be established with a corresponding basic point on the 3D color entity model as the origin, so that the three-dimensional positions of all first identification features and all second identification features can be obtained, and the pixel blocks selected in the box are the feature selection box.

[0067] Of course, in the specific application process, if the 3D color entity model and the 3D color standard model are inconsistent, it is necessary to convert the position coordinates through the relevant image calculation conversion method to obtain the corresponding feature selection box. In order to facilitate the comparison of features, this embodiment sets a model with the same body shape in the modeling stage, which can reduce the complexity of feature comparison operations and improve analysis accuracy and efficiency.

[0068] Step three, compare and analyze the identification feature and the feature selection box of the same coordinates. If the overlap degree is greater than a preset overlap threshold, it is determined that the identification feature and the feature selection box are the same.

[0069] Specifically, in this embodiment, the first identification feature of the same coordinates is compared and analyzed with the feature selection box. If the RGB pixel arrangement overlap is greater than 75%, the first identification feature is determined to be the same as the feature selection box; the second identification feature of the same coordinates is compared and analyzed with the feature selection box. If the brightness value distribution overlap is greater than 70%, the second identification feature is determined to be the same as the feature selection box. For example, when the 3D color entity model established by frog A is compared with the 3D color standard model with sequence 003 in the frog identification library, the RGB pixel arrangement overlap of the head is 82%, indicating that the head patterns of the two are the same; if the RGB pixel arrangement overlap of the back is 36%, it indicates that the back patterns of the two are different; and if the brightness value distribution overlap of the back is 50%, it indicates that the convex dot patterns of the two backs are different.

[0070] Step 4: If the matching degree of all identification features corresponding to the current 3D color standard model and the feature selection box of the 3D color entity model is greater than a preset matching threshold, the 3D color entity model is determined to be the same as the current 3D color standard model, otherwise the next 3D color standard model is replaced and the above steps 1 to 3 are repeated.

[0071] Specifically, in this embodiment, if the number of features that are determined to be the same as the feature selection box accounts for more than 80% of the number of all first identification features, then the 3D color entity model is determined to be the same as the current 3D color standard model. For example, in a comparison, among 10 first identification features, 8 are determined to be the same, accounting for 80%, indicating that the 3D color entity model is the same as the current 3D color standard model.

[0072] For special cases, if there are multiple 3D color standard models corresponding to a proportion greater than 80%, the 3D color standard model corresponding to the one with the largest proportion is selected to be the same as the 3D color entity model. For example, in the first comparison (such as comparison with 3D color standard model No. 001 in the identification library), its proportion is 80%, in the second comparison (such as comparison with 3D color standard model No. 003 in the identification library), its proportion is 90%, and in the third comparison (such as comparison with 3D color standard model No. 008 in the identification library), its proportion is 80%, then it is finally determined that the second comparison (i.e., 3D color standard model No. 003 in the identification library) is the same as the 3D color entity model.

[0073] Since the pattern is more obvious than the convex dot pattern in the imaging, the coincidence of the pattern (first identification feature) is confirmed first in this embodiment. In practical applications, it is also necessary to consider how to identify and match the frogs when the pattern comparison cannot confirm whether it matches.

[0074] This embodiment provides a method to solve the above-mentioned shortcomings:

[0075] If the number of features determined to be identical to the feature selection box for the first identification feature accounts for more than 50% and less than 80% of the number of all first identification features, then the ratio of the number of features determined to be identical to the feature selection box for the second identification feature accounts for the number of all second identification features;

[0076] If the number of features that are determined to be the same as the feature selection box in the second identification features accounts for more than 85% of all the second identification features, then it is determined that the 3D color entity model is the same as the current 3D color standard model.

[0077] The reason why the first identification feature is limited to a proportion greater than 50% is that if there is a clear lack of pattern, it is sufficient to determine that the two models do not correspond to the same frog. At this time, it is meaningless to consider the overlap of the convex points (second identification feature), which will increase the misjudgment of the match. In addition, when determining the second identification feature, the overlap ratio is set higher because the convex hull of the frog's surface is not obvious during shooting, so it is necessary to increase the overlap to improve the accuracy of the match.

[0078] This embodiment also provides a frog identification system based on image processing, such as Figure 3As shown, the frog identification system includes an image acquisition module, a 3D modeling module, a feature processing module and an identification binding module; the image acquisition module is used to acquire the color image of the frog; the 3D modeling module is used to construct a 3D color standard model and a 3D color entity model according to the color image; the feature processing module is used to extract features from the 3D color standard model and perform feature comparison analysis on the 3D color entity model; the identification binding module is used to generate an identification code for the 3D color standard model and match the identification code to the 3D color entity model according to the result of the feature comparison analysis.

[0079] Specifically, the 3D modeling module includes an image analysis unit, a posture correction unit and a standard model generation unit; the image analysis unit is used to analyze and process the color image of the frog to distinguish the structural characteristics of different areas of the frog; the posture correction unit is used to reconstruct the posture of the frog according to the acquired color image, so that all constructed 3D color standard models and 3D color entity models have the same standard posture; the standard model generation unit is used to construct the image features of different areas of the frog in the color image on the standard posture according to the analysis result of the image analysis unit, so as to generate the 3D color standard model and the 3D color entity model.

[0080] Furthermore, the feature processing module includes a standard model feature processing unit and a solid model feature processing unit; the standard model feature processing unit is used to extract features from the 3D color standard model to obtain an identification feature group; the solid model feature processing unit is used to compare and analyze the 3D color solid model with all the 3D color standard models in turn according to the identification feature group to determine whether the 3D color solid model is the same as the 3D color standard model.

[0081] In addition, the identification binding module includes an identification generation unit and an identification matching unit; the identification generation unit is used to generate a unique identification code for each 3D color standard model; the identification matching unit is used to match the identification code to the 3D color entity model according to the result of feature comparison analysis.

[0082] In the specific application process, the image acquisition module can be a high-definition camera, and the 3D modeling module, the feature processing module and the identification binding module can rely on a computer to form a software algorithm to realize the corresponding functions. Of course, in addition to this, the frog identification system based on image processing provided in this embodiment can also include an experimental data entry module, an experimental report generation module, etc., so that after obtaining the experimental data, in order to correspond the experimental data to the frog, after obtaining the frog's identification code, the corresponding experimental data can be directly entered into the system, and then the data can be automatically summarized and sorted through the system, and a complete experimental report can be output to avoid mistakes in the subsequent manual additional identification code and frog experimental data corresponding work.

[0083] In summary, the frog identification method and system based on image processing provided by the present embodiment include: obtaining a color image of a frog to be identified, and constructing a 3D color standard model according to the obtained color image; extracting features from the 3D color standard model to obtain an identification feature group; binding the identification feature group with an identification code; binding all the 3D color standard models with the corresponding identification feature group, and summarizing them to obtain a frog identification library; obtaining a color image of a frog to be identified, and constructing a 3D color entity model according to the obtained color image; comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in turn, wherein the basis for the comparison and analysis is the identification feature group corresponding to the current 3D color standard model; if the result of the comparison and analysis exceeds a preset matching threshold, it is determined that the frog corresponding to the 3D color entity model and the frog corresponding to the 3D color standard model are the same, and the identification code corresponding to the 3D color standard model is assigned to the frog corresponding to the 3D color entity model. Since the surface patterns of frogs change little during their growth, and different frogs have different surface patterns, different frogs can be distinguished by analyzing their surface images, and the frogs can be identified without harming them, thus solving the problem that existing frog identification methods cause great harm to frogs.

[0084] The above description is only a description of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention. Any changes or modifications made by a person skilled in the art in the field of the present invention based on the above disclosure shall fall within the scope of protection of the claims.

Claims

1. A frog identification method based on image processing, characterized in that: The frog identification method comprises: Acquire a color image of the frog to be identified, and construct a 3D color standard model based on the acquired color image; Extracting features from the 3D color standard model to obtain a set of identification features, wherein the parts from which the features are extracted include the head, the back, four fingers / toes, and four limbs; The method for extracting features from the 3D color standard model to obtain an identification feature group includes: Performing dot matrix pixel processing on the surface of the 3D color standard model; Select a×b pixel blocks with the largest RGB color difference values ​​from the head, back, four fingers / toes and four limbs as the first identification features, where the sizes of pixel blocks in different areas are not exactly the same; Select the pixel blocks of c×d pixels with the largest brightness difference from the head, back, four fingers / toes and four limbs as the second identification features, where the sizes of pixel blocks in different areas are not exactly the same; Aggregating all the first identification features to obtain a first identification feature group, aggregating all the second identification features to obtain a second identification feature group, and aggregating the first identification feature group and the second identification feature group to obtain an identification feature group; Binding the identification feature group to an identification code, wherein the identification codes corresponding to different identification feature groups are different and unique; Binding all the 3D color standard models with corresponding identification feature groups, and summarizing them to obtain a frog identification library; Acquire a color image of the frog to be identified, and construct a 3D color entity model based on the acquired color image; Comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in turn, wherein the basis for the comparative analysis is the identification feature group corresponding to the current 3D color standard model; If the result of the comparative analysis exceeds the preset matching threshold, it is determined that the frog corresponding to the 3D color entity model and the frog corresponding to the 3D color standard model are the same frog, and the identification code corresponding to the 3D color standard model is assigned to the frog corresponding to the 3D color entity model.

2. The frog identification method based on image processing according to claim 1, characterized in that: The method of comparing and analyzing the 3D color entity model with all the 3D color standard models in the frog identification library in sequence comprises: Scaling the 3D color entity model until the size of the 3D color entity model is consistent with the size of the 3D color standard model currently being compared; Set all pixel block coordinates in the identification feature group corresponding to the current 3D color standard model in the 3D color entity model to obtain a feature selection box on the 3D color entity model; Compare and analyze the identification feature and the feature selection box of the same coordinates. If the degree of overlap is greater than the preset overlap threshold, the identification feature is determined to be the same as the feature selection box. If the matching degree of all identification features corresponding to the current 3D color standard model and the feature selection box of the 3D color entity model is greater than a preset matching threshold, the 3D color entity model is determined to be the same as the current 3D color standard model, otherwise the next 3D color standard model is replaced and the above steps are repeated.

3. The frog identification method based on image processing according to claim 2, characterized in that: Methods for comparing and analyzing identification features and feature selection boxes of the same coordinates include: Compare and analyze the first identification feature and the feature selection box with the same coordinates. If the RGB pixel arrangement overlap is greater than 75%, the first identification feature is determined to be the same as the feature selection box. The second identification feature and the feature selection box with the same coordinates are compared and analyzed. If the brightness value distribution overlap is greater than 70%, it is determined that the second identification feature is the same as the feature selection box.

4. The frog identification method based on image processing according to claim 3 is characterized in that: The method for determining whether the 3D color entity model is the same as the current 3D color standard model comprises: If it is determined that the number of features that are identical to the first identification feature and the feature selection box accounts for more than 80% of the number of all first identification features, then it is determined that the 3D color solid model is identical to the current 3D color standard model; If there are multiple 3D color standard models corresponding to a proportion greater than 80%, the 3D color standard model corresponding to the one with the largest proportion is selected to be the same as the 3D color entity model.

5. The frog identification method based on image processing according to claim 4 is characterized in that: The method for determining whether the 3D color entity model is the same as the current 3D color standard model further includes: If the number of features determined to be identical to the feature selection box for the first identification feature accounts for more than 50% and less than 80% of the number of all first identification features, then the ratio of the number of features determined to be identical to the feature selection box for the second identification feature accounts for the number of all second identification features; If the number of features that are determined to be the same as the feature selection box in the second identification features accounts for more than 85% of all the second identification features, then it is determined that the 3D color entity model is the same as the current 3D color standard model.

6. A frog identification system based on image processing, used for performing the frog identification method based on image processing as claimed in any one of claims 1 to 5, characterized in that: The frog identification system includes an image acquisition module, a 3D modeling module, a feature processing module and an identification binding module; the image acquisition module is used to acquire a color image of a frog; the 3D modeling module is used to construct a 3D color standard model and a 3D color entity model according to the color image; the feature processing module is used to extract features from the 3D color standard model and perform feature comparison analysis on the 3D color entity model; the identification binding module is used to generate an identification code for the 3D color standard model and match the identification code to the 3D color entity model according to the result of the feature comparison analysis.

7. The frog identification system based on image processing according to claim 6, characterized in that: The 3D modeling module includes an image analysis unit, a posture correction unit and a standard model generation unit; the image analysis unit is used to analyze and process the color image of the frog to distinguish the structural characteristics of different areas of the frog; the posture correction unit is used to reconstruct the posture of the frog according to the acquired color image, so that all constructed 3D color standard models and 3D color entity models have the same standard posture; the standard model generation unit is used to construct the image features of different areas of the frog in the color image on the standard posture according to the analysis result of the image analysis unit, so as to generate a 3D color standard model and a 3D color entity model.

8. The frog identification system based on image processing according to claim 6, characterized in that: The feature processing module includes a standard model feature processing unit and a solid model feature processing unit; the standard model feature processing unit is used to extract features from the 3D color standard model to obtain an identification feature group; the solid model feature processing unit is used to compare and analyze the 3D color solid model with all the 3D color standard models in turn according to the identification feature group to determine whether the 3D color solid model is the same as the 3D color standard model.

9. The frog identification system based on image processing according to claim 6, characterized in that: The identification binding module includes an identification generation unit and an identification matching unit; the identification generation unit is used to generate a unique identification code for each 3D color standard model; the identification matching unit is used to match the identification code to the 3D color entity model according to the result of feature comparison analysis.

Images