Plant leaf intelligent detection method and system based on data analysis

Abstract

The application discloses a kind of plant leaf intelligent detection method and system based on data analysis, belong to plant leaf detection technical field.Extract leaf lesion and healthy area contour and repair, generate parameter set before and after repair;Two kinds of nidus density error coefficient are calculated;Coordinate point is constructed and the degree of nidus lesion is calculated by evaluation model;Lesion diffusion circle is described, diffusion path is identified and exhibited.The application improves the accuracy of error coefficient calculation by contour overlap repair and double parameter set construction, realizes lesion degree quantitative evaluation by constructing coordinate point with double error coefficient, identifies diffusion path based on diffusion circle overlap, solves the problems of low efficiency, poor precision and inability to trace lesion propagation in traditional detection, provides precise data support for plant disease control, and is suitable for intelligent detection scenarios of various plant leaves in agricultural production.

Description

Technical Field

[0001] This invention relates to the field of plant leaf detection technology, specifically to an intelligent plant leaf detection method and system based on data analysis. Background Technology

[0002] In agricultural production, timely and accurate detection of plant leaf diseases is a core element in ensuring crop yield and quality. Traditional leaf detection relies on manual visual observation, requiring inspectors to judge the area, extent, and spread trend of lesions based on experience. However, this method has many limitations: First, manual detection is extremely inefficient and cannot meet the batch detection needs of large-scale planting scenarios. Moreover, the detection results are easily affected by subjective experience, ambient light, and other factors, resulting in poor consistency and reliability. Second, manual detection cannot quantify the extent of lesions or accurately identify the spread path of lesions. This leads to a lack of scientific data support for the formulation of disease control measures, which can easily result in problems such as untimely control, excessive use of pesticides, or deviations in the scope of control, causing economic losses and environmental pollution.

[0003] With the development of intelligent detection technology, image recognition-based methods for detecting plant leaves are gradually being applied, but existing technologies still have key shortcomings. First, existing methods mostly extract leaf contours directly for analysis without considering the overlap between diseased and healthy areas, leading to inaccurate leaf parameter extraction and affecting the reliability of subsequent detection results. Second, when calculating the error coefficient related to lesions, existing technologies do not combine the differences in parameters before and after repair with the area of ​​the affected region for comprehensive calculation, resulting in insufficient specificity and accuracy of the error coefficient, failing to truly reflect the density characteristics of the lesions. Third, existing lesion severity assessments mostly use single indicators and do not construct multi-dimensional assessment models, making it difficult to achieve quantitative analysis of lesion severity, and the detection results lack intuitiveness and comparability. Finally, most existing methods can only identify the location of lesion areas and cannot trace the path of lesion spread, which is not conducive to users understanding the disease transmission patterns and making it difficult to formulate targeted prevention and control strategies. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for intelligent detection of plant leaves based on data analysis, so as to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] A data analysis-based intelligent detection system for plant leaves includes: a contour extraction and parameter generation module, an error coefficient calculation module, a lesion severity assessment module, and a diffusion path analysis and display module. The contour extraction and parameter generation module extracts and repairs the contours of lesions and healthy regions in leaf images, generating leaf parameter sets before and after repair. The error coefficient calculation module calculates the lesion density error coefficient based on the parameter set and the area of ​​the region contours. The lesion severity assessment module constructs lesion density error coordinate points and calculates the lesion severity of each lesion using an assessment model. The diffusion path analysis and display module delineates the lesion diffusion circle, identifies the lesion diffusion path, and sends it to a display terminal.

[0007] As a preferred embodiment of the present invention, the contour extraction and parameter generation module includes a region contour extraction unit, a contour overlap repair unit, and a parameter set generation unit; the region contour extraction unit is used to extract the contours of diseased areas and healthy areas from the acquired leaf images respectively; the contour overlap repair unit is used to overlap and repair the extracted two types of region contours to form a complete leaf contour; the parameter set generation unit is used to obtain the leaf parameters of the complete leaf contours before and after repair, and generate the corresponding basic parameter set and complete parameter set.

[0008] In a preferred embodiment of the present invention, the error coefficient calculation module includes a region area acquisition unit, a lesion feature error calculation unit, and a health feature error calculation unit; the region area acquisition unit is used to acquire the contour areas of the lesion region and the healthy region in the complete leaf contour after repair; the lesion feature error calculation unit calculates the lesion feature foci density error coefficient based on the basic parameter set, the complete parameter set, and the lesion region contour area; the health feature error calculation unit calculates the health feature foci density error coefficient based on the basic parameter set, the complete parameter set, and the healthy region contour area.

[0009] As a preferred embodiment of the present invention, the lesion severity assessment module includes a coordinate point construction unit and a lesion severity calculation unit; the coordinate point construction unit is used to construct a density error coordinate point for each lesion with the density error coefficient of the lesion feature lesion as the abscissa and the density error coefficient of the healthy feature lesion as the ordinate; the lesion severity calculation unit is used to construct a lesion severity assessment model, substitute the error coefficients corresponding to the coordinate points, and calculate the lesion severity of each lesion.

[0010] As a preferred embodiment of the present invention, the diffusion path analysis and display module includes a lesion diffusion circle characterization unit, a diffusion path identification unit, and a result sending unit; the lesion diffusion circle characterization unit is used to characterize the lesion diffusion circle of each lesion point with the density error coordinate point of each lesion point as the center and the corresponding lesion degree as the radius; the diffusion path identification unit is used to determine whether the lesion diffusion circles of adjacent lesions overlap, and construct and expand the lesion diffusion path identifier based on the overlap; the result sending unit is used to obtain the complete lesion diffusion path identifier and send it to the intelligent detection and display terminal.

[0011] A data analysis-based intelligent detection method for plant leaves, comprising the following steps:

[0012] Step S1: Extract the contours of the diseased and healthy areas in the acquired leaf images, and form a complete leaf contour after overlapping repair. Obtain the leaf parameters before and after repair to generate the corresponding parameter sets.

[0013] Step S2: Based on the parameter set and the contour area of ​​the lesion and healthy regions, evaluate the foci density error coefficients of the lesion features and healthy features of the leaf image, respectively.

[0014] Step S3: Construct lesion density error coordinate points using two types of error coefficients as coordinate values, and quantify the lesion degree of each lesion by constructing a lesion degree assessment model;

[0015] Step S4: Draw a lesion diffusion circle with the lesion density error coordinate point as the center and the corresponding lesion degree as the radius. Identify the lesion diffusion path based on the overlap of adjacent lesion diffusion circles and send the identified lesion diffusion path to the intelligent detection and display terminal.

[0016] As a preferred embodiment of the present invention, the specific implementation process of step S1 includes:

[0017] The contours of the diseased area and the healthy area in the leaf image of the plant are extracted separately. After the contours of the diseased area and the healthy area are repaired by overlapping, a complete leaf contour is formed. The complete leaf contour includes the contours of the diseased area and the healthy area.

[0018] For the complete blade profile formed after contour overlap repair, obtain the blade parameters of the complete blade profile before repair to generate the blade basic parameter set. Obtain the blade parameters of the complete blade outline after repair to generate a complete blade parameter set. Where x represents the leaf image number, i is the hearth number, and I is the total number of hearths. The diameter of the stove point before repair. The diameter of the repaired stove point.

[0019] As a preferred embodiment of the present invention, the specific implementation process of step S2 includes:

[0020] For the complete leaf outline formed after outline overlap repair, the outline area of ​​the diseased area and the outline area of ​​the healthy area are obtained respectively, and denoted as follows: and ;

[0021] Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the lesion region, the lesion feature foci density error coefficient of leaf image x is constructed. ;

[0022] Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the healthy region, the density error coefficient of the healthy feature foci in leaf image x is constructed. .

[0023] As a preferred embodiment of the present invention, the specific implementation process of step S3 includes:

[0024] Using the density error coefficient of lesion features as the horizontal coordinate value and the density error coefficient of healthy features as the vertical coordinate value, the lesion density error coordinate points of leaf image x are constructed. ;

[0025] Construct a lesion severity assessment model to evaluate the lesion severity of lesion point i in leaf image x. In the formula, max{} and min{} are the maximum and minimum value functions, respectively.

[0026] As a preferred embodiment of the present invention, the specific implementation process of step S4 includes:

[0027] Using the density error coordinates of the stove point Centered on the point of the circle, according to the degree of the lesion Use radius i to depict the lesion spread circle of lesion point i. ;

[0028] Among the lesions adjacent to lesion i, if the lesion spread circle of lesion i... The lesion spread circle of foci j If there is overlap, a lesion spread path marker is formed between lesion i and lesion j, with lesion i as the starting point for tracing the source. If the lesion spread circle of lesion i... The lesion spread circle of foci j If there is no overlap, it is determined that no lesion spread path marker can be formed between lesion point i and lesion point j, where i ≠ j. The degree of lesion at foci j in leaf image x;

[0029] For the lesion spread path marker formed between lesion i and lesion j, among the lesions adjacent to lesion j, trace the lesions other than lesion i that form the lesion spread path marker with lesion j, and determine whether lesion j can form a lesion spread path marker with lesion j other than lesion i. If it is determined that lesion j can form a lesion spread path marker with lesion j other than lesion i, then include lesion j and lesion j other than lesion i in the lesion spread path marker with lesion i as the starting point of tracing.

[0030] Obtain the lesion spread path identifier with foci i as the starting point for tracing, and send it to the intelligent detection and display terminal to display the tracing results of leaf image x.

[0031] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention takes the contour features and data analysis of leaf images as its core and constructs a full-process detection logic of "contour repair - parameter extraction - error calculation - severity assessment - path tracing". First, the contours of leaf lesions and healthy areas are separated through image processing technology, and contour overlap is repaired to eliminate overlap interference. Unlike existing schemes that directly extract contours, this method can eliminate the interference of contour overlap on parameter extraction. Second, leaf parameters before and after repair are extracted to form a dual parameter set. Combined with the area of ​​lesions and healthy areas, a targeted error coefficient calculation model is constructed to accurately characterize the lesion density features. Then, a two-dimensional coordinate point is constructed using the dual error coefficients as the horizontal and vertical axes. The lesion severity assessment model is constructed using the principle of extreme value normalization, which transforms complex lesion features into quantifiable values. Unlike existing single-dimensional assessment methods, this method makes the lesion severity intuitive and comparable, allowing users to quickly grasp the severity of the disease. Finally, based on the lesion severity, a diffusion circle is depicted. By the overlap relationship of adjacent diffusion circles, the lesion propagation path is traced, achieving an upgrade from "point detection" to "line tracing", which helps to solve the problem that existing technologies cannot trace the lesion propagation path. Attached Figure Description

[0032] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0033] Figure 1 This is a schematic diagram illustrating the steps of an intelligent detection method for plant leaves based on data analysis, as described in this invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] In this first embodiment: a plant leaf intelligent detection system based on data analysis is provided, the system comprising:

[0036] Please see Figure 1 In this second embodiment, a data analysis-based intelligent detection method for plant leaves is provided, applicable to the first embodiment above. This embodiment uses leafy vegetable leaf detection in a facility agriculture base as the application scenario. The base has a planting area of ​​50 mu and adopts a greenhouse planting mode, which is prone to leaf diseases such as early blight and late blight, requiring regular batch detection. Leafy vegetable leaf images are collected using a high-definition industrial camera (resolution 1920×1080), and a total of 100 leaf images containing different degrees of disease are collected (30 images of mild disease, 40 images of moderate disease, and 30 images of severe disease). After grayscale conversion and Gaussian denoising (standard deviation σ=0.8), the images are used as leaf image samples of this invention to ensure that the images are free from obvious noise interference.

[0037] The method includes the following steps:

[0038] Step S1: Extract the contours of the diseased and healthy areas in the acquired leaf images, and form a complete leaf contour after overlapping repair. Obtain the leaf parameters before and after repair to generate the corresponding parameter sets.

[0039] For example, the contours of the diseased area and the healthy area in the leaf image of the plant are extracted respectively. After the contours of the diseased area and the healthy area are repaired by overlapping, a complete leaf contour is formed. The complete leaf contour includes the contours of the diseased area and the healthy area.

[0040] For the complete blade profile formed after contour overlap repair, obtain the blade parameters of the complete blade profile before repair to generate the blade basic parameter set. Obtain the blade parameters of the complete blade outline after repair to generate a complete blade parameter set. Where x represents the leaf image number, i is the hearth number, and I is the total number of hearths. The diameter of the stove point before repair. The diameter of the repaired stoker point;

[0041] For example, the contours of the lesion area and the healthy area are extracted from each image. Taking a leaf image with moderate lesion as an example, a total of 8 lesion points were identified. There are 3 overlaps between the contours of the lesion area and the healthy area (the overlapping area is about 1.2 mm²). The contour fusion algorithm is used to repair the overlapping parts to form a complete leaf contour.

[0042] Step S2: Based on the parameter set and the contour area of ​​the lesion and healthy regions, evaluate the foci density error coefficients of the lesion features and healthy features of the leaf image, respectively.

[0043] For example, for the complete leaf outline formed after outline overlap repair, the outline area of ​​the diseased area and the outline area of ​​the healthy area are obtained respectively, and denoted as follows: and ;

[0044] Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the lesion region, the lesion feature foci density error coefficient of leaf image x is constructed. ;

[0045] Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the healthy region, the density error coefficient of the healthy feature foci in leaf image x is constructed. ;

[0046] For example, the outline area of ​​the lesion region of the moderately diseased leaf was measured to be 12.5 mm², and the outline area of ​​the healthy region was 87.5 mm², using the pixel point statistical method. Based on the dual parameter set and the lesion region area, the density error coefficient of the lesion feature lesion ...

[0047] Step S3: Construct lesion density error coordinate points using two types of error coefficients as coordinate values, and quantify the lesion degree of each lesion by constructing a lesion degree assessment model;

[0048] For example, the density error coordinates of lesion feature foci are constructed by using the horizontal coordinate value as the horizontal coordinate value and the density error coefficient of healthy feature foci as the vertical coordinate value. ;

[0049] Construct a lesion severity assessment model to evaluate the lesion severity of lesion point i in leaf image x. In the formula, max{} and min{} are the functions for maximizing and minimizing the values, respectively;

[0050] For example, using the lesion feature error coefficient of each lesion as the x-axis and the health feature error coefficient as the y-axis, eight lesion density error coordinate points are constructed; the lesion degree of each lesion is calculated by the evaluation model, and the result ranges from 0.32 to 0.78 (the larger the value, the more severe the lesion). Among them, the lesion degree of 3 lesions is ≥0.6 (judged as severe lesion lesions), and the lesion degree of 5 lesions is <0.6 (judged as moderate lesion lesions).

[0051] Step S4: Draw a lesion diffusion circle with the lesion density error coordinate point as the center and the corresponding lesion degree as the radius. Identify the lesion diffusion path based on the overlap of adjacent lesion diffusion circles and send the identified lesion diffusion path to the intelligent detection and display terminal.

[0052] For example, using the stove density error coordinate point Centered on the point of the circle, according to the degree of the lesion Use radius i to depict the lesion spread circle of lesion point i. ;

[0053] Among the lesions adjacent to lesion i, if the lesion spread circle of lesion i... The lesion spread circle of foci j If there is overlap, a lesion spread path marker is formed between lesion i and lesion j, with lesion i as the starting point for tracing the source. If the lesion spread circle of lesion i... The lesion spread circle of foci j If there is no overlap, it is determined that no lesion spread path marker can be formed between lesion point i and lesion point j, where i ≠ j. The degree of lesion at foci j in leaf image x;

[0054] For the lesion spread path marker formed between lesion i and lesion j, among the lesions adjacent to lesion j, trace the lesions other than lesion i that form the lesion spread path marker with lesion j, and determine whether lesion j can form a lesion spread path marker with lesion j other than lesion i. If it is determined that lesion j can form a lesion spread path marker with lesion j other than lesion i, then include lesion j and lesion j other than lesion i in the lesion spread path marker with lesion i as the starting point of tracing.

[0055] Obtain the lesion spread path identifier with foci i as the starting point for tracing the source, and send it to the intelligent detection and display terminal to display the source tracing results of leaf image x;

[0056] For example, using each coordinate point as the center and the corresponding lesion severity as the radius (radius range of 0.32-0.78 coordinate units), eight lesion diffusion circles are drawn. The overlap of diffusion circles of adjacent lesions is judged, and it is found that the diffusion circles of four adjacent lesions overlap (overlapping area ratio ≥30%). Taking the leftmost severe lesion as the starting point for tracing the source, the lesion diffusion path identifier is constructed and expanded, and the propagation path of "starting lesion → adjacent moderate lesion → distal moderate lesion" is successfully traced. The location of the lesion area of ​​the leaf, the severity of each lesion, and the diffusion path are sent to the intelligent detection and display terminal, and users can intuitively view the disease details through the terminal.

[0057] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0058] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent detection of plant leaves based on data analysis, characterized in that, The method includes the following steps: Step S1: Extract the contours of the diseased and healthy areas in the acquired leaf images, and form a complete leaf contour after overlapping repair. Obtain the leaf parameters before and after repair to generate the corresponding parameter sets. Step S2: Based on the parameter set and the contour area of ​​the lesion and healthy regions, evaluate the foci density error coefficients of the lesion features and healthy features of the leaf image, respectively. Step S3: Construct lesion density error coordinate points using two types of error coefficients as coordinate values, and quantify the lesion degree of each lesion by constructing a lesion degree assessment model; Step S4: Draw a lesion diffusion circle with the lesion density error coordinate point as the center and the corresponding lesion degree as the radius. Identify the lesion diffusion path based on the overlap of adjacent lesion diffusion circles and send the identified lesion diffusion path to the intelligent detection and display terminal. The specific implementation process of step S1 includes: The contours of the diseased and healthy areas in the leaf images of the collected plants are extracted separately. After the contours of the diseased and healthy areas are repaired by overlapping, a complete leaf contour is formed, which includes the contours of the diseased and healthy areas. For the complete blade profile formed after contour overlap repair, obtain the blade parameters of the complete blade profile before repair to generate the blade basic parameter set. Obtain the blade parameters of the complete blade outline after repair to generate a complete blade parameter set. Where x represents the leaf image number, i is the hearth number, and I is the total number of hearths. The diameter of the stove point before repair. The diameter of the repaired stoker point; The specific implementation process of step S2 includes: For the complete leaf outline formed after outline overlap repair, the outline area of ​​the diseased area and the outline area of ​​the healthy area are obtained respectively, and denoted as follows: and ; Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the lesion region, the lesion feature foci density error coefficient of leaf image x is constructed. ; Based on the leaf's basic parameter set, complete leaf parameter set, and the contour area of ​​the healthy region, the density error coefficient of the healthy feature foci in leaf image x is constructed. ; The specific implementation process of step S3 includes: Using the density error coefficient of lesion features as the horizontal coordinate value and the density error coefficient of healthy features as the vertical coordinate value, the lesion density error coordinate points of leaf image x are constructed. ; Construct a lesion severity assessment model to evaluate the lesion severity of lesion point i in leaf image x. In the formula, max{} and min{} are the maximum and minimum value functions, respectively.

2. The intelligent detection method for plant leaves based on data analysis according to claim 1, characterized in that, The specific implementation process of step S4 includes: Using the density error coordinates of the stove point Centered on the point of the circle, according to the degree of the lesion Use radius i to depict the lesion spread circle of lesion point i. ; Among the lesions adjacent to lesion i, if the lesion spread circle of lesion i... The lesion spread circle of foci j If there is overlap, a lesion spread path marker is formed between lesion i and lesion j, with lesion i as the starting point for tracing the source. If the lesion spread circle of lesion i... The lesion spread circle of foci j If there is no overlap, it is determined that no lesion spread path marker can be formed between lesion point i and lesion point j, where i ≠ j. The degree of lesion at foci j in leaf image x; For the lesion spread path marker formed between lesion i and lesion j, among the lesions adjacent to lesion j, trace the lesions other than lesion i that form the lesion spread path marker with lesion j, and determine whether lesion j can form a lesion spread path marker with lesion j other than lesion i. If it is determined that lesion j can form a lesion spread path marker with lesion j other than lesion i, then include lesion j and lesion j other than lesion i in the lesion spread path marker with lesion i as the starting point of tracing. Obtain the lesion spread path identifier with foci i as the starting point for tracing, and send it to the intelligent detection and display terminal to display the tracing results of leaf image x.

3. A data-analysis-based intelligent detection system for plant leaves, executing the data-analysis-based intelligent detection method for plant leaves as described in any one of claims 1-2, characterized in that, The system includes: a contour extraction and parameter generation module, an error coefficient calculation module, a lesion severity assessment module, and a diffusion path analysis and display module. The contour extraction and parameter generation module extracts and repairs the contours of lesion and healthy regions in leaf images, generating leaf parameter sets before and after repair. The error coefficient calculation module calculates the lesion density error coefficients for lesion and healthy features based on the parameter sets and the area of ​​the region contours. The lesion severity assessment module constructs lesion density error coordinate points and calculates the lesion severity of each lesion using an assessment model. The diffusion path analysis and display module delineates the lesion diffusion circle, identifies the lesion diffusion path, and sends it to the display terminal.

4. The intelligent plant leaf detection system based on data analysis according to claim 3, characterized in that, The contour extraction and parameter generation module includes a region contour extraction unit, a contour overlap repair unit, and a parameter set generation unit; the region contour extraction unit is used to extract the contours of diseased areas and healthy areas from the acquired leaf images, respectively. The contour overlap repair unit is used to repair the overlap of the extracted two types of region contours to form a complete leaf contour. The parameter set generation unit is used to obtain the blade parameters of the complete blade outline before and after repair, and generate the corresponding basic parameter set and complete parameter set.

5. The intelligent detection system for plant leaves based on data analysis according to claim 4, characterized in that, The error coefficient calculation module includes a region area acquisition unit, a lesion feature error calculation unit, and a health feature error calculation unit. The region area acquisition unit is used to acquire the contour areas of the lesion area and the healthy area in the complete leaf outline after repair. The lesion feature error calculation unit calculates the lesion feature foci density error coefficient based on the basic parameter set, the complete parameter set, and the lesion area contour area. The health feature error calculation unit calculates the health feature foci density error coefficient based on the basic parameter set, the complete parameter set, and the healthy area contour area.

6. The intelligent detection system for plant leaves based on data analysis according to claim 5, characterized in that, The lesion severity assessment module includes a coordinate point construction unit and a lesion severity calculation unit. The coordinate point construction unit is used to construct density error coordinate points for each lesion with the density error coefficient of lesion feature lesions as the abscissa and the density error coefficient of healthy feature lesions as the ordinate. The lesion severity calculation unit is used to construct a lesion severity assessment model, substitute the error coefficients corresponding to the coordinate points, and calculate the lesion severity of each lesion.

7. The intelligent detection system for plant leaves based on data analysis according to claim 6, characterized in that, The diffusion path analysis and display module includes a lesion diffusion circle characterization unit, a diffusion path identification unit, and a result sending unit. The lesion diffusion circle characterization unit is used to characterize the lesion diffusion circle of each lesion with the density error coordinate point of each lesion as the center and the corresponding lesion degree as the radius. The diffusion path identification unit is used to determine whether the lesion diffusion circles of adjacent lesions overlap, and construct and expand the lesion diffusion path identifier based on the overlap. The result sending unit is used to obtain the complete lesion diffusion path identifier and send it to the intelligent detection and display terminal.

Images