Garden pest control method based on infrared image

Abstract

The invention relates to the technical field of image processing, in particular to a garden disease and pest control method based on an infrared image, and the method comprises the steps: collecting the infrared image of a garden; the method comprises the following steps: dividing an infrared image into superpixel blocks, and determining a temperature difference influence coefficient and a local temperature difference coefficient of each superpixel block based on a dispersion degree and a change intensity degree of pixel values in each superpixel block; obtaining a temperature characteristic factor of each superpixel block; edge detection is carried out on each super-pixel block, the intersection condition of each edge in each super-pixel block and other super-pixel blocks is analyzed, and boundary feature factors of each super-pixel block are determined; in combination with the temperature characteristic factors and the boundary characteristic factors, determining a pest and disease damage probability index of each super-pixel block; based on the correlation between the disease and pest probability indexes of all the super-pixel blocks, obtaining a disease and pest area in the garden, and performing prevention and treatment on diseases and pests. The accuracy of identifying the pest and disease damage area in the garden is improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method for controlling garden pests and diseases based on infrared images. Background Art

[0002] Infrared imaging technology has many advantages in garden pest control, such as non-contact, rapid detection, precise positioning, reduced pesticide use and improved efficiency. It can detect pest problems early by monitoring plant temperature and physiological state changes, and take effective prevention and control measures in time, thereby improving the intelligent level of garden management, reducing environmental pollution, and promoting sustainable agricultural and horticultural development.

[0003] However, garden areas are often large, and the spatial distribution of pest and disease areas is uneven. Many pests and diseases do not show obvious symptoms in the early stages and cannot cause obvious temperature changes in infrared images. Therefore, how to accurately identify pest and disease areas in gardens is an urgent problem that needs to be solved. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a garden pest control method based on infrared images to solve the existing problems.

[0005] The garden pest control method based on infrared images of the present application adopts the following technical solutions: An embodiment of the present application provides a garden pest control method based on infrared images, the method comprising the following steps: Collect infrared images of the garden; The infrared image is divided into super-pixel blocks, and the temperature difference influence coefficient of each super-pixel block is determined based on the discrete degree and the drastic degree of change of the pixel values ​​in each super-pixel block; the grayscale difference between each super-pixel block and its adjacent super-pixel block is analyzed, and the local temperature difference coefficient of each super-pixel block is determined in combination with the temperature difference influence coefficient; Determine the temperature characteristic factor of each superpixel block by combining the temperature difference influence coefficient and the local temperature difference coefficient; perform edge detection on each superpixel block, obtain the related superpixel blocks of each edge line in each superpixel block, analyze the difference in temperature characteristic factor between each superpixel block and its related superpixel block, and the proportion of the number of pixel points on the corresponding edge line in the related superpixel block, and determine the boundary characteristic factor of each superpixel block; The temperature characteristic factor and the boundary characteristic factor are combined to determine the pest and disease probability index of each superpixel block; based on the correlation between the pest and disease probability indexes of all superpixel blocks, the pest and disease area in the garden is obtained, and the pests and diseases are prevented and controlled.

[0006] In one embodiment, the determination of the temperature difference influence coefficient includes: Calculate the grayscale mean of all pixels in each superpixel block, and calculate the difference between the grayscale value of each pixel in each superpixel block and the grayscale mean, which is recorded as the first difference; Based on the first difference of each pixel point in each superpixel block and the gradient of each pixel point in each superpixel block, the temperature difference influence coefficient of each superpixel block is determined, and the temperature difference influence coefficient is positively correlated with the first difference of each pixel point in each superpixel block, and negatively correlated with the gradient of each pixel point in each superpixel block.

[0007] In one embodiment, the temperature difference influence coefficient is calculated as follows: ; In the formula, is the temperature difference influence coefficient of the sth superpixel block, is the gray value of the i-th pixel in the s-th superpixel block, is the grayscale mean of all pixels in the sth superpixel block, is the gradient size of the i-th pixel in the s-th superpixel block, c is a preset value greater than 0, and n is the number of all pixels in the s-th superpixel block.

[0008] In one embodiment, the determination of the local temperature difference coefficient includes: Calculate the difference between the grayscale mean of each superpixel block and its adjacent superpixel blocks, take the inverse of the temperature difference influence coefficient of the adjacent superpixel blocks as the weight of the difference, weightedly fuse the differences between each superpixel block and all its adjacent superpixel blocks, and determine the local temperature difference coefficient of each superpixel block.

[0009] In one embodiment, the local temperature difference coefficient is calculated as follows: ; In the formula, is the local temperature difference coefficient of the sth superpixel block, is the grayscale mean of the sth superpixel block, is the grayscale mean of the jth superpixel block adjacent to the sth superpixel block, is the temperature difference influence coefficient of the jth superpixel block adjacent to the sth superpixel block, and m is the number of superpixel blocks adjacent to the sth superpixel block.

[0010] In one embodiment, the temperature characteristic factor is the product of the temperature difference influence coefficient of each super pixel block and the local temperature difference coefficient.

[0011] In one embodiment, obtaining the related super pixel block of each edge line in each super pixel block includes: For any edge line in any superpixel block, the related superpixel block of any edge line is a superpixel block in which pixel points on any edge line exist except for the superpixel block.

[0012] In one embodiment, the boundary characteristic factor is calculated as follows: ; In the formula, is the boundary feature factor of the sth superpixel block, u is the number of edge lines in the sth superpixel block, is the number of pixels on the vth edge line in the sth superpixel block, is the number of superpixel blocks related to the vth edge line in the sth superpixel block, is the number of edge pixels on the vth edge line in the rth related superpixel block of the vth edge line, is the absolute value of the difference between the temperature characteristic factors of the sth superpixel block and the rth related superpixel block.

[0013] In one embodiment, the pest probability index is the product of the temperature characteristic factor and the boundary characteristic factor of each super pixel block.

[0014] In one embodiment, obtaining the pest and disease area in the garden includes: Clustering is performed based on the pest and disease probability index of all superpixel blocks, all superpixel blocks are divided into clusters, and the average value of the pest and disease probability index of all superpixel blocks in each cluster is calculated. The pest and disease area is the corresponding area of ​​all superpixel blocks in the cluster with the maximum value of the average value in the garden.

[0015] This application has at least the following beneficial effects: The present application collects infrared images of gardens; divides the infrared images into super-pixel blocks, and determines the temperature difference influence coefficient of each super-pixel block based on the discrete degree of pixel values ​​in each super-pixel block and the severity of the change; the temperature difference influence coefficient reflects the degree of disorder of temperature change in each super-pixel block, reflects the possibility that the corresponding area of ​​the super-pixel block in the garden is a pest and disease area, improves the reliability of pest and disease area identification, analyzes the grayscale difference between each super-pixel block and its adjacent super-pixel blocks, and determines the local temperature difference coefficient of each super-pixel block in combination with the temperature difference influence coefficient; the local temperature difference coefficient reflects the degree of difference in temperature between each super-pixel block and its adjacent super-pixel blocks, and improves the accuracy of pest and disease area identification by comparing the temperature difference between the pest and disease area and the normal area; combines the temperature difference influence coefficient with the local temperature difference coefficient to determine the temperature characteristic factor of each super-pixel block; the temperature characteristic factor is overall The temperature reflects the degree of pest invasion in the corresponding area of ​​each superpixel block in the garden. Edge detection is performed on each superpixel block to obtain the related superpixel blocks of each edge line in each superpixel block. The difference in temperature characteristic factors between each superpixel block and its related superpixel block, as well as the proportion of the number of pixel points on the corresponding edge line in the related superpixel block are analyzed to determine the boundary characteristic factor of each superpixel block; the boundary characteristic factor reflects the pest invasion characteristics of the superpixel block from the edge information in the superpixel block, thereby improving the comprehensiveness of pest area identification; the pest probability index of each superpixel block is determined by combining the temperature characteristic factor and the boundary characteristic factor; based on the correlation between the pest probability indices of all superpixel blocks, the pest area in the garden is obtained, and the pests are prevented and controlled, thereby improving the accuracy of pest area identification and further improving the timeliness of pest control. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A flowchart of the steps of the garden pest control method based on infrared images provided in this application; Figure 2 Flowchart for determination of pest probability index. DETAILED DESCRIPTION

[0018] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of the garden pest control method based on infrared images proposed in the present application in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0020] The specific scheme of the garden pest control method based on infrared images provided by the present application is described in detail below with reference to the accompanying drawings.

[0021] An embodiment of the present application provides a garden pest control method based on infrared images. Specifically, the following garden pest control method based on infrared images is provided. Figure 1 , the method comprises the following steps: Step S001, collecting infrared images of the garden and performing preprocessing.

[0022] This embodiment uses a drone to patrol the garden. The drone is equipped with a 20-megapixel image sensor, can switch between RGB mode and infrared mode, and is equipped with a visual positioning system, which can perform stable flight or hovering at ultra-low altitudes.

[0023] This embodiment enables the drone to hover at a height of 20 meters above the patrol area, with good spatial resolution and image information. The camera maintains a vertical downward posture with a fixed shooting angle to obtain infrared images of the ground garden. The plant infrared image mainly uses the infrared band to capture the thermal radiation on the plant surface.

[0024] The implementer can set the drone hovering position according to the actual situation to facilitate image stitching, and obtain a panoramic infrared image of the entire garden while ensuring the image resolution and information volume as much as possible; The panoramic infrared image is subjected to filtering and noise reduction processing. In order to reduce the probability that the pest area is smoothed during the noise reduction process, the collected panoramic infrared image is processed using an edge-preserving filter. In this embodiment, the panoramic infrared image is subjected to noise reduction processing by non-local mean filtering. The implementer may select other existing feasible edge-preserving filters according to actual conditions, and this embodiment does not limit this.

[0025] The panoramic infrared image is a grayscale image, in which the pixel value of each pixel reflects the thermal radiation of the pixel at the corresponding position in the garden. The larger the pixel value, the higher the temperature of the area corresponding to the pixel.

[0026] Step S002, divide the infrared image into super-pixel blocks, and determine the temperature difference influence coefficient of each super-pixel block based on the discrete degree and the severity of the change of the pixel values ​​in each super-pixel block; analyze the grayscale difference between each super-pixel block and its adjacent super-pixel blocks, and determine the local temperature difference coefficient of each super-pixel block in combination with the temperature difference influence coefficient.

[0027] When plants are attacked by pests and diseases, their physiological metabolism will change, causing the temperature of the local area to rise or fall. For example, after the pathogens invade the plant tissue, the local temperature rises; or the pests and diseases destroy the plant's conductive tissue, affecting the transportation of water and nutrients, causing the local temperature to drop. Infrared images can capture these abnormal temperature changes and identify the pest and disease areas.

[0028] When drones collect infrared images of gardens, the temperature of different areas varies, which is usually shown as local areas with different grayscale value distributions in infrared images. In order to identify pest and disease areas, it is necessary to analyze the image features reflected in these local areas.

[0029] This embodiment uses the SLIC (Simple Linear Iterative Clustering) superpixel segmentation algorithm to perform block processing on the panoramic infrared image, and divides the panoramic infrared image into superpixel blocks. These superpixel blocks correspond to local garden areas with different temperature characteristics, such as normal vegetation areas, environmental areas, and pest and disease vegetation areas. The SLIC superpixel segmentation algorithm is an existing well-known technology, and the implementer can select other existing feasible superpixel segmentation algorithms according to actual conditions, and this embodiment does not limit it here.

[0030] The temperature of the surface of general plants is greatly affected by environmental factors, among which sunlight is the main factor affecting the surface temperature of plants. For normal vegetation or environmental areas, temperature changes are mainly related to meteorological factors, such as light, wind speed and other factors. The temperature of areas affected only by meteorological factors is usually relatively uniform; while the interior of pest and disease areas may present different temperature characteristics: pests and diseases of vegetation usually cause damage to the epidermis of vegetation. Therefore, in the pest and disease area, the water evaporation rate of the damaged plant epidermis is different from that of the undamaged plant epidermis. Therefore, there is usually a certain temperature difference inside the pest and disease area.

[0031] Based on the above analysis, the temperature difference influence coefficient of each superpixel block is calculated, specifically: the grayscale mean of all pixels in each superpixel block is calculated, and the difference between the grayscale value of each pixel in each superpixel block and the grayscale mean is calculated, which is recorded as the first difference. The temperature difference influence coefficient is positively correlated with the first difference of each pixel in each superpixel block, and negatively correlated with the gradient of each pixel in each superpixel block. It should be noted that the difference indicates the degree of difference between two variables, which can be calculated by the absolute value of the difference, the square of the difference, etc. A positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. A negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases.

[0032] In this embodiment, the temperature difference influence coefficient of each super pixel block may be calculated as follows: ; In the formula, is the temperature difference influence coefficient of the sth superpixel block, is the gray value of the i-th pixel in the s-th superpixel block, is the grayscale mean of all pixels in the sth superpixel block, is the gradient of the ith pixel in the sth superpixel block, c is a preset value greater than 0 to avoid the denominator being 0, and n is the number of all pixels in the sth superpixel block; in this embodiment, c=1, and the implementer can set it according to the actual situation, and this embodiment does not limit it here. Recorded as the first difference.

[0033] It should be understood that It reflects the degree of discreteness of the pixel value distribution within the superpixel block. The larger the value, the larger the overall temperature difference within the corresponding superpixel block. In addition, the temperature difference caused by pests and diseases is usually more concentrated, that is, in the superpixel block, the vegetation affected by pests and diseases is compared with the normal plant surface. The area with large temperature difference changes is usually concentrated in the local area where the vegetation epidermis is damaged. Therefore, the internal gradient value is usually lower. Correspondingly, the larger the temperature difference influence coefficient, the greater the possibility that the garden area corresponding to the superpixel block is a pest and disease area.

[0034] However, in the garden area, due to the influence of environmental factors, such as morning dew and frost, large temperature differences may appear in local areas of the infrared image. Therefore, the temperature difference influence coefficient alone cannot accurately identify the pest and disease areas in the garden.

[0035] Pests and diseases often affect the physiological functions of plants, such as transpiration, causing the temperature of the pest-affected area to be different from that of the healthy area. The surface temperature of the affected area may increase or decrease due to obstructed water transport, cell damage or abnormal metabolism. For example, some pests and diseases may cause local tissue necrosis, reduce water evaporation, and make the temperature of these areas relatively high; while the plant defense response caused by some pests and diseases may increase transpiration, resulting in a decrease in temperature.

[0036] In summary, the influence of pests and diseases will not only cause large differences in temperature within the superpixel block, but also cause large differences in temperature between superpixel blocks.

[0037] Based on the above analysis, the local temperature difference coefficient of each superpixel block is calculated. The specific calculation method is: ; In the formula, is the local temperature difference coefficient of the sth superpixel block, is the grayscale mean of the sth superpixel block, is the grayscale mean of the jth superpixel block adjacent to the sth superpixel block, is the temperature difference influence coefficient of the jth superpixel block adjacent to the sth superpixel block, and m is the number of superpixel blocks adjacent to the sth superpixel block.

[0038] It should be understood that through Comparing the grayscale mean difference between each superpixel block and its adjacent superpixel block indicates the overall temperature difference between the two superpixel blocks. In addition, in order to highlight the temperature difference between each superpixel block and its adjacent superpixel block, it is necessary to minimize the interference of superpixel blocks with large internal temperature differences on the grayscale mean calculation process. When the internal temperature difference is large, the average grayscale value is difficult to be used as a parameter for evaluating the normal temperature level of superpixel blocks. Therefore, the inverse of the temperature difference influence coefficient is used as a weight to adjust the proportion of the grayscale mean difference of the corresponding superpixel block in the calculation of the local temperature difference coefficient. The larger the local temperature difference coefficient, the greater the temperature difference between the corresponding superpixel block and its adjacent superpixel block, and the more likely the area of ​​the corresponding superpixel block in the garden is a pest and disease area.

[0039] Step S003, combining the temperature difference influence coefficient and the local temperature difference coefficient to determine the temperature characteristic factor of each superpixel block; performing edge detection on each superpixel block, obtaining the related superpixel blocks of each edge line in each superpixel block, analyzing the difference in temperature characteristic factors between each superpixel block and its related superpixel block, as well as the proportion of the number of pixel points on the corresponding edge line in the related superpixel block, to determine the boundary characteristic factor of each superpixel block.

[0040] Combined with the temperature difference influence coefficient and local temperature difference coefficient of each superpixel block, the temperature characteristic factor of each superpixel block is calculated. The specific calculation method is: ; In the formula, is the temperature characteristic factor of the sth superpixel block, is the temperature difference influence coefficient of the sth superpixel block, is the local temperature difference coefficient of the sth superpixel block.

[0041] Based on the above analysis, it should be understood that the larger the temperature characteristic factor of a superpixel block is, the more obvious the pest and disease characteristics of the corresponding area of ​​the superpixel block in the garden are.

[0042] Abnormal metabolic physiological activities of plant tissues caused by pest and disease invasion areas will also lead to differences in the boundary characteristics of pest and disease areas and normal areas. The boundary characteristics of normal areas are usually smooth, regular and continuous, while the boundaries of pest and disease areas are often irregular, blurred and have large temperature changes. In summary, by combining regional boundary characteristics, the accuracy of pest and disease area identification can be further increased.

[0043] This embodiment uses the Canny operator to obtain edge lines in each superpixel block, wherein the Canny operator is an existing well-known technology, and implementers can select other existing feasible edge detection algorithms according to actual conditions, and this embodiment does not limit this.

[0044] Calculate the boundary feature factor of each superpixel block. The specific calculation method is: ; In the formula, is the boundary feature factor of the sth superpixel block, u is the number of edge lines in the sth superpixel block, is the number of pixels on the vth edge line in the sth superpixel block, taking the vth edge line as an example, obtaining each related superpixel block of the vth edge line, wherein the related superpixel block is a superpixel block having pixels on the vth edge line except the sth superpixel block, is the number of superpixel blocks related to the vth edge line in the sth superpixel block, is the number of edge pixels on the vth edge line in the rth related superpixel block of the vth edge line, is the absolute value of the difference between the temperature characteristic factor of the sth superpixel block and the rth related superpixel block. The same acquisition method as that of the related superpixel blocks of the vth edge line is adopted to obtain the related superpixel blocks of each edge line in each superpixel block.

[0045] It should be understood that if The larger the value is, the more likely the rth related superpixel block is to be a normal area. If the difference in temperature characteristic factors between the sth superpixel block and the rth related superpixel block is greater, the greater the possibility that the sth superpixel block is a pest area. As a weight, the calculation ratio of the absolute value of the difference between the temperature characteristic factors of the rth related super-pixel block and the sth super-pixel block is amplified to improve the accuracy of determining the pest and disease area.

[0046] Step S004, combining the temperature characteristic factor and the boundary characteristic factor to determine the pest and disease probability index of each superpixel block; based on the correlation between the pest and disease probability indexes of all superpixel blocks, obtain the pest and disease area in the garden and prevent and control the pests and diseases.

[0047] Finally, the temperature characteristic factor of each superpixel block is combined with the boundary characteristic factor to obtain the pest probability index of each superpixel block. The specific calculation method is: ; In the formula, is the pest probability index of the sth superpixel block, is the boundary feature factor of the sth superpixel block, is the temperature characteristic factor of the sth superpixel block. The flow chart for determining the probability index of pests and diseases is as follows: Figure 2 shown.

[0048] The pest probability index of all superpixel blocks is used as the input of the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm, and the output is each cluster cluster. The mean of all pest probability indexes in each cluster cluster is calculated, and the cluster cluster corresponding to the maximum mean value is used as the pest cluster cluster, and the corresponding area of ​​all superpixel blocks in the pest cluster cluster in the garden is used as the pest area, so as to complete the identification of the pest area in the garden. Among them, the clustering parameters eps=2 and MinPts=5 of the DBSCAN clustering algorithm in this embodiment are the existing known technology, and the implementer can select other existing feasible clustering algorithms according to the actual situation, and this embodiment does not limit it here.

[0049] Biological control methods are used to treat the pest and disease areas identified in the garden. Specifically, by utilizing the predator-prey relationship between pests and their natural enemies, the population density of pests and diseases can be effectively reduced. According to the types of pests and diseases in the pest and disease areas, corresponding natural enemies are selected for biological control. This can not only reduce the use of chemical pesticides and avoid their potential threats to the environment and human health, but also help restore the natural order of the ecosystem.

[0050] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0051] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0052] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Modifications to the technical solutions recorded in the aforementioned embodiments, or equivalent replacement of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A garden pest control method based on infrared images, characterized in that: The method comprises the following steps: Collect infrared images of the garden; The infrared image is divided into super-pixel blocks, and the temperature difference influence coefficient of each super-pixel block is determined based on the discrete degree and the drastic degree of change of the pixel values ​​in each super-pixel block; the grayscale difference between each super-pixel block and its adjacent super-pixel block is analyzed, and the local temperature difference coefficient of each super-pixel block is determined in combination with the temperature difference influence coefficient; Determine the temperature characteristic factor of each superpixel block by combining the temperature difference influence coefficient and the local temperature difference coefficient; perform edge detection on each superpixel block, obtain the related superpixel blocks of each edge line in each superpixel block, analyze the difference in temperature characteristic factor between each superpixel block and its related superpixel block, and the proportion of the number of pixel points on the corresponding edge line in the related superpixel block, and determine the boundary characteristic factor of each superpixel block; The temperature characteristic factor and the boundary characteristic factor are combined to determine the pest and disease probability index of each superpixel block; based on the correlation between the pest and disease probability indexes of all superpixel blocks, the pest and disease area in the garden is obtained, and the pests and diseases are prevented and controlled.

2. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The determination of the temperature difference influence coefficient includes: Calculate the grayscale mean of all pixels in each superpixel block, and calculate the difference between the grayscale value of each pixel in each superpixel block and the grayscale mean, which is recorded as the first difference; Based on the first difference of each pixel point in each superpixel block and the gradient of each pixel point in each superpixel block, the temperature difference influence coefficient of each superpixel block is determined, and the temperature difference influence coefficient is positively correlated with the first difference of each pixel point in each superpixel block, and negatively correlated with the gradient of each pixel point in each superpixel block.

3. The method for controlling garden pests and diseases based on infrared images according to claim 2, characterized in that: The temperature difference influence coefficient is calculated as follows: ; In the formula, is the temperature difference influence coefficient of the sth superpixel block, is the gray value of the i-th pixel in the s-th superpixel block, is the grayscale mean of all pixels in the sth superpixel block, is the gradient size of the i-th pixel in the s-th superpixel block, c is a preset value greater than 0, and n is the number of all pixels in the s-th superpixel block.

4. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The determination of the local temperature difference coefficient includes: Calculate the difference between the grayscale mean of each superpixel block and its adjacent superpixel blocks, take the inverse of the temperature difference influence coefficient of the adjacent superpixel blocks as the weight of the difference, weightedly fuse the differences between each superpixel block and all its adjacent superpixel blocks, and determine the local temperature difference coefficient of each superpixel block.

5. The method for controlling garden pests and diseases based on infrared images according to claim 4, characterized in that: The local temperature difference coefficient is calculated as follows: ; In the formula, is the local temperature difference coefficient of the sth superpixel block, is the grayscale mean of the sth superpixel block, is the grayscale mean of the jth superpixel block adjacent to the sth superpixel block, is the temperature difference influence coefficient of the jth superpixel block adjacent to the sth superpixel block, and m is the number of superpixel blocks adjacent to the sth superpixel block.

6. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The temperature characteristic factor is the product of the temperature difference influence coefficient of each super pixel block and the local temperature difference coefficient.

7. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The obtaining of the related super pixel blocks of each edge line in each super pixel block includes: For any edge line in any superpixel block, the related superpixel block of any edge line is a superpixel block in which pixel points on any edge line exist except for the superpixel block.

8. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The boundary characteristic factor is calculated as follows: ; In the formula, is the boundary feature factor of the sth superpixel block, u is the number of edge lines in the sth superpixel block, is the number of pixels on the vth edge line in the sth superpixel block, is the number of superpixel blocks related to the vth edge line in the sth superpixel block, is the number of edge pixels on the vth edge line in the rth related superpixel block of the vth edge line, is the absolute value of the difference between the temperature characteristic factors of the sth superpixel block and the rth related superpixel block.

9. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The pest probability index is the product of the temperature characteristic factor and the boundary characteristic factor of each super pixel block.

10. The method for controlling garden pests and diseases based on infrared images according to claim 1, characterized in that: The method of obtaining the pest and disease area in the garden includes: Clustering is performed based on the pest and disease probability index of all superpixel blocks, all superpixel blocks are divided into clusters, and the average value of the pest and disease probability index of all superpixel blocks in each cluster is calculated. The pest and disease area is the corresponding area of ​​all superpixel blocks in the cluster with the maximum value of the average value in the garden.

Images