Insect intelligent monitoring and analysis system based on image analysis

Abstract

The invention discloses an intelligent insect monitoring and analysis system based on image analysis, which relates to the technical field of intelligent image analysis monitoring, and comprises an image acquisition module, a resource database, an insect image analysis module, a crop image analysis module and an insect display and early warning module; however, the existing image analysis insect monitoring system has some defects, such as low recognition accuracy under complex backgrounds, but through efficient image acquisition, preprocessing and analysis, accurate insect species recognition and crop health status assessment and multi-module collaboration are realized to achieve efficient insect and crop recognition, health status assessment and early warning, accurate identification of insect species and crop health status, and provide detailed analysis reports and prevention and control suggestions; through real-time monitoring and data analysis, measures can be taken in time to deal with pests and diseases, thereby improving crop yield and quality, reducing pesticide use, reducing production costs, and realizing green agricultural production.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent image analysis and monitoring, and in particular to an insect intelligent monitoring and analysis system based on image analysis. Background Art

[0002] The existing image analysis insect monitoring system uses computer vision and artificial intelligence technology to capture insect images through cameras and automatically identify and classify them to achieve real-time monitoring and early warning. The system has a wide range of applications in the fields of agriculture and environmental protection, and can effectively improve the efficiency of insect monitoring and reduce labor costs.

[0003] However, existing technologies have some defects, such as low recognition accuracy in complex backgrounds, insufficient adaptability to changes in lighting and weather, and challenges in large-scale data processing and storage; in addition, the similarities in features between different insect species also increase the difficulty of recognition. Summary of the invention

[0004] The purpose of the present invention is to propose an insect intelligent monitoring and analysis system based on image analysis in order to solve the problems of low recognition accuracy in complex backgrounds, insufficient adaptability to lighting and weather changes, and challenges in large-scale data processing and storage; in addition, the feature similarity between different insect species also increases the difficulty of recognition.

[0005] The object of the present invention can be achieved by the following technical solutions: an intelligent insect monitoring and analysis system based on image analysis, comprising an insect image analysis module and a crop image analysis module;

[0006] The insect image analysis module is used to analyze the collected and processed insect images, specifically:

[0007] Step 1: Identify the insects in the insect image through the insect image, and extract the image features through the convolutional neural network model and match the pre-stored image features to calculate the matching feature similarity. If the matching feature similarity is greater than or equal to 98%, output the type of insect and execute step 6; if the matching feature similarity is less than 98%, re-detect the image through the convolutional neural network model and output the secondary matching feature similarity. If the secondary matching feature similarity XSD is greater than or equal to 98%, output the type of insect and execute step 6; if the secondary matching feature similarity is less than 98%, mark the image as not containing insects and execute step 2;

[0008] Step 2: Collect insects by using an infrared camera, and generate an insect thermal image by detecting the infrared radiation emitted by the object, specifically: when the similarity of the collected insect features is less than 98%, the position information of the image collected by the image is analyzed to obtain the insect image coordinates in the insect thermal image; then the insect identification value is obtained by analyzing the insect thermal image, and when the insect identification value belongs to the insect identification value interval, the insect is the insect corresponding to S'; otherwise, step 3 is executed;

[0009] The crop image analysis module analyzes the crop image by using a high-resolution camera. The specific process is as follows:

[0010] Step 3: Use high-resolution camera to take pictures of crop images for identification and obtain similarity;

[0011] Step 4: Use high-resolution photography equipment to identify the crop categories in the collection area, mark the crop categories, and then obtain the insects corresponding to each crop; and also obtain the proportion of healthy leaves and the proportion of diseased leaves by statistically analyzing the pixels of the two categories;

[0012] Step 5: Obtain the final recognition coefficient by performing comprehensive analysis on the insect recognition value, matching feature similarity and similarity;

[0013] Step 6: Obtain the insect species, distinguish and number them, and understand their habits; then, based on the number of insect species in the collection area and the corresponding planting area and area of ​​crops, use the collected insect image area as the sampling site for calculation and analysis to obtain the number of insect species in the corresponding planting area of ​​crops.

[0014] As a preferred embodiment of the present invention, it also includes an image acquisition module, a resource database and an insect display warning module;

[0015] The image acquisition module collects images of crops by using insect photographing equipment, infrared cameras and high-resolution photographing equipment; the image acquisition module is also used to pre-process the images and store the collected and processed images in a resource database;

[0016] The resource database is used to store the required data of the system;

[0017] The insect display warning module generates warning information based on the results of the insect image analysis module and the crop image analysis module and displays it to the user.

[0018] As a preferred embodiment of the present invention, the crop images are photographed by a high-resolution photographing device for identification to obtain similarity, and the specific process is as follows:

[0019] By distinguishing the categories of the captured crop images, capturing the leaf features or other features of the crops and matching them with the leaf features or other features of the preset crop image categories, the corresponding crop image categories are obtained; then, the crop colors in the crop images are distinguished, the images are converted from RGB to HSV color space, the H channel is extracted from the converted HSV image, the hue channel is used to separate green and other colors, and then the threshold range of the hue channel is preset, the pixels whose pixel values ​​of the hue channel are within the threshold range are set to 1, and the other pixels are set to 0, to generate a binary image; then, by setting a target, the pixels in the image are classified into healthy leaves and diseased leaves; the number of cluster centers is set to 2, and the K-means clustering algorithm is applied to cluster the pixel colors in the image into two categories, and two categories of pixels are obtained, representing healthy leaves and diseased leaves respectively; finally, white and black crop images are obtained, and the black pixel areas existing in the white pixel areas or the intersection contours of the white pixel areas and the black pixel areas in the white and black crop images are obtained, wherein the black pixel areas and the intersection contours are compared with the contours of different insects gnawing on crops to calculate the similarity XSDZ.

[0020] As a preferred embodiment of the present invention, the healthy leaf ratio and the diseased leaf ratio are obtained by statistically analyzing the pixels of the two categories, specifically:

[0021] By counting the pixels of the two categories obtained, and by counting the number of pixels in each color category, the proportion of healthy and diseased leaves: the pixel ratio of healthy leaves and diseased leaves is calculated: healthy leaf ratio: number of healthy leaf pixels / total number of pixels; diseased leaf ratio: number of diseased leaf pixels / total number of pixels; the healthy leaf ratio is marked as JB and the diseased leaf ratio is marked as WB.

[0022] As a preferred embodiment of the present invention, the comprehensive analysis is performed by obtaining the insect identification value, matching feature similarity and similarity, specifically:

[0023] By formula Output the final recognition coefficient ZSB, where the matching feature similarity XSD is less than 98%; SBZ=μ has a peak value, SBZ belongs to (0-2.9), (3.1+∞), μ is the mean of the eigenvalue, σ is the standard deviation of the eigenvalue, c1, c2 and c3 are preset weight factors; when the final recognition coefficient ZSB belongs to the preset recognition interval, execute step six.

[0024] As a preferred implementation of the present invention, the area of ​​the captured insect image is used as a sampling area for computational analysis, specifically:

[0025] According to the number of insect species in the collection area and the corresponding crop planting area and area, the area where the insect images were collected was taken as the sampling area, and the formula Output the number of insect species DYZS in the planting area corresponding to the crop; is the number of insect species in a certain collection area, is the average number of insect species in the same crop collection area, M is the total planting area of ​​the corresponding crop, M' is the planting area of ​​the corresponding crop in a certain collection area, and a3 is the preset correction factor.

[0026] As a preferred embodiment of the present invention, the insect display warning module generates warning information based on the results of the insect image analysis module and the crop image analysis module and displays it to the user, specifically:

[0027] By insect species, the number of insects harmful to crops DYZS 害 and beneficial insect population DYZS 益 , and the proportion of healthy leaves is marked as JB and the proportion of diseased leaves is marked as WB, respectively, to calculate the execution warning coefficient one and the warning coefficient two, which are compared with the preset execution thresholds respectively. If the warning coefficient one is greater than or equal to the preset execution threshold, the operation command one is executed; when the operation command one is executed, a detailed text report is generated, including insect species and quantity statistics, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a warning prompt is popped up in the email and system; if the warning coefficient one is greater than or equal to the preset execution threshold, the operation command two is executed; when the operation command two is executed, a detailed text report is generated, including crop health status assessment, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a short message is sent.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. The present invention realizes accurate insect species identification and crop health status assessment through efficient image acquisition, preprocessing and analysis through the insect intelligent monitoring and analysis system based on image analysis. The system can automatically switch the acquisition equipment under different lighting conditions, and combine the convolutional neural network model to perform multiple feature matching to improve the recognition accuracy, and finally generate detailed early warning reports and prevention and control recommendations. Through this intelligent and automated monitoring method, potential crop diseases and pests can be discovered and responded to in a timely manner, significantly improving the efficiency of agricultural production and the health level of crops.

[0030] 2. The present invention uses an intelligent insect monitoring and analysis system based on image analysis to achieve efficient insect and crop identification, health status assessment and early warning through multi-module collaboration. The system uses advanced image processing and convolutional neural network technology to accurately identify insect species and crop health status, and provide detailed analysis reports and prevention and control recommendations. Through real-time monitoring and data analysis, timely measures can be taken to deal with pests and diseases, thereby increasing crop yield and quality, reducing pesticide use, reducing production costs, and achieving green agricultural production.

[0031] 3. The present invention provides accurate early warning information through the insect display early warning module through the intelligent insect monitoring and analysis system, combined with the results of insect and crop image analysis. When the number of harmful insects exceeds the threshold or the health of crops deteriorates, the system will automatically generate a detailed report containing insect species, quantity statistics, potential risk analysis and prevention and control suggestions, and promptly notify the user through emails, text messages and system pop-ups. This system not only improves the timeliness and accuracy of crop disease and insect pest control, but also realizes intelligent and precise agricultural management. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0033] Figure 1 It is a principle block diagram of the present invention. DETAILED DESCRIPTION

[0034] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] It should be understood that the terms "include" and "comprising" used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0036] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms of "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations.

[0037] See also Figure 1 As shown, an insect intelligent monitoring and analysis system based on image analysis includes an image acquisition module, a resource database, an insect image analysis module, a crop image analysis module and an insect display and warning module;

[0038] The image acquisition module uses insect photographing equipment and infrared cameras to collect insect images; it also uses high-resolution photographing equipment to collect images of crops; it selects and analyzes the collected image equipment, and collects the ambient brightness of the scene through the ambient light sensing equipment, and marks it as L, and compares it with the preset ambient brightness threshold L'. When L is less than or equal to L', the insect image is collected through the infrared camera; when L is greater than L', the insect image is collected through the insect photographing equipment;

[0039] Preprocess the insect images and crop images by first reading the image data from the file and loading it as a grayscale image, then reducing the image noise by using a Gaussian filter, and finally enhancing the image contrast by using histogram equalization; store the collected and processed images in a resource database;

[0040] The insect image analysis module analyzes insect images. The specific steps are as follows:

[0041] Step 1: Identify insects in insect images through insect images; when the insect images are collected by insect photographing equipment, convert the collected insect images to the same ratio according to the image resolution and the ratio of the image collected by the insect photographing equipment to the collection environment, and split the insect images into several images, and each image has the same resolution as the preset collected insect image in the resource database, and then compare each image with the preset collected insect image; first obtain each image for detection, and extract the features of each image by the convolutional neural network model; similarly, obtain the resource database in advance through the convolutional neural network model to extract the image features of different insects and different angles, and match the extracted features of each image with the pre-stored image features to calculate the feature similarity, and calculate the feature similarity by the formula Output the matching feature similarity XSD, where X and X' are respectively the total value of the coordinates of the spatial pixel area corresponding to the insect feature and the preset total value of the coordinates of the spatial pixel area corresponding to the insect feature, and Z and Z' are respectively the number of insect feature mark points identified and the preset number of insect feature mark points; a1 and a2 are both preset weight correction factors, and their sizes are customized by technicians in this field, such as 0.6 and 0.4; if the matching feature similarity XSD is greater than or equal to 98%, the type of insect is output, and step six is ​​executed; if the matching feature similarity XSD is less than 98%, the image is re-detected through the convolutional neural network model, and the second matching feature similarity is output. If the second matching feature similarity XSD is greater than or equal to 98%, the type of insect is output, and step six is ​​executed; if the second matching feature similarity is less than 98%, the image is marked as having no insects, and step two is executed;

[0042] Step 2: Collect insects through infrared cameras, and generate insect thermal images by detecting infrared radiation emitted by the object. Then, when the similarity of the collected insect features is less than 98%, the position information of the image collected by the image is obtained. The infrared camera collects the position information, and obtains the position coordinates of the insect collection image where the collected insect features are located through the position information. The insect thermal image of the infrared camera and the insect image of the professional insect shooting equipment are adjusted to the same frame ratio, and finally the insect image coordinates of the insect thermal image are obtained. Then, the insect thermal image is analyzed, specifically as follows:

[0043] Then, based on the convolutional neural network, the body shape, wing structure and activity state of the insect are captured and matched with the characteristic data of the limb structure and activity state of different types of insects preset in the resource database; the body shape of the insect is obtained by marking the resolution of the insect thermal image to obtain the ratio between the pixels of the insect thermal image and the reality when the infrared camera is shot at one times, and then the number of pixels occupied by the insect in the insect thermal image is obtained, and the number of pixels occupied is the body shape of the insect; the folding methods of the wing structure include elytra covering, overlapping folding, longitudinal folding and transverse folding, etc., and the folding method characteristics of the wing structure of the insect in the image are captured by the convolutional neural network to match and obtain the overlap; the activity state is obtained by obtaining the insect habits and divided into nocturnal activity (20:00-4:00), diurnal activity (5:00-19:00) and diurnal activity (0:00-24:00); the activity state is assigned a value, if the activity state is completely consistent with the corresponding insect habit, it is assigned a value of y1; otherwise, it is assigned a value of y2; and y1=1, y2=0; then through the formula Output the insect identification value SBZ, S is the number of pixels occupied in the insect thermal image, u=1 or 2, S' is the number of pixels occupied by the corresponding insect, CHD is the coincidence degree obtained by matching the insect wing structure, b1, b2 and b3 are corresponding preset correction factors, and their sizes are customized by technicians in this field; and by setting the insect identification value interval, the insect identification value interval is [2.9-3.1]. When the insect identification value SBZ belongs to the insect identification value interval [2.9-3.1], the insect is the insect corresponding to S'; otherwise, execute step three;

[0044] The crop image analysis module analyzes crop images by using high-resolution photography equipment; the specific process is as follows:

[0045] Step 3: Use high-resolution camera equipment to take crop images for identification. First, classify the crop images, capture the leaf features or other features of the crops, and match them with the preset leaf features or other features of the crop image category to obtain the corresponding crop image category; then, distinguish the color of the crops in the crop image, convert the image from RGB to HSV color space, extract the H (hue) channel from the converted HSV image, separate the hue channel into green (healthy leaves) and other colors (disease parts), and then set the threshold range of the hue channel (such as green: 30-90) to set the pixels whose pixel values ​​are within the threshold range. The pixel is set to 1 (white) and the other pixels are set to 0 (black) to generate a binary image; then, by setting a target, the pixels in the image are classified into healthy leaves and diseased leaves; the number of cluster centers is set to 2 (healthy leaves and diseased leaves), and the K-means clustering algorithm is applied to cluster the pixel colors in the image into two categories, and two categories of pixels are obtained, representing healthy leaves and diseased leaves respectively; finally, white and black crop images are obtained, and the black pixel area existing in the white pixel area or the intersection contour of the white pixel area and the black pixel area in the white and black crop images is obtained, wherein the black pixel area and the intersection contour are compared with the contours of different insects eating crops to calculate the similarity XSDZ;

[0046] Step 4: Use high-resolution photography equipment to identify the crop categories in the collection area, mark the crop categories, w1, w2, w3, ..., wp; p = 1, 2, ..., o; then obtain the insects corresponding to each crop; also count the pixels (white) and (black) of the two categories obtained, and count the number of pixels in each color category, the healthy and diseased ratio: calculate the pixel ratio of healthy leaves and diseased leaves: healthy leaf ratio: number of healthy leaf pixels / total number of pixels; diseased leaf ratio: number of diseased leaf pixels / total number of pixels; mark the healthy leaf ratio as JB and the diseased leaf ratio as WB;

[0047] Step 5: Perform comprehensive analysis by obtaining the insect identification value SBZ, matching feature similarity XSD and similarity XSDZ, and use the formula Output the final recognition coefficient ZSB, where the matching feature similarity XSD is less than 98%; SBZ=μ has a peak value, SBZ belongs to (0-2.9), (3.1+∞), μ is the mean of the feature value, σ is the standard deviation of the feature value, c1, c2 and c3 are preset weight factors, and their sizes are customized by technicians in this field; when the final recognition coefficient ZSB belongs to the preset recognition interval, execute step 6;

[0048] Step 6: By obtaining the insect species, the insect species are distinguished and numbered as Kn, n = 1, 2, ..., 2ρ, and according to the insect habits, the insects that are beneficial to crops are numbered as K2, K4, K6, ..., K (2ρ), and the insects that are harmful to crops are numbered as K1, K3, K5, ..., K (2ρ-1); then, according to the number of insect species in the collection area and the corresponding crop planting area and area, the area where the insect images are collected is the sampling area, and the formula is used Output the number of insect species DYZS in the planting area corresponding to the crop; is the number of insect species in a certain collection area, is the average number of insect species in the same crop collection area, M is the total planting area of ​​the corresponding crop, M' is the planting area of ​​the corresponding crop in a certain collection area, and a3 is a preset correction factor, the size of which is customized by technicians in this field;

[0049] The insect display warning module generates warning information based on the results of the insect image analysis module and the crop image analysis module and displays it to the user. Specifically:

[0050] By insect species, the number of insects harmful to crops DYZS 害 and beneficial insect population DYZS 益 , and the early warning coefficient will be implemented = the number of beneficial insects DYZS 益 / Number of harmful insects DYZS 害 , compare the early warning coefficient 1 with the preset execution threshold value, and if the early warning coefficient 1 is greater than or equal to the preset execution threshold value, execute the operation command 1; when the operation command 1 is executed, a detailed text report is generated including insect species and quantity statistics, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a warning prompt pops up in the email and system;

[0051] Then, the proportion of healthy leaves is marked as JB and the proportion of diseased leaves is marked as WB; and the warning coefficient 2 is executed = the proportion of healthy leaves is marked as JB / the proportion of diseased leaves is marked as WB, and the warning coefficient 2 is compared with the preset execution threshold. If the warning coefficient 1 is greater than or equal to the preset execution threshold, the operation command 2 is executed; when the operation command 2 is executed, a detailed text report is generated including the crop health status assessment, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a text message is sent and a warning prompt pops up in the system;

[0052] If operation command 1 and operation command 2 are executed at the same time, warning information including insect species, quantity, distribution location, harm, and suggestions will be generated, and the alarm mechanism will be triggered, including sending text messages, emails, and popping up warning prompts in the system.

[0053] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent insect monitoring and analysis system based on image analysis, comprising an insect image analysis module and a crop image analysis module, characterized in that: The insect image analysis module is used to analyze the collected and processed insect images, specifically: Step 1: Identify the insects in the insect image, extract image features, match them with pre-stored image features and calculate the matching feature similarity. If the matching feature similarity is greater than or equal to 98%, output the type of insect and execute step 6; if the matching feature similarity is less than 98%, re-detect twice and output the matching feature similarity of the two detections. If the two matching feature similarities XSD is greater than or equal to 98%, output the type of insect and execute step 6; if the two matching feature similarities are less than 98%, mark the image as having no insects and execute step 2; Step 2: Collect insects and generate insect thermal images. When the similarity of the collected insect features is less than 98%, analyze the location information of the collected insect images to obtain the insect image coordinates in the insect thermal image; based on the convolutional neural network, capture the insect's body shape, wing structure and activity state and match and compare with the characteristic data of different types of insect limb structure and activity state preset in the resource database; the insect body shape is marked by marking the resolution of the insect thermal image to obtain the ratio between the insect thermal image pixels and the reality when the infrared camera is shot at one times, and then obtain the number of pixels occupied by the insect in the insect thermal image, and the number of pixels occupied is the insect's body shape; the folding methods of the wing structure include elytra covering, overlapping folding, longitudinal folding and transverse folding, and the folding method characteristics of the insect's wing structure in the image are captured by the convolutional neural network to match and obtain the overlap; The activity state is obtained by obtaining the insect habits and is divided into nocturnal activity, diurnal activity and diurnal activity. The value is assigned according to the activity state. If the activity state is completely consistent with the corresponding insect habits, the value is assigned to y1. If not, assign the value to y2; And y1=1, y2=0; then through the formula Output the insect identification value SBZ, S is the number of pixels occupied by the insect thermal image, u=1 or 2, S' is the number of pixels occupied by the corresponding insect, CHD is the coincidence degree obtained by matching the insect wing structure, b1, b2 and b3 are all corresponding preset correction factors, when the insect identification value belongs to the insect identification value interval, the insect is the insect corresponding to S'; otherwise, execute step three; The crop image analysis module analyzes the crop image, and the specific process is as follows: Step 3: Take images of crops and identify them to obtain similarity; Step 4: Use high-resolution photography equipment to identify the crop categories in the collection area, mark the crop categories, and then obtain the insects corresponding to each crop; and also obtain the proportion of healthy leaves and the proportion of diseased leaves by statistically analyzing the pixels of the two categories; Step 5: The final recognition coefficient is obtained by comprehensive analysis of the insect recognition value, matching feature similarity and similarity; Step 6: Obtain the insect species and distinguish and number the insect species; then, based on the number of insect species in the collection area and the corresponding crop planting area and area, analyze the collected insect image area as the sampling site to obtain the number of insect species in the corresponding crop planting area.

2. The insect intelligent monitoring and analysis system based on image analysis according to claim 1 is characterized in that: It also includes an image acquisition module, a resource database, and an insect display and warning module; The image acquisition module collects images of crops by using insect photographing equipment, infrared cameras and high-resolution photographing equipment; The image acquisition module is also used to pre-process the image and store the acquired and processed image in the resource database; The resource database is used to store the required data of the system; The insect display warning module generates warning information based on the results of the insect image analysis module and the crop image analysis module and displays it to the user.

3. The insect intelligent monitoring and analysis system based on image analysis according to claim 1 is characterized in that: The crop images are photographed by high-resolution photographing equipment to obtain similarity through identification, and the specific process is as follows: By classifying the captured crop images, capturing the leaf features or other features of the crops and matching them with the leaf features or other features of the preset crop image categories, the corresponding crop image categories are obtained; the crop colors in the crop images are distinguished, the images are converted from RGB to HSV color space, the H channel is extracted from the converted HSV image, the hue channel is used to separate green and other colors, and then the threshold range of the hue channel is preset, the pixels whose pixel values ​​of the hue channel are within the threshold range are set to 1, and the other pixels are set to 0, so as to generate a binary image; Then, by setting a target, the pixels in the image are classified into healthy leaves and diseased leaves; the number of cluster centers is set to 2, and the K-means clustering algorithm is applied to cluster the pixel colors in the image into two categories, and two categories of pixels are obtained, representing healthy leaves and diseased leaves respectively; finally, white and black crop images are obtained, and the black pixel areas existing in the white pixel areas in the white and black crop images or the intersection contours of the white pixel areas and the black pixel areas are obtained, wherein the black pixel areas and the intersection contours are compared with the contours of different insects eating crops to calculate the similarity XSDZ.

4. The insect intelligent monitoring and analysis system based on image analysis according to claim 3 is characterized in that: The healthy leaf ratio and the diseased leaf ratio are obtained by statistically analyzing the pixels of the two categories, specifically: By counting the pixels of the two obtained categories, the healthy and diseased proportions were calculated by counting the number of pixels in each color category: the pixel ratio of healthy leaves and diseased leaves was calculated: healthy leaf proportion: number of healthy leaf pixels / total number of pixels; Proportion of diseased leaves: number of diseased leaf pixels / total number of pixels; The proportion of healthy leaves was marked as JB, and the proportion of diseased leaves was marked as WB.

5. The insect intelligent monitoring and analysis system based on image analysis according to claim 4 is characterized in that: The comprehensive analysis is performed by obtaining the insect identification value, matching feature similarity and similarity, specifically: By formula Output the final recognition coefficient ZSB, where the matching feature similarity XSD is less than 98%; SBZ is the insect recognition value, and SBZ is expressed as the mean μ of the feature value, and there is a peak value, SBZ belongs to (0-2.9), (3.1+∞), μ is the mean of the insect recognition value, σ is the standard deviation of the insect recognition value, c1, c2 and c3 are preset weight factors; when the final recognition coefficient ZSB belongs to the preset recognition interval, execute step six.

6. The insect intelligent monitoring and analysis system based on image analysis according to claim 5, characterized in that: The area where the insect image is collected is used as the sampling area for analysis, specifically: According to the number of insect species in the collection area and the corresponding crop planting area and area, the area where the insect images were collected was taken as the sampling area, and the formula Output the number of insect species DYZS in the planting area corresponding to the crop; is the number of insect species in a certain collection area, is the average number of insect species in the same crop collection area, M is the total planting area of ​​the corresponding crop, M' is the planting area of ​​the corresponding crop in a certain collection area, and a3 is the preset correction factor.

7. The insect intelligent monitoring and analysis system based on image analysis according to claim 2, characterized in that: The insect display warning module generates warning information based on the results of the insect image analysis module and the crop image analysis module and displays it to the user, specifically: By insect species, the number of insects harmful to crops DYZS 害 and beneficial insect population DYZS 益 , and the proportion of healthy leaves is marked as JB and the proportion of diseased leaves is marked as WB to calculate the execution warning coefficient 1 and the warning coefficient 2 respectively, and compare them with the preset execution threshold value respectively. If the warning coefficient 2 is greater than or equal to the preset execution threshold value, the operation command 1 is executed; when the operation command 1 is executed, a detailed text report is generated, including insect species and quantity statistics, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a warning prompt pops up in the email and system; If the warning coefficient 1 is greater than or equal to the preset execution threshold, the operation command 2 is executed; when the operation command 2 is executed, a detailed text report is generated, including crop health status assessment, potential risk analysis and recommended prevention and control measures, and the alarm mechanism is triggered, and a text message is sent and a warning prompt pops up.

Images