Portable corn rust disease detection device and method

Abstract

The invention discloses a portable corn rust disease detection device and method, and relates to the field of corn rust disease detection. The image acquisition module acquires information data of target corn leaves; the main control module performs corn rust disease detection according to the spectral information based on a corn rust detection model, and determines a pixel point detection result; the pixel point detection result is a healthy point, an early disease point or a disease point; the corn rust detection model adopts a support vector machine classification model; performing pseudo-color enhancement processing on the image information by adopting a pseudo-color technology to obtain processed image information; segmenting the processed image information based on a YOLOv5 segmentation model, and determining a target leaf disease rate according to the segmented image information; and determining an overall leaf detection result corresponding to the corn rust disease according to the target leaf disease rate. The method is easy and convenient to operate, and the detection efficiency of the corn rust disease can be improved.

Description

Technical Field

[0001] The present application relates to the field of corn rust detection, and in particular to a portable corn rust disease detection device and method. Background Art

[0002] Corn rust is a devastating fungal disease whose spores can be spread hundreds or even thousands of kilometers away through air currents, rapidly spreading to large areas of corn planting. Once infected, the pathogen forms dense piles of rust-colored spores on corn leaves, seriously affecting photosynthesis, leading to poor plant growth, shriveled kernels, and ultimately causing a serious decline in corn quality and yield. Therefore, early and accurate detection of corn rust is crucial for judging the development trend of the disease, taking effective prevention and control measures, and reducing economic losses.

[0003] Traditional corn rust detection mainly relies on agricultural experts or experienced farmers to judge the disease by observing the lesions on corn leaves with the naked eye. This judgment method is highly subjective and difficult to identify early symptoms. Naked eye observation is prone to missed diagnosis or misdiagnosis, which affects the accuracy of detection. For large-scale corn planting areas, relying on manual plant-by-plant detection is time-consuming and laborious, and cannot meet the needs of rapid and efficient detection, often resulting in delayed disease discovery and missed best prevention and control opportunities. Although there are some portable detectors on the market, they can often only make a diagnosis when corn rust symptoms are obvious, and cannot effectively detect the disease in the early stage, missing the best prevention and control opportunities. Moreover, these instruments are mainly based on point measurement, which can only reflect the status of the sampling point, and cannot fully reflect the occurrence of the disease on the entire leaf or plant. In addition, these instruments also have problems such as low detection sensitivity, complex operation, and high cost, which are difficult to be widely used in actual production and cannot meet the needs of farmers.

[0004] In summary, the development of a portable point-to-surface fusion corn rust detection device that is easy to operate, fast and accurate in detection, and low in cost is of great significance to the corn planting industry and even food security. Summary of the invention

[0005] The purpose of this application is to provide a portable corn rust disease detection device and method, which is easy to operate and can improve the detection efficiency of corn rust disease.

[0006] To achieve the above objectives, this application provides the following solutions.

[0007] In the first aspect, the present application provides a portable corn rust disease detection device, comprising: an image acquisition module and a main control module; the image acquisition module is connected to the main control module; the image acquisition module is used to collect information data of corn target leaves; the information data includes: image information and spectral information; the main control module is used to: based on a corn rust detection model, perform corn rust disease detection according to the spectral information, and determine the pixel point detection result; the pixel point detection result is a healthy point, an early onset point or a diseased point; wherein the healthy point, the early onset point and the diseased point have different pixel gray value intervals; the corn rust detection model adopts a support vector machine classification model; the image information is pseudo-color enhanced by pseudo-color technology to obtain processed image information; the pseudo-color technology is used to assign corresponding color identification to the image information according to the pixel point detection result; the processed image information is segmented based on the YOLOv5 segmentation model, and the target leaf disease rate is determined according to the segmented image information; the overall leaf detection result corresponding to the corn rust disease is determined according to the target leaf disease rate.

[0008] In a second aspect, the present application provides a portable corn rust disease detection method, which is implemented by the portable corn rust disease detection device described above; the portable corn rust disease detection method comprises: obtaining information data of target corn leaves; the information data comprises: image information and spectral information; based on a corn rust detection model, performing corn rust disease detection according to the spectral information, and determining a pixel point detection result; the pixel point detection result is a healthy point, an early onset point or a diseased point; wherein the grayscale value interval range of the pixel points corresponding to the healthy point, the early onset point and the diseased point is different; the corn rust detection model adopts a support vector machine classification model; pseudo-color technology is used to perform pseudo-color enhancement processing on the image information to obtain processed image information; the pseudo-color technology is used to assign a corresponding color identification to the image information according to the pixel point detection result; the processed image information is segmented based on the YOLOv5 segmentation model, and the target leaf disease rate is determined according to the segmented image information; the overall leaf detection result corresponding to the corn rust disease is determined according to the target leaf disease rate.

[0009] According to the specific embodiments provided in this application, this application has the following technical effects.

[0010] The present application provides a portable corn rust disease detection device and method, the device includes an image acquisition module and a main control module. The image acquisition module collects information data of the target corn leaves, and the main control module analyzes and processes the information data based on the corn rust detection model to determine the pixel detection results, and then segmentation is performed based on the YOLOv5 segmentation model to finally determine the overall leaf detection results. Since the present application can determine the detection results based on the image acquisition module and the main control module, it has the advantage of simple operation. In addition, the present application can more comprehensively characterize the disease characteristics and improve the accuracy of detection through the measurement of the point-surface fusion of pixel points and the overall leaf, and can monitor the spread of the disease and improve the accuracy of early diagnosis. It can be seen that the present application is easy to operate and can improve the detection efficiency of corn rust diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be 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 paying creative work.

[0012] Figure 1 This is a diagram of the internal relationship structure of a portable corn rust disease detection device.

[0013] Figure 2 This is a schematic diagram of the overall structure of a portable corn rust disease detection device.

[0014] Figure 3 This is the overall rear view of the portable corn rust disease detection device.

[0015] Figure 4 This is a cross-sectional view of a portable corn rust disease detection device.

[0016] Figure 5 This is the working flow chart of the portable corn rust disease detection device.

[0017] Figure 6 Build a flow chart for the corn rust detection model.

[0018] Figure 7 This is a schematic diagram of the appearance and structure of a portable corn rust disease detection device.

[0019] Figure 8 Schematic diagram of the process of portable corn rust disease detection method.

[0020] Figure numerals: spectral camera-1, visible light camera-2, ranging module-3, embedded main control board-4, touch screen-5, wireless communication positioning module-6, battery-7, handle-8, GPIO button-9 and power button-10. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0023] In an exemplary embodiment, Figure 1 As shown, a portable corn rust disease detection device is provided.

[0024] The portable corn rust disease detection device includes: an image acquisition module and a main control module. The image acquisition module is connected to the main control module; the image acquisition module is used to collect information data of corn target leaves; the information data includes: image information and spectral information.

[0025] The main control module is used to detect corn rust disease based on the corn rust detection model and spectrum information, and determine the pixel detection results; the pixel detection results are healthy points, early onset points or diseased points; the grayscale value intervals of the pixels corresponding to healthy points, early onset points and diseased points are different; the grayscale value of the pixels corresponding to healthy points is greater than 0.8; the grayscale value of the pixels corresponding to early onset points is between 0.7 and 0.8; the grayscale value of the pixels corresponding to diseased points is less than 0.7. The corn rust detection model adopts the support vector machine classification model.

[0026] The main control module is used to perform pseudo-color enhancement processing on the image information using pseudo-color technology to obtain processed image information; the pseudo-color technology is used to give corresponding color identification to the image information according to the pixel detection result.

[0027] The main control module is used to segment the processed image information based on the YOLOv5 segmentation model, and determine the target leaf disease rate according to the segmented image information.

[0028] The main control module is also used to determine the overall leaf detection result corresponding to the corn rust disease according to the target leaf disease rate.

[0029] In one embodiment, the portable corn rust disease detection device further includes: a wireless communication positioning module and a terminal device; the wireless communication positioning module is connected to the terminal device.

[0030] Among them, the wireless communication positioning module in the communication positioning module includes a GPS positioning module and a wireless transmission module; the wireless communication positioning module can also be a 4G communication positioning sensor; the GPS positioning module and the wireless transmission module are connected; and the wireless transmission module is connected to the terminal device.

[0031] The GPS positioning module is used to obtain geographic location information; geographic location information is the location information corresponding to when the information data is obtained.

[0032] The wireless transmission module is used to transmit the geographical location information and the overall detection results of the blade to the terminal device.

[0033] The portable corn rust disease detection device also includes: a distance measuring module; the distance measuring module is connected to the main control module; the distance measuring module is used to emit laser, measure the distance between the laser emission point and the corn leaf and transmit it to the main control module.

[0034] The portable corn rust disease detection device also includes: a power module; the power module uses a battery with a capacity of 5000mah; the power module is used to provide power.

[0035] The portable corn rust disease detection device also includes: a display module; the display module uses a 4.3-inch LCD touch screen and is connected to the main control module via a MIPI interface; the display module is used to display the detection results.

[0036] The main control module includes an embedded core main control board, which is firmly fixed inside the device by screws as the core control unit of the entire device. The acquired information data is input into the preset corn rust disease judgment model (corn rust detection model) and the analysis results of the characteristic band combination are output, that is, the pixel point detection results.

[0037] The image acquisition module is located on the front panel of the device and consists of a visible light camera and a spectral camera. They transmit data to the main control module through the USB2.0 interface. The visible light camera is responsible for acquiring image information on the leaf surface, while the spectral camera is responsible for acquiring the spectral characteristics of the leaf and obtaining spectral information. The two work together to provide a data basis for early identification of corn rust.

[0038] In order to keep the distance consistent when detecting the target blade and avoid the influence of distance on the detection result, a distance measurement module is installed in the front panel area of ​​the device, which can emit laser and read the distance value autonomously to make the detection result more accurate. The distance measurement module can use a distance measurement sensor and communicate with the main control module through the TTL to USB interface to realize the display of the real-time distance between the device and the target object.

[0039] The power module includes a 5000mah removable lithium battery and a voltage conversion board installed under the device. The power module provides sufficient power for the device.

[0040] In order to realize the real-time data transmission and positioning function, the top of the device is connected to the wireless communication positioning module through the USB interface, so as to realize the real-time data transmission and positioning function. And the antenna is external to facilitate signal reception.

[0041] In order to improve the user experience and protect the internal components, a shell is installed on the outside of the device. A 4.3-inch LCD touch screen is installed at the center of the top inner wall of the shell through the MIPI interface of the main control module as a display module, which is convenient for users to operate and view the test results in real time.

[0042] The upper part of the device is connected to the lithium battery at the lower part through a handle, ensuring the compactness and portability of the structure; a GPIO button is installed on the handle to facilitate users to save data, improving the convenience and efficiency of use.

[0043] Combination Figure 2 , Figure 3 , Figure 4 and Figure 7 The structure of the device mentioned in this application is described. Figure 2 For the overall structure, Figure 3 For the rear view, Figure 4 The upper part is a cross-sectional view, revealing the layout of the internal components.

[0044] The image acquisition module is integrated on the front of the device, which consists of a spectral camera 1 and a visible light camera 2. Both spectral camera 1 and visible light camera 2 communicate with the main control module through the USB2.0 interface to achieve high-speed data transmission. The two cameras work together to obtain multispectral images (spectral information) and high-resolution visible light images (image information) of corn leaves, and transmit the collected information data to the main control module in real time, laying the foundation for subsequent image processing and result analysis.

[0045] The distance measuring module 3 is also placed on the front panel of the device. The distance measuring module 3 can autonomously emit a laser to project onto the target leaf, thereby realizing real-time acquisition of the distance between the device and the target leaf. The TTL interface of the distance measuring module 3 is converted into a USB interface to communicate with the main control module, and the real-time acquired distance information is displayed on the corn rust detection software interface on the display module for easy user operation.

[0046] The 4.3-inch LCD touch screen 5 on the back of the device serves as a display module and is connected to the main control module through the MIPI interface. It not only provides users with a convenient and intuitive operation interface, but also presents the test results in real time, realizing the visualization of human-computer interaction.

[0047] The main control module is located inside the touch screen 5, with the embedded main control board 4 as the core, and undertakes the key tasks of image data processing, algorithm execution and result output, and is the calculation and decision-making center of the entire device.

[0048] The power module is a 5000mAh large-capacity removable lithium battery 7, which provides stable and reliable power support for the device, ensuring its efficient operation in complex environments such as fields. A power button 10 is provided in the center of the back of the device, which allows users to easily control the start and stop of the device. The battery 7 is designed to support 4 to 5 hours of continuous working time. For situations that require long-term operation, the battery can be replaced at any time to ensure continuity and flexibility of work.

[0049] The power module is tightly connected to the upper part of the device through a handle 8, which adopts an ergonomic hand-held design, which is convenient for users to hold and operate. A GPIO button 9 is provided at the upper end of the handle near the device body, and users can easily realize the data storage function by lightly pressing this button. This design not only improves the convenience of operation, but also ensures the efficiency and accuracy of data recording.

[0050] The wireless communication positioning module 6 in the communication positioning module is located at the top of the device, and the wireless communication positioning module 6 is mounted through the USB interface. The module integrates the GPS positioning module and the wireless transmission module, and has the functions of positioning and wireless transmission. It can not only obtain the geographical location information when collecting data in real time, but also transmit the detection data (such as the overall detection results of the leaves) to the cloud or other terminal devices through the wireless network to realize remote sharing and analysis of data. Users can realize the synchronous storage and cloud upload functions of GPS information, thereby ensuring accurate and timely prevention and control measures for the diseased areas. This function helps to improve the efficiency and effectiveness of disease management.

[0051] Through the above structural design, the present application realizes early, rapid and accurate detection of corn rust, providing strong technical support for agricultural production. The working principle of the present application is simple and efficient, which fully reflects its portability and ease of use.

[0052] like Figure 5 As shown in the flow chart, the specific steps are as follows.

[0053] Step 1. When the user conducts corn rust detection in the field, press the power button to start the device and the corn rust detection software interface, turn on the spectral camera and the visible light camera at the same time through the interface, and align the device vertically with the target to be tested. The visible light camera and the spectral camera will synchronously capture the image information and spectral information of the target.

[0054] Step 2: The user can start the distance measurement module and emit laser through the corn rust detection software interface to obtain the target distance in real time. To ensure the measurement accuracy, whiteboard calibration is required. During the calibration process, the user needs to align the device vertically with the standard whiteboard. The user can adjust the exposure time of the spectral camera through the interface slider to obtain a clear image. After confirming that the image quality meets the requirements, click the "Save" button, and the system will save the whiteboard image, spectral data, and the current distance measurement value and exposure time to provide basic data for subsequent analysis.

[0055] Step 3: To ensure the reliability of the experimental results, when collecting images of the target leaves, the vertical distance between the device and the leaves must be strictly controlled, and the exposure time of the spectral camera must be kept the same as that during whiteboard calibration. At the same time, the leaves to be tested should be flat and flat to avoid obstruction, so as to obtain clear and complete leaf images and achieve full coverage of the target area.

[0056] Step 4: Open the GPS positioning module through the corn rust detection software interface to confirm the current time, longitude and latitude to facilitate subsequent precise prevention and control of the diseased area.

[0057] Step 5. When the user confirms that the image is clear and the distance is appropriate, press the save button on the detection software interface or press the GPIO button on the device handle, and the current image (image information), spectral data (spectral information) and GPS information will be saved.

[0058] Step 6: Press the result analysis button on the corn rust detection software interface, and the main control module will immediately start the preset analysis model (corn rust detection model) to process and calculate the collected data.

[0059] The steps for establishing the corn rust detection model are as follows: Figure 6 As shown, the specific steps are as follows.

[0060] S61, data set construction, select the corn rust data collected in advance, mainly select the healthy leaves collected according to the health points, and select the early disease points that were not diseased in the early stage and the diseased points in the late stage or the middle area between two disease points by comparing the color image. The disease point is also the disease point, so the existing spore area is selected to extract data and establish the corn rust data set.

[0061] S62, data preprocessing, selected equal amounts of healthy, early-stage diseased, and diseased corn leaf area samples, and used Savitzky-Golay (SG) filtering and standard normal variant (SNV) to preprocess the data and eliminate noise.

[0062] S63, data division, divides the preprocessed data into a training set and a validation set using the KS algorithm.

[0063] S64, feature extraction, uses recursive feature elimination (RFE) and genetic algorithm (GA) to screen feature bands; RFE selects features by recursively removing features and building a model until the required number of features is reached; GA gradually optimizes feature selection in multiple iterations and eventually finds a set of optimal or near-optimal feature wavelengths; the feature wavelengths finally screened out are mainly concentrated around 800nm ​​in the near-infrared band and around 600nm (orange light) in the visible light band.

[0064] S65, model establishment and optimization, based on the extracted feature bands, constructs a support vector machine (SVM) classification model, and by comparing the model performance of different feature extraction methods, selects the optimal model and saves it to the device to achieve accurate detection of corn rust.

[0065] S66, real-time detection, based on the characteristic bands selected by the established model, found that early disease points and healthy disease points showed obvious differences in the near-infrared band. Healthy disease points had higher reflectivity in the near-infrared band, and obvious disease points had higher reflectivity in the visible light band and lower reflectivity in the near-infrared band. Based on the above characteristics, the characteristic bands were selected for combination to quickly determine the disease status of corn leaves and provide a basis for timely prevention and control measures.

[0066] .

[0067] in, is the pixel detection result, is the image corresponding to the near infrared band in the combined band image, This is the image corresponding to the orange light band in the combined band image. The grayscale value of the pixel point greater than 0.8 is a healthy point, between 0.7 and 0.8 is an early disease point, and less than 0.7 is a disease point.

[0068] Step 7: When processing the spectral image after combining the characteristic bands, the pseudo-color technology is used to assign corresponding color labels according to different disease conditions, so as to effectively distinguish and identify the disease conditions and provide intuitive data support for subsequent analysis and research.

[0069] Step 8: Input the pseudo-color processed image into the pre-trained YOLOv5 segmentation model to achieve accurate segmentation of the target leaves and set the background to black. This process helps improve the accuracy of target recognition and simplifies the image background.

[0070] Step 9, displaying the leaf disease diagnosis result obtained by analysis, that is, the overall leaf detection result, in the result display part of the corn rust detection software interface, and generating a targeted prevention and control plan in real time to provide intuitive and timely feedback to users.

[0071] S91, quantitatively analyze the proportion of diseased areas in leaves for the image information that has been band combined and pseudo-color enhanced and segmented by the YOLOv5 model. The specific calculation formula is as follows.

[0072] .

[0073] in, is the target leaf disease rate; is the area sum of the early onset points and diseased points in the segmented image information; is the total area of ​​the target leaf.

[0074] S92, according to The leaves are divided into four categories: healthy, mild, moderate and severe. When the disease rate of the target leaf is less than 1%, the overall detection result of the leaf is healthy. When the target leaf disease rate is within the range of 1% to 10%, the overall leaf detection result is a lightly diseased leaf; when the target leaf disease rate is When the target leaf disease rate is between 10% and 40%, the overall leaf detection result is a moderately diseased leaf; when the target leaf disease rate is When it is greater than 40%, the overall leaf detection result is a severely diseased leaf. According to the severity of the disease, scientific prevention and control measures are formulated to achieve precise prevention and control.

[0075] Step 10: After entering the "Single Point Analysis" module, the user can click on any pixel on the target leaf image to trigger the system to conduct an in-depth analysis of the spectral data of that point. The analysis results will be presented in real time, intuitively displaying the spectral reflectance of each band and the disease status of the point.

[0076] S101, after the user selects the target area point in the device interface, the system will calculate the spectral data of the point. By comparing with the standard whiteboard spectrum, the reflectivity of the point is calculated and displayed intuitively in the form of a curve. That is, in the "Single Point Analysis" function interface, after the user clicks on a certain point in the screen, the preset program will save the spectral value of the point. By calculating the ratio of the stored whiteboard spectral value to the spectral value of the point, the program will derive the reflectivity of the point and display the reflectivity curve on the interface. The specific calculation formula is as follows.

[0077] .

[0078] in, For the end The reflectivity result of the point, is the spectral value at that point, is the spectral value of the photographed white board.

[0079] S102, input the calculated reflectivity into the corn rust detection model of step 6, and the model will judge the corn rust disease condition of the point according to the reflectivity characteristics, that is, obtain the pixel point detection result.

[0080] Embodiment 2, as Figure 8 As shown, an embodiment of the present application provides a portable corn rust disease detection method, which is implemented using the portable corn rust disease detection device in Example 1; the portable corn rust disease detection method includes the following steps.

[0081] Step 100, obtaining information data of a target corn leaf; the information data includes: image information and spectral information.

[0082] Step 200, based on the corn rust detection model, corn rust disease detection is performed according to the spectral information to determine the pixel point detection result; the pixel point detection result is a healthy point, an early onset point or a diseased point; wherein the healthy point, the early onset point and the diseased point have different pixel point gray value intervals; the corn rust detection model adopts a support vector machine classification model.

[0083] Step 300, using pseudo color technology to perform pseudo color enhancement processing on the image information to obtain processed image information; the pseudo color technology is used to give the image information a corresponding color label according to the pixel detection result.

[0084] Step 400, segmenting the processed image information based on the YOLOv5 segmentation model, and determining the target leaf disease rate according to the segmented image information.

[0085] Step 500, determining the overall leaf detection result corresponding to corn rust disease according to the target leaf disease rate.

[0086] After pseudo-color enhancement, leaves with different disease severity show color differences. The processed image information is input into the YOLOv5 segmentation model for semantic segmentation, accurately segmenting the target leaf area and removing background interference. By calculating the ratio of the diseased area to the total leaf area, the target leaf disease rate is obtained, and the degree of disease can be quantified.

[0087] The calculation formula for the target leaf disease rate is as follows.

[0088] .

[0089] in, is the target leaf disease rate; is the area sum of the early onset points and diseased points in the segmented image information; is the total area of ​​the target leaf.

[0090] Determining the overall leaf detection result corresponding to corn rust disease according to the target leaf disease rate specifically includes the following steps.

[0091] When the target leaf disease rate When it is less than 1%, the overall detection result of the leaf is a healthy leaf.

[0092] When the target leaf disease rate When it is in the range of 1% to 10%, the overall leaf detection result is a slightly diseased leaf.

[0093] When the target leaf disease rate When the leaf density is in the range of 10% to 40%, the overall leaf detection result is moderately diseased.

[0094] When the target leaf disease rate When it is greater than 40%, the overall leaf detection result is a severely diseased leaf.

[0095] In one embodiment, based on a corn rust detection model, corn rust disease detection is performed according to spectral information to determine pixel point detection results, which specifically includes the following steps.

[0096] Recursive feature elimination method and genetic algorithm are used to extract features from spectral information and determine characteristic wavelengths.

[0097] Based on the corn rust detection model, the bands are combined according to the characteristic wavelengths to obtain the combined band image, and the formula is used Determine the pixel detection result.

[0098] in, is the pixel detection result, is the image corresponding to the near infrared band in the combined band image, This is the image corresponding to the orange light band in the combined band image.

[0099] Among them, the method for determining the corn rust detection model specifically includes the following steps.

[0100] Construct a data set; the data set includes: spectral information of known pixel detection results.

[0101] SG filtering and standard normal transformation were used to preprocess the data set to eliminate noise.

[0102] The KS algorithm is used to partition the preprocessed data set to obtain partitioned data; the partitioned data includes: a training set and a validation set.

[0103] The recursive feature elimination method and genetic algorithm are used to extract features from the partitioned data and obtain the training feature bands.

[0104] Build an initial support vector machine classification model.

[0105] The trained feature bands are used to train and optimize the initial support vector machine classification model to obtain a trained support vector machine classification model.

[0106] The trained support vector machine classification model is determined as a corn rust detection model; the corn rust detection model is used to characterize the mapping relationship between characteristic bands, reflectivity and pixel point detection results.

[0107] The device mentioned in this application has a compact structure, clear division of labor among modules, and efficient coordination. The main control module adopts an embedded core main control board, which is responsible for the coordination between modules; the image acquisition module includes a visible light camera and a spectral camera, which cooperates with the main control module to realize the synchronous acquisition of visible light images (image information) and spectral image data (spectral information); the LCD touch screen is used as a display module to provide an intuitive and convenient operation interface, and at the same time has the function of real-time display of corn rust detection results, so that users can obtain disease information instantly; the ranging module can turn on the laser and calculate the distance between the target and the device in real time to ensure that the distance is consistent when detecting each target object.

[0108] The wireless communication positioning module has a GPS positioning function, that is, the current detection location information (geographic location information) is obtained through the GPS positioning module, and data is transmitted through the included wireless transmission module or 4G network. The power module includes a 5000mAh removable lithium battery, which can provide stable power support for the entire device to ensure its reliability in field operations. By comprehensively analyzing image information and spectral information, and analyzing images under characteristic bands, accurate diagnosis of early corn rust can be achieved. The present application has the characteristics of simple operation and rapid detection. It can complete accurate diagnosis of corn rust within seconds, providing an efficient technical means for the early prevention and control of corn rust.

[0109] This application makes full use of spectral imaging technology to achieve early and accurate detection of corn rust, thereby buying precious time for timely prevention and control measures and effectively reducing losses. The device is compact in design and easy to operate, and does not require professional knowledge. Ordinary farmers can also use it easily to achieve rapid detection in the fields.

[0110] The design of this application cleverly solves many inconveniences in traditional detection methods. In the past, farmers needed to carry bulky instruments to the fields, which was cumbersome to operate and had low detection efficiency. The present invention realizes one-handed holding and one-button detection, which greatly simplifies the operation process and improves detection efficiency. At the same time, the function of real-time display and analysis of results enables users to instantly understand the health status of corn leaves, providing a scientific basis for timely prevention and control measures.

[0111] The point-surface fusion measurement can more comprehensively characterize the disease characteristics, improve the accuracy of detection, monitor the spread of the disease, improve the accuracy of early diagnosis, and thus formulate more accurate disease prevention and control measures to improve the prevention and control effect. This device can not only help farmers detect corn rust in real time and accurately in the field, take prevention and control measures as early as possible, effectively reduce disease losses, and ensure corn yields, but also promote agricultural production to develop in a modern and intelligent direction, improve production efficiency, and increase economic benefits. It has broad application prospects.

[0112] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A portable corn rust disease detection device, characterized in that: The portable corn rust disease detection device comprises: an image acquisition module and a main control module; The image acquisition module is connected to the main control module; The image acquisition module is used to collect information data of corn target leaves; the information data includes: image information and spectral information; The main control module is used for: Based on the corn rust detection model, corn rust disease detection is performed according to the spectral information to determine the pixel point detection result; the pixel point detection result is a healthy point, an early onset point or a diseased point; wherein the healthy point, the early onset point and the diseased point have different pixel point gray value intervals; the corn rust detection model adopts a support vector machine classification model; Performing pseudo-color enhancement processing on the image information by using pseudo-color technology to obtain processed image information; the pseudo-color technology is used to give a corresponding color identification to the image information according to the pixel point detection result; The processed image information is segmented based on the YOLOv5 segmentation model, and the target leaf disease rate is determined based on the segmented image information; The overall leaf detection result corresponding to the corn rust disease is determined according to the target leaf disease rate.

2. The portable corn rust disease detection device according to claim 1, characterized in that: The portable corn rust disease detection device further includes: a wireless communication positioning module and a terminal device; the wireless communication positioning module is connected to the terminal device; Wherein, the wireless communication positioning module includes: a GPS positioning module and a wireless transmission module; The GPS positioning module is connected to the wireless transmission module; the wireless transmission module is connected to the terminal device; The GPS positioning module is used to obtain geographical location information; the geographical location information is the location information corresponding to when the information data is obtained; The wireless transmission module is used to transmit the geographical location information and the overall blade detection result to the terminal device.

3. The portable corn rust disease detection device according to claim 1, characterized in that: The portable corn rust disease detection device also includes: a distance measurement module; The distance measuring module is connected to the main control module; the distance measuring module is used to emit laser, measure the distance between the laser emission point and the corn leaf, and transmit the distance to the main control module.

4. The portable corn rust disease detection device according to claim 1, characterized in that: The portable corn rust disease detection device also includes: a power module; The power module adopts a battery with a capacity of 5000mah; the power module is used to provide electric power.

5. The portable corn rust disease detection device according to claim 1, characterized in that: The portable corn rust disease detection device also includes: a display module; The display module adopts a 4.3-inch LCD touch screen and is connected to the main control module via a MIPI interface; the display module is used to display the detection results.

6. The portable corn rust disease detection device according to claim 1, characterized in that: The grayscale value of the pixel corresponding to the healthy point is greater than 0.8; the grayscale value of the pixel corresponding to the early disease point is between 0.7 and 0.8; the grayscale value of the pixel corresponding to the diseased point is less than 0.

7.

7. A portable corn rust disease detection method, characterized in that: The portable corn rust disease detection method is implemented by using the portable corn rust disease detection device according to any one of claims 1 to 6; The portable corn rust disease detection method comprises: Acquire information data of target corn leaves; the information data includes: image information and spectral information; Based on the corn rust detection model, corn rust disease detection is performed according to the spectral information to determine the pixel point detection result; the pixel point detection result is a healthy point, an early onset point or a diseased point; wherein the healthy point, the early onset point and the diseased point have different pixel point gray value intervals; the corn rust detection model adopts a support vector machine classification model; Performing pseudo-color enhancement processing on the image information by using pseudo-color technology to obtain processed image information; the pseudo-color technology is used to give a corresponding color identification to the image information according to the pixel point detection result; The processed image information is segmented based on the YOLOv5 segmentation model, and the target leaf disease rate is determined based on the segmented image information; The overall leaf detection result corresponding to the corn rust disease is determined according to the target leaf disease rate.

8. The portable corn rust disease detection method according to claim 7, characterized in that: Based on the corn rust detection model, corn rust disease detection is performed according to the spectral information to determine the pixel point detection result, specifically including: Using a recursive feature elimination method and a genetic algorithm, feature extraction is performed on the spectral information to determine a characteristic wavelength; Based on the corn rust detection model, the bands are combined according to the characteristic wavelengths to obtain a combined band image, and the formula is used Determine pixel detection results; in, is the pixel detection result, is the image corresponding to the near infrared band in the combined band image, It is the image corresponding to the orange light band in the combined band image; The method for determining the corn rust detection model specifically includes: Constructing a data set; the data set includes: spectral information of known pixel point detection results; The data set is preprocessed using SG filtering and standard normal transformation to eliminate noise; The preprocessed data set is divided using the KS algorithm to obtain divided data; the divided data includes: a training set and a validation set; Using a recursive feature elimination method and a genetic algorithm, feature extraction is performed on the divided data to obtain a training feature band; Build an initial support vector machine classification model; The trained characteristic bands are used to train and optimize the initial support vector machine classification model to obtain a trained support vector machine classification model; The trained support vector machine classification model is determined as a corn rust detection model; the corn rust detection model is used to characterize the mapping relationship between characteristic bands, reflectivity and pixel point detection results.

9. The portable corn rust disease detection method according to claim 7, characterized in that: The calculation formula of the target leaf disease rate is: ； in, is the target leaf disease rate; is the area sum of the early onset points and diseased points in the segmented image information; is the total area of ​​the target leaf.

10. The portable corn rust disease detection method according to claim 9, characterized in that: Determining the overall leaf detection result corresponding to the corn rust disease according to the target leaf disease rate specifically includes: When the target leaf disease rate When it is less than 1%, the overall leaf detection result is a healthy leaf; When the target leaf disease rate When it is in the range of 1% to 10%, the overall leaf test result is a slightly diseased leaf; When the target leaf disease rate When it is in the range of 10% to 40%, the overall leaf test result is a moderately diseased leaf; When the target leaf disease rate When it is greater than 40%, the overall leaf detection result is a severely diseased leaf.

Images