An early diagnosis method for pine wilt disease based on temporal hyperspectral data

Abstract

The present invention provides an early diagnosis method for pine wilt disease based on temporal hyperspectral data, belonging to the technical field of vegetation growth monitoring. The method includes: for the crown of the same target pine tree, obtaining multiple pine tree crown images, where the multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown images are hyperspectral images; based on the multiple pine tree crown images, determining the comprehensive water index and the normalized difference vegetation index corresponding to each pine tree crown image; determining the maximum value of the water index slope and the maximum value of the normalized difference vegetation index slope; based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope, analyzing the changes in leaf water content and chlorophyll content, and being able to achieve efficient diagnosis of pine wilt disease.

Description

Technical Field

[0001] The present invention relates to the technical field of vegetation growth monitoring, and particularly to an early diagnosis method for pine wilt disease based on time-series hyperspectral data. Background Art

[0002] Pine wilt disease, a pine tree disease caused by Bursaphelenchus xylophilus, has a fast transmission speed and a high lethality rate. After a healthy pine tree is infected with this disease, it will die within about 4 weeks at the fastest from the appearance of symptoms to death, and the entire pine forest will be destroyed within 3 to 5 years. Pine wilt disease seriously threatens the forest ecological environment and also causes huge economic losses. At present, there is no effective prevention and control means, and only by felling can the rapid spread and spread of the epidemic be curbed. Therefore, timely and effective diagnosis of infected diseased trees is the key to epidemic prevention and control. How to efficiently diagnose pine wilt disease is an urgent problem to be solved in the industry at present. Summary of the Invention

[0003] In view of the problems existing in the prior art, an embodiment of the present invention provides an early diagnosis method for pine wilt disease based on time-series hyperspectral data.

[0004] In a first aspect, the present invention provides an early diagnosis method for pine wilt disease based on time-series hyperspectral data, including:

[0005] For the crown of the same target pine tree, obtain multiple pine tree crown images, where the multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown images are hyperspectral images;

[0006] Based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image;

[0007] Based on the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image, perform index slope analysis to determine the maximum value of the moisture index slope and the maximum value of the normalized difference vegetation index slope;

[0008] Based on the maximum value of the moisture index slope, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the moisture index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope, analyze the changes in leaf moisture content and chlorophyll content to determine the diagnosis result of pine wilt disease for the crown of the target pine tree;

[0009] Wherein, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum value of the moisture index slope, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum value of the moisture index slope.

[0010] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, based on the maximum value of the slope of the moisture index, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the slope of the moisture index, and the second acquisition time corresponding to the maximum value of the slope of the normalized vegetation index, analyzing the changes in leaf moisture content and chlorophyll content to determine the diagnosis result of pine wilt disease for the target pine tree crown, including:

[0011] Compare the maximum value of the slope of the moisture index with a preset slope threshold to determine the first comparison result;

[0012] Compare the comprehensive moisture index corresponding to the first pine tree crown image with the comprehensive moisture index corresponding to the second pine tree crown image to determine the second comparison result;

[0013] Compare the time sequence before and after between the first acquisition time corresponding to the maximum value of the slope of the moisture index and the second acquisition time corresponding to the maximum value of the slope of the normalized vegetation index to determine the third comparison result;

[0014] If the first comparison result indicates that the maximum value of the slope of the moisture index is greater than the preset slope threshold, and the second comparison result indicates that the comprehensive moisture index corresponding to the first pine tree crown image is greater than the comprehensive moisture index corresponding to the second pine tree crown image, and the first acquisition time corresponding to the maximum value of the slope of the moisture index is earlier than the second acquisition time corresponding to the maximum value of the slope of the normalized vegetation index, then it is determined that the target pine tree crown is in the early stage of pine wilt disease infection.

[0015] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive moisture index and the normalized vegetation index corresponding to each pine tree crown image, including:

[0016] Determine the comprehensive moisture index corresponding to the pine tree crown image through the following comprehensive moisture index calculation formula:

[0017]

[0018] where WI represents the comprehensive moisture index, R m represents the spectral reflectance at wavelength m.

[0019] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, for the spectral reflectance analysis based on the multiple pine tree crown images, determining the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image includes:

[0020] Determining the normalized difference vegetation index corresponding to the pine tree crown image through the following normalized difference vegetation index calculation formula:

[0021]

[0022] where NDVI represents the normalized difference vegetation index, NIR represents the reflectance of the near-infrared band, and R red represents the reflectance of the red band.

[0023] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, obtaining multiple pine tree crown images for the same target pine tree crown includes:

[0024] Obtaining multiple hyperspectral images for the same area, where the multiple hyperspectral images respectively correspond to different acquisition times;

[0025] Based on the multiple hyperspectral images, by delineating the edge of the pine tree crown, determining the position of the target pine tree crown in each hyperspectral image;

[0026] Based on the positions of the target pine tree crown in each hyperspectral image, cropping the multiple hyperspectral images to determine the multiple pine tree crown images.

[0027] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, for determining the position of the target pine tree crown in each hyperspectral image by delineating the edge of the pine tree crown based on the multiple hyperspectral images, includes:

[0028] Calibrating and matching the multiple hyperspectral images to obtain multiple hyperspectral images with aligned images;

[0029] Based on the first hyperspectral image, by delineating the edge of the pine tree crown, determining the position of the target pine tree crown in the first hyperspectral image, where the first hyperspectral image is the hyperspectral image with the earliest acquisition time among the multiple hyperspectral images;

[0030] Based on the multiple hyperspectral images with aligned images and the position of the target pine tree crown in the first hyperspectral image, determining the position of the target pine tree crown in each hyperspectral image.

[0031] Optionally, according to an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, for the same area, obtaining multiple hyperspectral images includes:

[0032] For the same area, based on a preset time interval, aerial photography is carried out by a drone equipped with a hyperspectral sensor to obtain multiple hyperspectral images, and the preset time interval represents the interval duration between the acquisition times corresponding to two adjacent hyperspectral images in time series.

[0033] In a second aspect, the present invention also provides an early diagnosis device for pine wilt disease based on temporal hyperspectral data, including:

[0034] An image acquisition module, configured to obtain multiple pine tree crown images for the same target pine tree crown, the multiple pine tree crown images respectively corresponding to different acquisition times, and the pine tree crown images being hyperspectral images;

[0035] A first determination module, configured to perform spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive water index and the normalized difference vegetation index corresponding to each pine tree crown image;

[0036] A second determination module, configured to perform index slope analysis based on the comprehensive water index and the normalized difference vegetation index corresponding to each pine tree crown image to determine the maximum value of the water index slope and the maximum value of the normalized difference vegetation index slope;

[0037] A third determination module, configured to analyze the changes in leaf water content and chlorophyll content based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index, and determine the diagnosis result of pine wilt disease of the target pine tree crown;

[0038] Wherein, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum value of the water index slope, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum value of the water index slope.

[0039] In a third aspect, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the early diagnosis method of pine wilt disease based on temporal hyperspectral data as described in any one of the above.

[0040] Fourthly, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the early diagnosis method of pine wilt disease based on temporal hyperspectral data as described in any one of the above is implemented.

[0041] The early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention obtains multiple pine tree crown images for the same target pine tree crown, and can perform spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive water index and normalized vegetation index corresponding to each pine tree crown image. Furthermore, index slope analysis can be performed to determine the maximum value of the water index slope and the maximum value of the normalized vegetation index slope. Furthermore, based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, through the combined judgment of the water index and the vegetation index, it is possible to detect the infected diseased wood in a timely and effective manner when the pine is in the early stage of nematode disease infection, and achieve an efficient diagnosis of pine wilt disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0043] Figure 1 is one of the flow diagrams of the early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention;

[0044] Figure 2 is another flow diagram of the early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention;

[0045] Figure 3 is the structural diagram of the early diagnosis device of pine wilt disease based on temporal hyperspectral data provided by the present invention;

[0046] Figure 4 is the structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] To facilitate a clearer understanding of the embodiments of the present invention, some relevant background knowledge is introduced as follows.

[0048] In the related art, the diagnosis of pine wilt disease mainly relies on single-temporal or multi-temporal visible light / hyperspectral images. Sensitive bands are screened or vegetation indices are constructed, and traditional machine learning and neural network classifiers are used to diagnose pine wilt disease at an early stage.

[0049] There are many deficiencies in the method of diagnosing pine wilt disease based on single-temporal or multi-temporal visible light / hyperspectral images.

[0050] (1) The diagnosis of pine wilt disease by single-temporal visible light / hyperspectral methods mainly focuses on the middle and late stages of the disease occurrence. At this time, the canopy area of the infected pine trees has changed color significantly. Due to the extremely fast spread speed and harm of pine wilt disease, it cannot meet the requirements of "early detection and early treatment". In addition, pine trees that die due to other pests and diseases and natural reasons are very similar in color characteristics to diseased logs, and it is difficult to distinguish them in single-temporal images, which is prone to misjudgment.

[0051] (2) There is a problem that the sensitivity of early diagnosis of pine wilt disease by multi-temporal visible light is not high enough. Since there is no band sensitive to canopy physiological changes, the monitoring results of visible light data are often in the middle and late stages of the infection process. Therefore, it is difficult to meet the requirements of early diagnosis using multi-temporal visible light.

[0052] (3) For the early diagnosis of pine wilt disease based on multi-temporal hyperspectral images, the existing methods usually collect image data at relatively large intervals, generally about one month. The infected pine trees may be completely infected and discolored within the above time, and the process of the pine trees from the disease occurrence area of the crown to the complete infection and discoloration of the entire canopy cannot be observed. In addition, the early diagnosis methods such as vegetation indices proposed based on multi-temporal spectral data have the defect of being greatly affected by tree species and environmental factors.

[0053] In order to overcome the above defects, the present invention provides an early diagnosis method for pine wilt disease based on time-series hyperspectral data. By analyzing the changes in leaf water content and chlorophyll content, the efficient diagnosis of pine wilt disease can be achieved.

[0054] To make the purpose, technical solution and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0055] Figure 1 is one of the flow schematic diagrams of the early diagnosis method for pine wilt disease based on time-series hyperspectral data provided by the present invention, as Figure 1As shown, the execution subject of the early diagnosis method for pine wilt disease based on time-series hyperspectral data can be an electronic device, such as a server, etc. The method includes:

[0056] Step 101, for the crown of the same target pine tree, obtain multiple pine tree crown images, where the multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown images are hyperspectral images.

[0057] Specifically, in order to achieve efficient diagnosis of pine wilt disease, multiple pine tree crown images can be obtained. The images of multiple pine tree crowns are collected for the crown of the same target pine tree, and the multiple pine tree crown images respectively correspond to different acquisition times. By comparing the multiple pine tree crown images, the growth status and change trend of the pine tree crown in different time periods can be analyzed.

[0058] Step 102, based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive moisture index and normalized difference vegetation index corresponding to each pine tree crown image.

[0059] Specifically, based on the multiple pine tree crown images, by performing spectral analysis on each image, the reflectance data of different bands can be extracted, and then the comprehensive moisture index and normalized difference vegetation index can be calculated. The comprehensive moisture index is an index for judging the moisture status of plants by comparing the reflectance data of different bands. The normalized difference vegetation index is an index for evaluating the vegetation coverage degree by comparing the reflectance data of red light and near-infrared light.

[0060] It can be understood that the comprehensive moisture index can be determined by statistical analysis of multiple moisture indices.

[0061] Step 103, based on the comprehensive moisture index and normalized difference vegetation index corresponding to each pine tree crown image, perform index slope analysis to determine the maximum value of the moisture index slope and the maximum value of the normalized difference vegetation index slope.

[0062] Specifically, by performing index slope analysis on the comprehensive moisture index and normalized difference vegetation index of each pine tree crown image, the maximum value of the moisture index slope and the maximum value of the normalized difference vegetation index slope are determined. In this process, mathematical methods can be used to calculate the slope of each index, and then the maximum value of the slope is found. These maximum values can assist in analyzing the growth status and water use situation of pine trees.

[0063] Optionally, based on the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image, a plurality of moisture index slope values and a plurality of normalized difference vegetation index slope values are determined. The moisture index slope value is used to characterize the slope between the comprehensive moisture indexes corresponding to two adjacent pine tree crown images in time series, and the normalized difference vegetation index slope value is used to characterize the slope between the normalized difference vegetation indexes corresponding to two adjacent pine tree crown images in time series. Based on the plurality of moisture index slope values and the plurality of normalized difference vegetation index slope values, the maximum moisture index slope value and the maximum normalized difference vegetation index slope value are determined.

[0064] Step 104, based on the maximum moisture index slope value, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum moisture index slope value, and the second acquisition time corresponding to the maximum normalized difference vegetation index slope value, analyze the changes in leaf moisture content and chlorophyll content, and determine the diagnosis result of pine wilt disease for the target pine tree crown.

[0065] Among them, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum moisture index slope value, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum moisture index slope value.

[0066] Specifically, based on the maximum moisture index slope value, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum moisture index slope value, and the second acquisition time corresponding to the maximum normalized difference vegetation index slope value, the change trends of leaf moisture content and chlorophyll content can be further analyzed. By comparing the data at different acquisition times, the diagnosis result of pine wilt disease for the target pine tree crown can be determined.

[0067] The early diagnosis method for pine wilt disease based on time series hyperspectral data provided by the present invention, by acquiring multiple pine tree crown images for the same target pine tree crown, can perform spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image, and then can perform index slope analysis to determine the maximum moisture index slope value and the maximum normalized difference vegetation index slope value. Furthermore, based on the maximum moisture index slope value, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum moisture index slope value, and the second acquisition time corresponding to the maximum normalized difference vegetation index slope value, through the combined judgment of the moisture index and the vegetation index, it is possible to timely and effectively detect the infected diseased wood when the pine is in the early stage of nematode disease infection, and achieve an efficient diagnosis of pine wilt disease.

[0068] Optionally, for an early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention, based on the maximum value of the slope of the moisture index, the comprehensive moisture index corresponding to the first pine tree canopy image, the comprehensive moisture index corresponding to the second pine tree canopy image, the first acquisition time corresponding to the maximum value of the slope of the moisture index, and the second acquisition time corresponding to the maximum value of the slope of the normalized difference vegetation index, analyzing the changes in leaf moisture content and chlorophyll content to determine the diagnosis result of pine wilt disease for the target pine tree canopy, including:

[0069] Comparing the maximum value of the slope of the moisture index with a preset slope threshold to determine a first comparison result;

[0070] Comparing the comprehensive moisture index corresponding to the first pine tree canopy image with the comprehensive moisture index corresponding to the second pine tree canopy image to determine a second comparison result;

[0071] Comparing the time sequence before and after between the first acquisition time corresponding to the maximum value of the slope of the moisture index and the second acquisition time corresponding to the maximum value of the slope of the normalized difference vegetation index to determine a third comparison result;

[0072] If the first comparison result indicates that the maximum value of the slope of the moisture index is greater than the preset slope threshold, and the second comparison result indicates that the comprehensive moisture index corresponding to the first pine tree canopy image is greater than the comprehensive moisture index corresponding to the second pine tree canopy image, and the first acquisition time corresponding to the maximum value of the slope of the moisture index is earlier than the second acquisition time corresponding to the maximum value of the slope of the normalized difference vegetation index, then it is determined that the target pine tree canopy is in the early stage of pine wilt disease infection.

[0073] Specifically, by using a comparison method, the maximum value of the slope of the moisture index can be compared with a preset slope threshold to determine the first comparison result. The comprehensive moisture index corresponding to the first pine tree canopy image can be compared with the comprehensive moisture index corresponding to the second pine tree canopy image to determine the second comparison result. The time sequence before and after between the first acquisition time corresponding to the maximum value of the slope of the moisture index and the second acquisition time corresponding to the maximum value of the slope of the normalized difference vegetation index can be compared to determine the third comparison result. If the first comparison result indicates that the maximum value of the slope of the moisture index is greater than the preset slope threshold, and the second comparison result indicates that the comprehensive moisture index corresponding to the first pine tree canopy image is greater than the comprehensive moisture index corresponding to the second pine tree canopy image, and the first acquisition time corresponding to the maximum value of the slope of the moisture index is earlier than the second acquisition time corresponding to the maximum value of the slope of the normalized difference vegetation index, then it can be determined that the target pine tree canopy is in the early stage of pine wilt disease infection.

[0074] It is understandable that after the pine wood nematode infects the pine tree, it blocks the vessels and prevents the transportation of water, resulting in a gradual weakening of photosynthesis. However, since the lack of water precedes the destruction of chlorophyll, diagnosing whether the pine wilt disease is in the early stage by detecting the change in the water content of the leaves is more sensitive than conventional indicators such as calculating vegetation indices that reflect chlorophyll content, that is, the disease can be detected more effectively and earlier. Therefore, using time-series hyperspectral image data, by detecting the canopy water content, it is mainly reflected in observing the change in the water index, rather than simply using the vegetation index as in the past, which has more physiological significance. By combining the water index and the vegetation index for judgment, early diagnosis of the disease is ensured under the premise that it is difficult to detect obvious differences in the canopy color change by visual observation.

[0075] Therefore, based on the high-resolution hyperspectral data of diseased infected trees with time-series changes, early diagnosis of pine wilt disease is achieved by comprehensively observing the changes in the water index and the vegetation index.

[0076] Optionally, according to an early diagnosis method for pine wilt disease based on time-series hyperspectral data provided by the present invention, for the spectral reflectance analysis based on the multiple pine tree canopy images to determine the comprehensive water index and the normalized vegetation index corresponding to each pine tree canopy image, it includes:

[0077] The comprehensive water index corresponding to the pine tree canopy image is determined through the following comprehensive water index calculation formula:

[0078]

[0079] where WI represents the comprehensive water index, R m represents the spectral reflectance at a wavelength of m.

[0080] Specifically, the spectral reflectance R at a wavelength of 950 nm 950 and the spectral reflectance R at a wavelength of 900 nm 900 are taken to determine the first water index WI1. The spectral reflectance R at a wavelength of 695 nm 695 and the spectral reflectance R at a wavelength of 760 nm 760 are taken to determine the second water index WI2. The spectral reflectance R at a wavelength of 740 nm 740 and the spectral reflectance R at a wavelength of 720 nm 720 are taken to determine the third water index WI3. Furthermore, the comprehensive water index WI can be determined by taking the average of the first water index WI1, the second water index WI2, and the third water index WI3.

[0081] It is understandable that by calculating the average value of the three moisture indices to establish a comprehensive moisture index, the error in indicating canopy moisture information by a single moisture index can be overcome.

[0082] Optionally, according to an early diagnosis method of pine wilt disease based on time-series hyperspectral data provided by the present invention, for performing spectral reflectance analysis on the multiple pine tree crown images and determining the comprehensive moisture index and normalized difference vegetation index corresponding to each pine tree crown image, it includes:

[0083] Determining the normalized difference vegetation index corresponding to the pine tree crown image through the following normalized difference vegetation index calculation formula:

[0084]

[0085] where NDVI represents the normalized difference vegetation index, NIR represents the reflectance in the near-infrared band, and R red represents the reflectance in the red band.

[0086] Optionally, according to an early diagnosis method of pine wilt disease based on time-series hyperspectral data provided by the present invention, for obtaining multiple pine tree crown images for the same target pine tree crown, it includes:

[0087] Obtaining multiple hyperspectral images for the same area, where the multiple hyperspectral images respectively correspond to different acquisition times;

[0088] Based on the multiple hyperspectral images, by depicting the edge of the pine tree crown, determining the position of the target pine tree crown in each hyperspectral image;

[0089] Based on the position of the target pine tree crown in each hyperspectral image, cropping the multiple hyperspectral images to determine the multiple pine tree crown images.

[0090] Specifically, in order to study the target area more precisely, multiple hyperspectral images can be obtained, and these images respectively correspond to different acquisition times. By analyzing these images, the edge of the pine tree crown can be depicted, thereby determining the position of the target pine tree crown in each hyperspectral image. Then, using this position information, the multiple hyperspectral images can be cropped to obtain multiple pine tree crown images. In this way, the vegetation information of the target area can be analyzed more precisely, providing a more reliable data basis for subsequent research.

[0091] Optionally, according to an early diagnosis method of pine wilt disease based on time-series hyperspectral data provided by the present invention, for determining the position of the target pine tree crown in each hyperspectral image by depicting the edge of the pine tree crown based on the multiple hyperspectral images, it includes:

[0092] Calibrate and match the multiple hyperspectral images to obtain multiple aligned hyperspectral images;

[0093] Based on the first hyperspectral image, determine the position of the target pine tree crown in the first hyperspectral image by depicting the edge of the pine tree crown. The first hyperspectral image is the hyperspectral image with the earliest acquisition time among the multiple hyperspectral images;

[0094] Based on the multiple aligned hyperspectral images and the position of the target pine tree crown in the first hyperspectral image, determine the position of the target pine tree crown in each hyperspectral image.

[0095] Specifically, in the specific implementation, it is necessary to calibrate and match multiple hyperspectral images to obtain multiple aligned hyperspectral images. This process can be achieved through various image registration algorithms, such as feature point-based registration algorithms, phase correlation-based registration algorithms, etc. After the image alignment is completed, the position of the target pine tree crown can be determined based on the first hyperspectral image. Specifically, its position can be determined by depicting the edge of the pine tree crown, and this process can be completed manually or automatically. The first hyperspectral image is the hyperspectral image with the earliest acquisition time among the multiple hyperspectral images, so it can be used as a reference for other hyperspectral images. Next, using the multiple aligned hyperspectral images and the position of the target pine tree crown in the first hyperspectral image, the position of the target pine tree crown in each hyperspectral image can be determined. This process can be achieved by calculating the relative positions of each hyperspectral image and the first hyperspectral image. Finally, the position information of the target pine tree crown in each hyperspectral image can be obtained, providing a basis for subsequent data analysis and processing.

[0096] Optionally, according to an early diagnosis method for pine wilt disease based on time-series hyperspectral data provided by the present invention, the obtaining of multiple hyperspectral images for the same area includes:

[0097] For the same area, based on a preset time interval, conduct aerial photography with a drone equipped with a hyperspectral sensor to obtain multiple hyperspectral images. The preset time interval represents the interval duration between the acquisition times corresponding to two adjacent hyperspectral images in the time series.

[0098] Specifically, during aerial photography, the drone will fly based on a preset time interval, along a preset flight path and at a preset altitude. Meanwhile, the hyperspectral sensor will continuously record and collect spectral information within the area. This spectral information will be converted into digital signals to form hyperspectral images. By processing and analyzing these images, information such as spectral reflectance and radiation values at different wavelengths within the area can be obtained, thereby obtaining more detailed and accurate surface features and environmental conditions.

[0099] Optionally, Figure 2 is the second schematic flow diagram of the early diagnosis method for pine wilt disease based on temporal hyperspectral data provided by the present invention. As Figure 2 shown, the method includes steps 201 to 209.

[0100] Step 201, use a drone to collect hyperspectral images in a fixed area, and collect image data a total of n times.

[0101] Specifically, a drone can be used to collect hyperspectral images in a fixed area for about one month, collect image data a total of n times, and the collection interval is 4 - 6 days. The drone is equipped with a hyperspectral sensor, and the parameter indicators of the hyperspectral sensor include: the spectral range is 450nm - 950nm, the sampling interval is 4nm, the spectral resolution is 8nm, and there are a total of 125 channels.

[0102] Optionally, before collecting the hyperspectral images, radiation processing needs to be carried out. The specific process is as follows: use a black and white board to perform radiation calibration on the hyperspectral sensor to correct the spectral reflectance of the collection area. The calculation formula for radiation calibration is as follows:

[0103]

[0104] DN target and DN white respectively represent the digital quantization values (DN values) of the target object and the standard white board in the original hyperspectral image. Ref target and Ref white respectively represent the reflectances of the target object and the reference white board, and DN dark represents the systematic error.

[0105] Step 202, perform image calibration and matching on the n hyperspectral images collected from the diseased and infected trees.

[0106] Specifically, using the RPC Orthorectification Using ReferenceImage function in ENVI 5.3 SP1 software, select the first hyperspectral image as the reference image, and use the i-th image as the image to be calibrated and matched, where i takes values of 2, 3, …, n. First, uniformly select 9 control points on the first image, and then find the positions of the control points on the first image on the remaining n - 1 images, and calibrate the control points accordingly. Click the next program to perform automatic calculations to align the n - 1 images with the first image.

[0107] Step 203: Obtain the image of a single crown of the Chinese fir with canker.

[0108] Specifically, in the ENVI 5.3 SP1 software, click the Region of Interest (ROI) Tool, depict the edge part of the crown of the Chinese fir with canker in the first image, and save it as an.xml file. Then click the Subset Data from ROIs tool, first open the saved.xml file, crop the n images, and then export and save the image with only a single crown of the Chinese fir with canker.

[0109] Step 204: Determine the comprehensive moisture index of the crown part of the Chinese fir with canker.

[0110] Specifically, determine the comprehensive moisture index corresponding to the image of the pine tree crown through the following comprehensive moisture index calculation formula:

[0111]

[0112] Among them, WI (Water index) represents the comprehensive moisture index. R m represents the spectral reflectance at wavelength m.

[0113] Step 205: Determine the normalized difference vegetation index of the crown part of the Chinese fir with canker.

[0114] Specifically, determine the normalized difference vegetation index corresponding to the image of the pine tree crown through the following normalized difference vegetation index calculation formula:

[0115]

[0116] Among them, NDVI represents the normalized difference vegetation index, NIR represents the reflectance in the near-infrared band, and R red represents the reflectance in the red band.

[0117] Step 206: Determine multiple moisture index slope values and multiple normalized difference vegetation index slope values.

[0118] Specifically, for the same discolored diseased tree, calculate the slope KW of the water index WI between two adjacent time-series images i , and calculate the normalized vegetation index KN between two adjacent time-series images i , where i ∈ (1, n - 1).

[0119] Step 207: Determine the maximum value of the slope of the water index and the maximum value of the slope of the normalized vegetation index.

[0120] Specifically, take the maximum value of the absolute values of KW i and KN i , and denote them as KW t1_max and KN t2_max respectively. t1 and t2 are the sampling periods corresponding to the maximum absolute values, where t1 ∈ (1, n) and t2 ∈ (1, n).

[0121] Step 208: Determine the comprehensive water index corresponding to the first pine tree crown image and the comprehensive water index corresponding to the second pine tree crown image.

[0122] Specifically, calculate the average values of WI before and after t1 respectively, denoted as WI before_t1 and WI after_t1 respectively.

[0123] Step 209: Determine whether the pine tree is in the early stage of Bursaphelenchus xylophilus infection.

[0124] Specifically, use the following conditions to determine whether the pine tree is in the early stage of Bursaphelenchus xylophilus infection:

[0125]

[0126] In the formula, KW t1_max > 0.5 indicates that there is an obvious change in water content, WI before_t1 > WI after_t1 ensures that the water content is still decreasing after this change occurs at time t1, and t1 < t2 ensures that the water change occurs earlier than the change in chlorophyll.

[0127] It can be understood that after Bursaphelenchus xylophilus infects a pine tree, it will block the ducts and hinder the transportation of water, resulting in a gradual weakening of photosynthesis. However, since the lack of water appears earlier than the destruction of chlorophyll, by detecting the change in the water content of the leaves, the early stage of Bursaphelenchus xylophilus disease can be detected more sensitively. In contrast, conventional indicators such as calculating vegetation indices that reflect chlorophyll content are not so effective. Therefore, using time-series hyperspectral image data, by detecting the canopy water content, mainly reflected in observing the change in the water index rather than simply using the vegetation index as in the past, has more physiological significance. By combining the water index and the vegetation index for judgment, the disease can be detected early before obvious differences are difficult to detect by visual observation.

[0128] The early diagnosis method of pine wilt disease based on temporal hyperspectral data provided by the present invention can obtain multiple pine tree crown images for the same target pine tree crown. Based on the multiple pine tree crown images, spectral reflectance analysis can be carried out to determine the comprehensive moisture index and normalized vegetation index corresponding to each pine tree crown image. Furthermore, index slope analysis can be carried out to determine the maximum value of the moisture index slope and the maximum value of the normalized vegetation index slope. Furthermore, based on the maximum value of the moisture index slope, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the moisture index slope, and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, through the combined judgment of the moisture index and the vegetation index, it is possible to detect diseased infected trees in a timely and effective manner when the pine is in the early stage of nematode disease infection, and achieve an efficient diagnosis of pine wilt disease.

[0129] The early diagnosis device of pine wilt disease based on temporal hyperspectral data provided by the present invention will be described below. The early diagnosis device of pine wilt disease based on temporal hyperspectral data described below can be mutually referred to the early diagnosis method of pine wilt disease based on temporal hyperspectral data described above.

[0130] Figure 3 is a schematic structural diagram of the early diagnosis device of pine wilt disease based on temporal hyperspectral data provided by the present invention, as Figure 3 shown, the early diagnosis device of pine wilt disease based on temporal hyperspectral data includes: an image acquisition module 301, a first determination module 302, a second determination module 303, and a third determination module 304, wherein:

[0131] The image acquisition module 301 is used to acquire multiple pine tree crown images for the same target pine tree crown. The multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown image is a hyperspectral image;

[0132] The first determination module 302 is used to perform spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive moisture index and normalized vegetation index corresponding to each pine tree crown image;

[0133] The second determination module 303 is used to perform index slope analysis based on the comprehensive moisture index and normalized vegetation index corresponding to each pine tree crown image to determine the maximum value of the moisture index slope and the maximum value of the normalized vegetation index slope;

[0134] A third determination module 304, configured to analyze the changes in leaf water content and chlorophyll content based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, and determine the diagnosis result of pine wilt disease of the target pine tree crown;

[0135] Wherein, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum value of the water index slope, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum value of the water index slope.

[0136] Figure 4 is a schematic structural diagram of an electronic device provided by the present invention, as Figure 4 shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440. Among them, the processor 410, the communication interface 420, and the memory 430 complete mutual communication through the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an early diagnosis method for pine wilt disease based on time-series hyperspectral data. The method includes:

[0137] For the same target pine tree crown, obtain multiple pine tree crown images, and the multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown images are hyperspectral images;

[0138] Based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive water index and the normalized vegetation index corresponding to each pine tree crown image;

[0139] Based on the comprehensive water index and the normalized vegetation index corresponding to each pine tree crown image, perform index slope analysis to determine the maximum value of the water index slope and the maximum value of the normalized vegetation index slope;

[0140] Based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, analyze the changes in leaf water content and chlorophyll content, and determine the diagnosis result of pine wilt disease of the target pine tree crown;

[0141] Among them, the acquisition moment corresponding to the first pine tree crown image is the previous acquisition moment of the first acquisition moment corresponding to the maximum value of the moisture index slope, and the acquisition moment corresponding to the second pine tree crown image is the next acquisition moment of the first acquisition moment corresponding to the maximum value of the moisture index slope.

[0142] In addition, when the logical instructions in the above-mentioned memory 430 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0143] On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the early diagnosis method of pine wilt disease based on time-series hyperspectral data provided by the above methods. The method includes:

[0144] For the same target pine tree crown, obtain multiple pine tree crown images, and the multiple pine tree crown images respectively correspond to different acquisition moments. The pine tree crown images are hyperspectral images;

[0145] Based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive moisture index and normalized difference vegetation index corresponding to each pine tree crown image;

[0146] Based on the comprehensive moisture index and normalized difference vegetation index corresponding to each pine tree crown image, perform index slope analysis to determine the maximum value of the moisture index slope and the maximum value of the normalized difference vegetation index slope;

[0147] Based on the maximum value of the moisture index slope, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition moment corresponding to the maximum value of the moisture index slope, and the second acquisition moment corresponding to the maximum value of the normalized difference vegetation index slope, analyze the changes in leaf moisture content and chlorophyll content to determine the diagnosis result of pine wilt disease of the target pine tree crown;

[0148] Among them, the acquisition moment corresponding to the first pine tree crown image is the previous acquisition moment of the first acquisition moment corresponding to the maximum value of the slope of the moisture index, and the acquisition moment corresponding to the second pine tree crown image is the next acquisition moment of the first acquisition moment corresponding to the maximum value of the slope of the moisture index.

[0149] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0150] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An early diagnosis method for pine wilt disease based on temporal hyperspectral data, characterized in that, Including: For the same target pine tree crown, obtain multiple pine tree crown images, where the multiple pine tree crown images respectively correspond to different acquisition times, and the pine tree crown images are hyperspectral images; Based on the multiple pine tree crown images, perform spectral reflectance analysis to determine the comprehensive water index and normalized difference vegetation index corresponding to each pine tree crown image; Based on the comprehensive water index and normalized difference vegetation index corresponding to each pine tree crown image, perform index slope analysis to determine the maximum value of the water index slope and the maximum value of the normalized difference vegetation index slope; Based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope, analyze the changes in leaf water content and chlorophyll content to determine the diagnosis result of pine wilt disease for the target pine tree crown; Among them, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum value of the water index slope, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum value of the water index slope; The step of analyzing the changes in leaf water content and chlorophyll content based on the maximum value of the water index slope, the comprehensive water index corresponding to the first pine tree crown image, the comprehensive water index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the water index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope to determine the diagnosis result of pine wilt disease for the target pine tree crown includes: Compare the maximum value of the water index slope with a preset slope threshold to determine the first comparison result; Compare the comprehensive water index corresponding to the first pine tree crown image with the comprehensive water index corresponding to the second pine tree crown image to determine the second comparison result; Compare the time sequence between the first acquisition time corresponding to the maximum value of the water index slope and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope to determine the third comparison result; If the first comparison result indicates that the maximum value of the water index slope is greater than the preset slope threshold, and the second comparison result indicates that the comprehensive water index corresponding to the first pine tree crown image is greater than the comprehensive water index corresponding to the second pine tree crown image, and the first acquisition time corresponding to the maximum value of the water index slope is earlier than the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope, then determine that the target pine tree crown is in the early stage of pine wilt disease infection.

2. The early diagnosis method of pine wilt disease based on temporal hyperspectral data according to claim 1, characterized in that The step of performing spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive water index and normalized difference vegetation index corresponding to each pine tree crown image includes: Determine the comprehensive water index corresponding to the pine tree crown image through the following comprehensive water index calculation formula: Among them, WI represents the comprehensive moisture index, R m represents the spectral reflectance at a wavelength of m.

3. The early diagnosis method of pine wilt disease based on temporal hyperspectral data according to claim 1, wherein The step of performing spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive water index and normalized difference vegetation index corresponding to each pine tree crown image includes: Determine the normalized difference vegetation index (NDVI) corresponding to the pine tree crown image through the following NDVI calculation formula: Among them, NDVI represents the Normalized Difference Vegetation Index, NIR represents the reflectance in the near-infrared band, and R red represents the reflectance in the red band.

4. The early diagnosis method of pine wilt disease based on temporal hyperspectral data according to claim 1, characterized in that, For the same target pine tree crown, obtain multiple pine tree crown images, including: For the same area, obtain multiple hyperspectral images, where the multiple hyperspectral images correspond to different acquisition times respectively; Based on the multiple hyperspectral images, determine the positions of the target pine tree crowns in each hyperspectral image by delineating the edges of the pine tree crowns; Based on the positions of the target pine tree crowns in each hyperspectral image, crop the multiple hyperspectral images to determine the multiple pine tree crown images.

5. The early diagnosis method of pine wilt disease based on time-series hyperspectral data according to claim 4, characterized in that, The step of determining the positions of the target pine tree crowns in each hyperspectral image by delineating the edges of the pine tree crowns based on the multiple hyperspectral images includes: Calibrate and match the multiple hyperspectral images to obtain multiple aligned hyperspectral images; Based on the first hyperspectral image, determine the position of the target pine tree crown in the first hyperspectral image by delineating the edges of the pine tree crown, where the first hyperspectral image is the hyperspectral image with the earliest acquisition time among the multiple hyperspectral images; Based on the multiple aligned hyperspectral images and the position of the target pine tree crown in the first hyperspectral image, determine the positions of the target pine tree crowns in each hyperspectral image.

6. The early diagnosis method of pine wilt disease based on temporal hyperspectral data according to claim 4, characterized in that The step of obtaining multiple hyperspectral images for the same area includes: For the same area, based on a preset time interval, conduct aerial photography using a drone equipped with a hyperspectral sensor to obtain multiple hyperspectral images, where the preset time interval represents the time interval between the acquisition times corresponding to two adjacent hyperspectral images in the time series.

7. An early diagnosis device for pine wilt disease based on time - series hyperspectral data, characterized in that, Including: An image acquisition module, configured to obtain multiple pine tree crown images for the same target pine tree crown, where the multiple pine tree crown images correspond to different acquisition times respectively, and the pine tree crown images are hyperspectral images; A first determination module, configured to perform spectral reflectance analysis based on the multiple pine tree crown images to determine the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image; A second determination module, configured to perform index slope analysis based on the comprehensive moisture index and the normalized difference vegetation index corresponding to each pine tree crown image to determine the maximum value of the moisture index slope and the maximum value of the normalized difference vegetation index slope; A third determination module, configured to analyze the changes in leaf moisture content and chlorophyll content based on the maximum value of the moisture index slope, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the moisture index slope, and the second acquisition time corresponding to the maximum value of the normalized difference vegetation index slope, and determine the diagnosis result of pine wilt disease for the target pine tree crown; Wherein, the acquisition time corresponding to the first pine tree crown image is the previous acquisition time of the first acquisition time corresponding to the maximum value of the moisture index slope, and the acquisition time corresponding to the second pine tree crown image is the next acquisition time of the first acquisition time corresponding to the maximum value of the moisture index slope; Analyze the changes in leaf water content and chlorophyll content based on the maximum value of the moisture index slope, the comprehensive moisture index corresponding to the first pine tree crown image, the comprehensive moisture index corresponding to the second pine tree crown image, the first acquisition time corresponding to the maximum value of the moisture index slope, and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, and determine the diagnosis result of pine wilt disease for the target pine tree crown, including: Compare the maximum value of the moisture index slope with a preset slope threshold to determine a first comparison result; Compare the comprehensive moisture index corresponding to the first pine tree crown image with the comprehensive moisture index corresponding to the second pine tree crown image to determine a second comparison result; Compare the time sequence before and after between the first acquisition time corresponding to the maximum value of the moisture index slope and the second acquisition time corresponding to the maximum value of the normalized vegetation index slope to determine a third comparison result; If the first comparison result indicates that the maximum value of the moisture index slope is greater than the preset slope threshold, and the second comparison result indicates that the comprehensive moisture index corresponding to the first pine tree crown image is greater than the comprehensive moisture index corresponding to the second pine tree crown image, and the first acquisition time corresponding to the maximum value of the moisture index slope is earlier than the second acquisition time corresponding to the maximum value of the normalized vegetation index slope, then determine that the target pine tree crown is in the early stage of pine wilt disease infection.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein When the processor executes the program, it implements the early diagnosis method for pine wilt disease based on temporal hyperspectral data according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the early diagnosis method for pine wilt disease based on temporal hyperspectral data according to any one of claims 1 to 6.

Images