Method for judging damage degree of rat-damaged Chinese pines based on hyperspectrum and physiological and biochemical parameters

Abstract

The invention discloses a method for judging the damage degree of rodent-damaged Chinese pines based on hyperspectrum and physiological and biochemical parameters, which comprises the following steps of: determining a target area of the rodent-damaged Chinese pines according to rodent damage characteristics, partitioning the target area, and collecting needle leaves of the Chinese pines in the target area; acquiring spectral reflectivity data of the pinus tabulaeformis needle leaves by using a spectral radiometer, and acquiring chlorophyll content and water content of the pinus tabulaeformis needle leaves; processing the spectral reflectivity data by using a convolution smoothing method, and removing abnormal values in the chlorophyll content and the water content; performing first-order derivative processing on the smooth spectral reflectivity data by using a spectral differential technology; determining a target vegetation index and a target red edge parameter; training the immediate forest model to obtain a pinus tabulaeformis damage degree estimation model; and outputting the chlorophyll content and the water content of the pinus tabulaeformis needle leaves to be detected through the pinus tabulaeformis damage degree estimation model, and determining the damage degree of the pinus tabulaeformis needle leaves to be detected. According to the method, the accuracy of evaluating the damage degree of the rat-damaged Chinese pines can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of forestry harmful organism monitoring and prevention, and in particular to a method for determining the damage degree of Pinus tabulaeformis based on ground object hyperspectral and physiological parameters. Background Art

[0002] In the returning farmland to forest area of ​​the Loess Plateau, Pinus tabulaeformis is the main planted tree species, and its growth and health are crucial to the stability and sustainability of the ecosystem in the region. However, the Gansu zokor gnaws the roots of Pinus tabulaeformis, seriously hindering its growth and even causing its death, thus causing the degradation of the forest ecosystem. Effective rodent monitoring is the key to rodent control. Accurately classifying the degree of damage to Pinus tabulaeformis and taking corresponding management measures is of great significance to maintaining the ecological environment.

[0003] At present, the damage assessment of Gansu zokor to Pinus tabulaeformis mainly relies on manual field surveys, but this method is not only difficult to accurately identify slightly damaged Pinus tabulaeformis, but also requires a lot of manpower and material resources. Therefore, a more efficient method is urgently needed. With the rapid development of remote sensing technology, it has gradually been applied in forest pest monitoring due to its advantages of fast, efficient and low cost. The present invention proposes a rodent damage monitoring method based on remote sensing technology to solve the limitations of traditional monitoring methods and provide technical support for the accurate assessment and prevention of rodent damage to Pinus tabulaeformis. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the method for determining the degree of damage to Chinese pine caused by rodent infestation based on hyperspectral and physiological and biochemical parameters provided by the present invention solves the problem of inaccurate assessment of the degree of damage to Chinese pine caused by rodent infestation in the prior art.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a method for determining the degree of damage to Pinus tabulaeformis caused by rodents based on hyperspectral and physiological and biochemical parameters, comprising the following steps:

[0006] S1. Determine the target area of ​​rodent-damaged pine according to the characteristics of rodent damage, divide the target area into zones, and collect pine needles in the target area;

[0007] S2. Based on the Chinese pine needles, a spectroradiometer is used to obtain the spectral reflectance data of the Chinese pine needles, and the Chinese pine needles are processed to obtain the chlorophyll content and moisture content of the Chinese pine needles;

[0008] S3, using a convolution smoothing method to process outliers in the spectral reflectance data to obtain smoothed spectral reflectance data, and removing outliers in the chlorophyll content and moisture content to obtain processed chlorophyll content and moisture content;

[0009] S4, using spectral differentiation technology to perform first-order derivative processing on the smoothed spectral reflectance data to obtain first-order differential spectral reflectance data of Pinus tabulaeformis needles;

[0010] S5. Determine a candidate vegetation index and a candidate red edge parameter according to the smoothed spectral reflectance data and the first-order differential spectral reflectance data, and determine a target vegetation index and a target red edge parameter according to the candidate vegetation index and the candidate red edge parameter and the processed chlorophyll content and water content using a correlation calculation formula;

[0011] S6, taking the target vegetation index and the target red edge parameter as input, and taking the processed chlorophyll content and water content as output, training the random forest model, and obtaining a Pinus tabulaeformis damage degree estimation model;

[0012] S7. Input the target vegetation index and target red edge parameter of the pine needles to be tested into the pine damage degree estimation model, output the chlorophyll content and water content of the pine needles to be tested through the pine damage degree estimation model, and determine the damage degree of the pine needles to be tested based on the chlorophyll content and water content of the pine needles to be tested.

[0013] The beneficial effects of the above scheme are as follows: the present invention uses hyperspectral remote sensing of ground objects to obtain the spectral reflectance of Pinus tabulaeformis needles, calculates the spectral index, and analyzes the spectral characteristics of damaged plants and healthy plants; at the same time, the chlorophyll content and moisture content of the needles are measured, and an inversion model of the chlorophyll and moisture content of the needles and the spectral characteristics is established, and the degree of damage to the Pinus tabulaeformis needles to be tested is determined through the spectral characteristics and the chlorophyll content of the needles. The damage of Pinus tabulaeformis zokor to Pinus tabulaeformis is graded by using hyperspectral remote sensing of ground objects, providing basic data for the construction of large-scale remote sensing monitoring of zokor, and improving the scientificity, accuracy and efficiency of rodent control.

[0014] Further, in step S2, based on the Chinese pine needles, a spectroradiometer is used to obtain spectral reflectance data of the Chinese pine needles, which specifically includes:

[0015] Based on the Chinese pine needles, the Chinese pine needles are measured at least 20 times using a spectroradiometer, and the average value of the measurement results of at least 20 times is taken as the spectral reflectance data.

[0016] The beneficial effect of the above further scheme is that slight changes in environmental conditions or differences in operator techniques may cause fluctuations in measurement results. Multiple measurements help smooth out these fluctuations, obtain a more stable average value, reduce the impact of random errors, and improve data reliability.

[0017] Further, in step S2, the pine needles are processed to obtain the chlorophyll content of the pine needles, which specifically includes:

[0018] S21. Treating the needles of Pinus tabulaeformis according to the test method;

[0019] S22, using a UV2600 spectrophotometer to measure the absorbance values ​​of the treated pine needles at 645 nm and 663 nm;

[0020] S23. Determine the chlorophyll content according to the 645nm absorbance value and the 663nm absorbance value and the chlorophyll content calculation formula.

[0021] Furthermore, in step S23, the chlorophyll content calculation formula is:

[0022] Ca=12.7A 663 -2.69A 645

[0023] Cb=22.9A 645 -4.68A 663

[0024] C 总 =Ca+Cb

[0025] Among them, Ca represents the chlorophyll a content, A 663 Indicates the absorbance at 663 nm, A 645 represents the absorbance at 645nm, Cb represents the chlorophyll b content, C 总 Indicates the chlorophyll content.

[0026] The beneficial effect of the above further scheme is that chlorophyll is a key pigment for photosynthesis of plants, and its content directly reflects the photosynthetic capacity and health of plants. By measuring the chlorophyll content, the growth status of plants can be effectively evaluated, and the extent to which the needles of Chinese pine are affected by rodent damage can be discovered in time.

[0027] Furthermore, in step S5, there are 15 vegetation indexes to be selected and 8 target red edge parameters to be selected.

[0028] Furthermore, in step S5, the correlation calculation formula is:

[0029]

[0030] Among them, r represents the correlation, x represents the measured value of any candidate vegetation index or any candidate red edge parameter, y represents the measured value of chloroplast pigment content or water content, and n represents the number of samples.

[0031] The beneficial effect of the above further scheme is that since the spectral curve of the canopy of Pinus tabulaeformis shows significant differences in different bands with the change of rodent damage level, the reflectance change law of the differential spectral curve of Pinus tabulaeformis in different health conditions is obvious in each band. This specific change in spectral reflectance is mainly due to the difference in chlorophyll content and water content of needles. Through correlation analysis, the target vegetation index and red edge parameter with a high correlation with the damage degree of Pinus tabulaeformis can be screened out from the candidate vegetation index and red edge parameter, thereby improving the ability to distinguish the damage degree of Pinus tabulaeformis.

[0032] Furthermore, the method further comprises:

[0033] The determination coefficient, root mean square error and mean relative error were used to evaluate the damage extent estimation model of Pinus tabulaeformis.

[0034] The beneficial effects of the above further scheme are: the comprehensive use of evaluation indicators can comprehensively evaluate the performance of the Pinus tabulaeformis damage estimation model, guide model optimization, ensure that the model can accurately and reliably predict the damage status of Pinus tabulaeformis in practical applications, provide a scientific basis for forest management and rodent control, and promote the sustainable management and protection of forestry resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The figure is a flow chart of a method for determining the degree of damage to Pinus tabulaeformis caused by rodents based on hyperspectral and physiological and biochemical parameters. DETAILED DESCRIPTION

[0036] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0037] like Figure 1 As shown, a method for determining the degree of damage to Pinus tabulaeformis caused by rodents based on hyperspectral and physiological and biochemical parameters includes the following steps:

[0038] S1. According to the characteristics of rodent damage, determine the target area for rodent-damaged pine, divide the target area into zones, and collect pine needles in the target area.

[0039] For example, the target area may be the Pinus tabulaeformis planting area in Maotao Village, Yuanzhou District, Guyuan City, Ningxia. The target area may be divided into 10 plots of 30m×30m, and a systematic random sampling method may be used to collect 6 healthy and damaged Pinus tabulaeformis needles (mild, moderate, severe and dead), for a total of 30 trees as research objects, and each tree is divided into upper, middle and lower parts for sampling to ensure that the damage status of the entire plant can be reflected.

[0040] S2. Based on the Chinese pine needles, a spectroradiometer is used to obtain the spectral reflectance data of the Chinese pine needles, and the Chinese pine needles are processed to obtain the chlorophyll content and moisture content of the Chinese pine needles.

[0041] In this embodiment, in step S2, based on the Chinese pine needles, a spectroradiometer is used to obtain spectral reflectance data of the Chinese pine needles, which specifically includes:

[0042] Based on the Chinese pine needles, the Chinese pine needles are measured at least 20 times using a spectroradiometer, and the average value of the measurement results of at least 20 times is taken as the spectral reflectance data.

[0043] In this embodiment, in step S2, the pine needles are processed to obtain the chlorophyll content of the pine needles, which specifically includes:

[0044] S21. Treating the needles of Pinus tabulaeformis according to the test method;

[0045] S22, using a UV2600 spectrophotometer to measure the absorbance values ​​of the treated pine needles at 645 nm and 663 nm;

[0046] S23. Determine the chlorophyll content according to the 645nm absorbance value and the 663nm absorbance value and the chlorophyll content calculation formula.

[0047] In this embodiment, the chlorophyll content calculation formula is:

[0048] Ca=12.7A 663 -2.69A 645

[0049] Cb=22.9A 645 -4.68A 663

[0050] C 总 =Ca+Cb

[0051] Among them, Ca represents the chlorophyll a content, A 663 Indicates the absorbance at 663 nm, A 645 represents the absorbance at 645nm, Cb represents the chlorophyll b content, C 总 Indicates the chlorophyll content.

[0052] Exemplarily, step S21 may specifically be:

[0053] Cut off all the sample needles collected in the sealed bag, wash and dry them, cut the needles into pieces (the size of the broken needles should not exceed 1.5mm), weigh 0.3g and put it into a test tube, add 25ml 95% ethanol, seal and protect from light. After 36h of extraction, filter the solution into a volumetric flask, make up to 25ml, shake well and number it. Pour the liquid into a centrifuge tube, protect from light for testing, and use 95% ethanol as a control.

[0054] Secondly, a UV2600 spectrophotometer can be used to measure the 645nm absorbance and 663nm absorbance of the treated Pinus tabulaeformis needles, and the average value can be taken.

[0055] For example, the pine needles are processed to obtain the moisture content of the pine needles as follows:

[0056] Step 1: Weigh the fresh weight (FW) of all collected samples.

[0057] Step 2: Place the weighed sample in a drying oven and incubate at 105°C for 30 minutes, then adjust the temperature to 80°C and dry to constant weight, then weigh the dry weight (DW).

[0058] Step 3: Calculate the moisture content of needles. The calculation formula can be: leaf moisture content = (FW-DW) / FW×100%.

[0059] For example, based on Chinese pine needles, using a spectroradiometer to obtain the spectral reflectance data of Chinese pine needles can be specifically as follows: bring the collected numbered pine needle samples back to the room for needle spectrum measurement, and place the pine needle samples on a black curtain. During the measurement process, reflectance correction is performed in a timely manner so that the reflectance of the whiteboard is always 1. To ensure accuracy, 20 spectrum measurements are performed on each sample, and the average value is used as the spectral reflectance of the sample.

[0060] The spectrum was collected using the ASD Field Spec 4 ground feature spectroradiometer produced by ASD Corporation of the United States. The band range of the ASD ground feature spectroradiometer is 350-2500nm, with a total of 2151 bands.

[0061] S3. Use the convolution smoothing method to process the outliers of the spectral reflectance data to obtain smoothed spectral reflectance data, and remove the outliers in the chlorophyll content and moisture content to obtain the processed chlorophyll content and moisture content.

[0062] In this embodiment, during the collection of hyperspectral data, some abnormal data will be generated due to the influence of environmental and equipment noise, so it is very important to remove the abnormal data. Before using the data, the data needs to be removed for abnormal points.

[0063] S4. Use spectral differentiation technology to perform first-order derivative processing on the smoothed spectral reflectance data to obtain first-order differential spectral reflectance data of Pinus tabulaeformis needles.

[0064] In the present embodiment, the SG (Savitzky-Golay) convolution smoothing method is used to process outliers, and quadratic polynomial fitting is used, and the number of smoothing points is 7, thereby obtaining smooth spectral reflectance data. Spectral differential technology is to reduce the interference of baseline drift, background interference and other components in the sample by differential processing of spectral data, and improve the accuracy of detection. Therefore, based on the smoothing spectral data, first-order derivative processing is performed to obtain the first-order differential spectral reflectance of each sample Pinus tabulaeformis.

[0065] S5. Determine the candidate vegetation index and the candidate red edge parameter based on the smoothed spectral reflectance data and the first-order differential spectral reflectance data, and determine the target vegetation index and the target red edge parameter using a correlation calculation formula based on the candidate vegetation index and the candidate red edge parameter and the processed chlorophyll content and water content.

[0066] In this embodiment, in step S5, there are 15 vegetation indices to be selected and 8 target red edge parameters to be selected.

[0067] Furthermore, in step S5, the correlation calculation formula is:

[0068]

[0069] Among them, r represents the correlation, x represents the measured value of any candidate vegetation index or any candidate red edge parameter, y represents the measured value of chloroplast pigment content or water content, and n represents the number of samples.

[0070] S6. Using the target vegetation index and target red edge parameter as input and the processed chlorophyll content and moisture content as output, the random forest model is trained to obtain a model for estimating the damage degree of Pinus tabulaeformis.

[0071] In the random forest model, multiple decision trees are constructed to improve prediction accuracy through ensemble learning. During the training process, the performance of the model can be optimized by setting the number of decision trees in the random forest (such as the ntree parameter) and the minimum number of samples when each tree splits a node (such as the nodesize parameter). In order to ensure the generalization ability of the model and avoid overfitting, the depth of the tree and other parameters can also be adjusted for tuning. In the prediction stage, the trained random forest model can be used to predict the test data set, and the actual value can be compared with the predicted value to evaluate the performance of the model. "

[0072] S7. Input the target vegetation index and target red edge parameter of the pine needles to be tested into the pine damage degree estimation model, output the chlorophyll content and water content of the pine needles to be tested through the pine damage degree estimation model, and determine the damage degree of the pine needles to be tested based on the chlorophyll content and water content of the pine needles to be tested.

[0073] For example, the damage degree of the tested Pinus tabulaeformis needles can be divided into healthy, mild, moderate, severe, and dead according to the actual situation, as shown in Table 1. No specific limitation is made here.

[0074] Table 1 Classification of damage degree of Pinus tabulaeformis

[0075]

[0076] In this embodiment, the method further includes:

[0077] The determination coefficient, root mean square error and mean relative error were used to evaluate the damage extent estimation model of Pinus tabulaeformis.

[0078] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the invention.

Claims

1. A method for determining the degree of damage to Pinus tabulaeformis caused by rodents based on hyperspectral and physiological and biochemical parameters, characterized in that: The method comprises: S1. Determine the target area of ​​rodent-damaged pine according to the characteristics of rodent damage, divide the target area into zones, and collect pine needles in the target area; S2. Based on the Chinese pine needles, using a spectroradiometer to obtain spectral reflectance data of the Chinese pine needles, and processing the Chinese pine needles to obtain the chlorophyll content and moisture content of the Chinese pine needles; S3, using a convolution smoothing method to process outliers on the spectral reflectance data to obtain smoothed spectral reflectance data, and removing outliers in the chlorophyll content and the moisture content to obtain processed chlorophyll content and moisture content; S4, performing first-order derivative processing on the smoothed spectral reflectance data using spectral differential technology to obtain first-order differential spectral reflectance data of the Chinese pine needles; S5. Determine a vegetation index and a red edge parameter to be selected according to the smoothed spectral reflectance data and the first-order differential spectral reflectance data, and determine a target vegetation index and a target red edge parameter according to the vegetation index and the red edge parameter to be selected and the processed chlorophyll content and water content using a correlation calculation formula; S6, taking the target vegetation index and target red edge parameter as input, and taking the processed chlorophyll content and moisture content as output, training the random forest model to obtain a Pinus tabulaeformis damage degree estimation model; S7. Input the target vegetation index and target red edge parameter of the pine needles to be tested into the pine damage degree estimation model, output the chlorophyll content and moisture content of the pine needles to be tested through the pine damage degree estimation model, and determine the damage degree of the pine needles to be tested based on the chlorophyll content and moisture content of the pine needles to be tested.

2. The method according to claim 1, characterized in that In step S2, based on the Chinese pine needles, using a spectroradiometer to obtain the spectral reflectance data of the Chinese pine needles specifically includes: Based on the Chinese pine needles, the Chinese pine needles are measured at least 20 times using a spectroradiometer, and the average value of the measurement results of at least 20 measurements is taken as the spectral reflectance data.

3. The method according to claim 1, characterized in that In step S2, the processing of the Chinese pine needles to obtain the chlorophyll content of the Chinese pine needles specifically includes: S21, treating the Chinese pine needles according to the test method; S22, using a UV2600 spectrophotometer to measure the absorbance value of the treated pine needles at 645 nm and 663 nm respectively; S23. Determine the chlorophyll content according to the 645nm absorbance value and the 663nm absorbance value and the chlorophyll content calculation formula.

4. The method according to claim 3, characterized in that In step S23, the chlorophyll content calculation formula is: <h2 style=";text-align:left;direction:ltr">Ca = 12.7A<h2 style=";text-align:left;direction:ltr"> 663 <h2 style=";text-align:left;direction:ltr"> -2.69A<h2 style=";text-align:left;direction:ltr"> 645 <h2 style=";text-align:left;direction:ltr">Cb = 22.9A<h2 style=";text-align:left;direction:ltr"> 645 <h2 style=";text-align:left;direction:ltr"> -4.68A<h2 style=";text-align:left;direction:ltr"> 663 C 总 =Ca+Cb Among them, Ca represents the chlorophyll a content, A 663 Indicates the absorbance at 663 nm, A 645 represents the absorbance at 645nm, Cb represents the chlorophyll b content, C 总 Indicates the chlorophyll content.

5. The method according to claim 1, characterized in that In step S5, there are 15 vegetation indexes to be selected and 8 target red edge parameters to be selected.

6. The method according to claim 1, characterized in that In step S5, the correlation calculation formula is: Among them, r represents the correlation, x represents the measured value of any candidate vegetation index or any candidate red edge parameter, y represents the measured value of chloroplast pigment content or water content, and n represents the number of samples.

7. The method according to claim 1, characterized in that The method further comprises: The determination coefficient, root mean square error and mean relative error were used to evaluate the damage degree estimation model of Pinus tabulaeformis.

Images