Method and system for measuring curl index of corn leaf and evaluating drought stress degree

Abstract

The invention relates to a method and a system for measuring a curl index of a corn leaf and evaluating a drought stress degree. According to the method, a foundation laser radar is used for collecting point cloud data of a corn plant in the morning and the afternoon under different drought stress degrees in a large bell mouth period; threshold segmentation and stem leaf segmentation are carried out; performing optimization preprocessing operation on the blade point cloud; cutting the point cloud of the blade by using a gradually expanded cylinder, and calculating the vein point of the blade; carrying out equidistant plane segmentation on the corn leaf from the tip to the base of the leaf; calculating the Euclidean distance between two points which are farthest from each other in each section of blade section curve point cloud of the two sample sets; fitting the blade section curve point cloud by using a greedy sorting algorithm based on the shortest distance and a cubic spline curve, and calculating a spline curve fitting length; calculating a blade curl index and an average actual blade curl index of the blade; and finally performing drought stress degree grading evaluation.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of agricultural production, and relates to a method and system for measuring a corn leaf rolling index and evaluating a drought stress degree based on point cloud data. BACKGROUND

[0002] Under the background of climate change, the growth and yield of corn are often affected by drought stress and other abiotic stresses. In order to stabilize the corn yield in China, it is urgent to reduce the impact of drought stress on the growth and yield of corn. Timely determination of the drought stress degree of corn plants is of great significance for guiding intelligent irrigation and breeding of excellent drought-resistant varieties. Corn drought indicators include meteorological drought indicators, hydrological drought indicators and plant morphological indicators, among which: meteorological and hydrological indicators need to measure environmental variables such as rainfall, evaporation, air temperature and soil moisture to construct a model, which has the problems of many input parameters, complex calculation and inability to continuously measure; plant morphological indicators are also important drought indicators, which can evaluate the drought stress degree of corn through plant morphology, and have the advantages of simplicity, real-time and the like.

[0003] Drought stress will affect plant morphological parameters such as leaf rolling, leaf angle and plant height. Existing researches have shown that leaf rolling is one of the most sensitive responses of corn plants to drought stress, and by measuring leaf rolling and exploring the response mechanism between leaf rolling and drought, the drought stress degree of corn plants can be quickly and accurately detected. Existing researches have found that corn plants show a certain regularity within a day after being subjected to drought stress with changes in temperature and light. In the morning, the leaves are naturally flat and the leaf width is maximum, at noon when the temperature is highest, the leaves are most rolled and the leaf width is minimum, and with the decrease of temperature and light intensity at sunset, the rolled leaves begin to slowly flatten.

[0004] The degree of leaf rolling can be represented by a leaf rolling index (LRI), LRI = [(Lw-Ln) / Lw]x100%, wherein Lw is the maximum width of the flat leaf, and Ln is the Euclidean distance of the leaf edge at the corresponding position. At present, the measurement of LRI often adopts a manual method, and Lw and Ln are measured at a single position by subjective judgment of the experimenter. The experimenter first judges the maximum leaf width of the narrow leaf, and then manually measures the Euclidean distance (Ln) of the leaf edge at the position and the maximum width (Lw) of the flat leaf. The manual judgment of the reasonable measurement point has subjectivity, the measurement of the single position LRI has unreliability in measuring the overall rolling degree of the narrow leaf, and there are the disadvantages of long working time, harsh environment and large task.

[0005] With the development of computer and sensor technologies, three-dimensional point cloud data of crops, which reflects the spatial morphology of crops and is obtained using sensors such as structured light, time-of-flight cameras, and laser scanners, can provide data support for the non-destructive and accurate acquisition of crop morphological phenotypic parameters. The ground-based lidar used in this invention acquires crop plant point clouds using the time-of-flight method, which has advantages such as large range, high accuracy, and low susceptibility to light interference.

[0006] Therefore, this invention proposes a method and system for measuring the leaf curling index and evaluating the degree of drought stress in maize plants based on point cloud data. Point cloud data of maize plants at two times (morning and noon) under different drought stress levels during the large trumpet stage are collected using ground-based lidar. Computer software and algorithms are used to segment the two sample sets into individual plants and organs, respectively. Leaf width is measured at regular intervals from tip to base for the same maize leaf. An index quantifying the degree of leaf curling caused by drought stress in maize plants—the average actual leaf curling index—is proposed and calculated. This indicator was used to classify and evaluate the degree of drought stress on maize. Summary of the Invention

[0007] To address the problems of low efficiency in measuring leaf curling index and unclear indicators and methods for judging the degree of drought stress in maize in existing methods, the purpose of this invention is to provide a method and system for measuring the leaf curling index and evaluating the degree of drought stress in maize based on point cloud data.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for measuring maize leaf curl index and evaluating drought stress level includes the following steps:

[0010] Step 1: Use ground-based lidar to collect point cloud data of maize plants at two times of day, morning and noon, under different drought stress levels during the large trumpet stage;

[0011] Among them, different drought stress levels include five drought stress levels: no drought, mild drought, moderate drought, severe drought, and extreme drought; point cloud data of maize leaves were collected at two times, one in the morning and one at noon, when the maize leaves were naturally extended and the other when the maize leaves were most curled, and these were used as sample set 1 and sample set 2, respectively.

[0012] Step 2: Threshold segmentation is performed on the population corn point clouds of the two collected sample sets to obtain the point cloud of a single corn plant. Then, manual stem and leaf segmentation is performed on each single corn plant point cloud to obtain the stem point cloud and the leaf point cloud. The leaf point cloud of at least one corn leaf at the top of the corn plant is selected and the leaf point cloud is optimized by smoothing, filtering, upsampling and downsampling to obtain the optimized smooth, uniform leaf point cloud with no obvious noise.

[0013] Step 3: Use an increasingly larger cylinder to cut the leaf point cloud to obtain small segments of leaf point cloud. Find the geometric center of each small segment of leaf point cloud to obtain the leaf vein points.

[0014] Step 4: Based on the leaf vein points, the corn leaf is divided into equidistant plane segments from the leaf tip to the base to obtain multiple leaf cross-section curve point clouds.

[0015] Step 5: Calculate the Euclidean distance between the two furthest points in the leaf cross-section curve point clouds of each of the two sample sets. This yields the naturally curled leaf width Ln1 when the maize leaf is naturally extended and the drought-stressed curled leaf width Ln2 when the maize leaf is most curled, both distributed at equal intervals along the leaf vein direction. Using a greedy sorting algorithm based on the shortest distance and a cubic spline curve fitting algorithm, fit the leaf cross-section curve point cloud of sample set 1 to calculate the spline curve fitting length, thus obtaining the fully extended leaf width Lw1 when the maize leaf is naturally extended. Based on the naturally curled leaf width Ln1, the drought-stressed curled leaf width Ln2, and the fully extended leaf width Lw1, calculate the leaf curl index LRI1 and LRI2 corresponding to each segment of the leaf cross-section curve point cloud. Finally, calculate the average actual leaf curl index.

[0016] Step 6: Assessment of drought stress severity.

[0017] The specific steps for Step 3 are as follows:

[0018] Step 3.1: Read in the point cloud of corn leaves and stalks, and obtain the maximum and minimum values ​​of the X, Y, and Z axis coordinates of the stalk point cloud. smax y smax z smax x smin y smin z min Obtain the minimum Z-coordinate value z of the leaf point cloud. lmin L, the point with the largest Z coordinate zmax (x lmax ,y lmax ,z lmax );

[0019] Step 3.2: The coordinates of a point on the axis of the cylinder being cut are C. cylinder (x cylinder ,y cylinder ,z cylinder ), calculate its coordinates according to formula (1);

[0020]

[0021] In the formula, x cylinder y cylinder z cylinder x represents the X-axis, Y-axis, and Z-axis coordinates of a point on the cylindrical axis; smax x smin y smax y smin These represent the maximum and minimum values ​​of the X-axis and Y-axis coordinates of the stem point cloud, respectively; z lmin The minimum Z-coordinate value of the leaf point cloud;

[0022] Step 3.3: Calculate the projection vector G(x) of the leaf growth direction onto the XOY plane. lmax -x cylinder ,y lmax -y cylinder ,0), the cylindrical axial vector is the perpendicular vector of the growth direction vector K(1,k,0), where k is calculated according to formula (2);

[0023]

[0024] In the formula, x cylinder y cylinder x and y are the X and Y coordinates of a point on the cylindrical axis, respectively; x lmax y lmax , , are the X and Y coordinates of the point with the largest Z coordinate in the leaf point cloud, respectively; k is the Y coordinate of the perpendicular vector to be obtained;

[0025] Step 3.4: Using vector K and a point C on the cylinder axis... cylinder Determine the axis of the cylinder, set the cylinder radius to 0.02*j (in meters), obtain the blade point cloud within the cylinder, and save it as a blade annular point cloud PC. cj Other leaf point clouds are updated to new leaf point clouds, j starts from 1, j = 1, 2, 3, ... until all leaf point clouds are divided by cylinders from smallest to largest;

[0026] Step 3.5: Read in all the leaf ring point clouds, calculate the three-dimensional coordinates of the geometric center point of the bounding box of all the leaf ring point clouds, and the three-dimensional coordinates of the geometric center point are the coordinates of the leaf vein points.

[0027] Step 3.6: Find the point L with the largest Z-coordinate in the leaf point cloud. zmax (x lmax ,y lmax ,z lmax The three-dimensional coordinates of the center points of the leaf rings obtained above are sorted in reverse order and stored sequentially in a new point cloud, forming all the leaf vein points C from the tip to the base of the leaf. gi, i = 0, 2, 3, ..., n-1, where n is the total number of leaf veins.

[0028] Step 4 involves the following specific steps:

[0029] Step 4.1: Let the plane equation of the cross section be Ax + By + Cz + D = 0. The four parameters of the plane equation form four arrays A[], B[], C[], D[]; In the first set of parameters of the plane equation, A[0], B[0], and C[0] are the X-axis, Y-axis, and Z-axis coordinates of the first leaf vein vector V0(x0, y0, z0), respectively. The first point of the plane equation passes through the first leaf vein point, that is, the point L with the largest Z-coordinate in the leaf point cloud. zmax The initial A[0], B[0], C[0], D[0] are calculated according to formula (3), and the subsequent parameters are calculated from step 4.2 to step 4.8.

[0030]

[0031] In the formula:

[0032] A[0],B[0],C[0],D[0] are the parameters of the first set of plane equations;

[0033] x0, y0, z0 are the X-axis, Y-axis, and Z-axis coordinates of the first leaf vein vector;

[0034] x g0 ,y g0 ,z g0 These are the X-axis, Y-axis, and Z-coordinates of the first vein point on the leaf, respectively.

[0035] Step 4.2: Generate the leaf vein from the i-th point C gi (x gi ,y gi ,z gi ) to the (i+1)th leaf vein point C gi+1 (x gi+1 ,y gi+1 ,z gi+1 The vector V i (x i ,y i ,z i () represents the leaf vein vector, and the distance between two adjacent leaf vein points constitutes a small segment of leaf length l. i , i = 0, 2, 3, ..., n-2, where n is the total number of leaf veins;

[0036] Step 4.3: Set the segmentation interval H = 0.02, the initial segmentation distance H0 = 0.02, and the initial index m = 1 for the parameters of the plane equation to be determined;

[0037] Step4.4: Starting from i = 1, if i > n, execute step4.9; otherwise, execute step4.5;

[0038] Step4.5: If l i <H0, it means there is no segmentation plane behind this vein point, update the segmentation distance H = H - l i , and then execute Step4.8; otherwise, execute step4.6;

[0039] Step4.6: If l i <H + H0, it means there is one segmentation plane behind this vein point, calculate its parameters according to formula (4) and update the segmentation distance H, and then execute Step4.8; otherwise, it means there are two segmentation planes behind this vein point, execute step4.7;

[0040]

[0041] In the formula:

[0042] A[m], B[m], C[m], D[m] are the parameters of the m-th group of plane equations to be solved;

[0043] x i , y i , z i are the X-axis, Y-axis, and Z-axis coordinates of the i-th vein vector;

[0044] x gi , y gi , z gi are the X-axis, Y-axis, and Z-axis coordinates of the i-th vein point;

[0045] H is the segmentation interval, and its unit is m;

[0046] H0 is the initial segmentation distance, and its unit is m;

[0047] l i is the length of the small leaf segment between the i-th and the (i + 1)-th vein points, and its unit is m;

[0048] Step4.7: Calculate the first segmentation plane behind the i-th vein point according to formula (4) and update the segmentation distance, increment m, calculate the second segmentation plane behind the i-th vein point according to formula (5) and update the segmentation distance, increment m;

[0049] ​​​​Step 4.8: Increment i, calculate the segmentation plane after the next leaf vein point, and execute step 4.4;

[0052] Step 4.9: All the dividing planes have been calculated and stored in four parameter arrays. Use the obtained plane equations to extract the blade point cloud. For each set of parameters, i.e. m0 = 0, 1, 2...m, m is the number of planes obtained in the above steps. Substitute the X-axis, Y-axis and Z-axis coordinates of the blade point cloud into formula (6) in sequence. If the conditions are met, export and store it as a new point cloud, i.e., the blade section curve point cloud.

[0053]

[0054] In the formula:

[0055] x, y, z refer to the X-axis, Y-axis, and Z-axis coordinates of all points on the blade section curve substituted into the inequality;

[0056] A[m0], B[m0], C[m0], D[m0] are the corresponding parameters of all the obtained plane equations, m0 = 0, 1, 2...m, where m is the number of planes obtained in the above steps;

[0057] Step 4.10: End the process of equidistantly cropping the leaf point cloud.

[0058] In Step 5, the leaf cross-section curve point cloud between the 4th and 13th segments is selected, that is, the leaf range from 6cm to 24cm from the leaf tip, which is 18cm from the middle of the leaf, as the study area.

[0059] Step 5 uses a cubic spline curve with segmented parameters to fit the cross-sectional curve. The smoothing parameters of the segments are set according to formula (7). The fitted curve is represented by 1000 interpolation points. The distance between each pair of interpolation points is calculated and the spline curve fitting length is accumulated. That is, the fully flat leaf width Lw1 when the corn leaf is naturally spread out is calculated.

[0060]

[0061] In the formula:

[0062] smooth_param is the smoothing parameter for fitting cubic spline curves;

[0063] Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaf is naturally unfolded in the morning, and its unit is m.

[0064] In Step 5, the fully flat leaf width of sample set 2 is calculated using the fully flat leaf width Lw1 of the corn leaf when it is naturally spread out, as calculated in sample set 1.

[0065] In Step 5, if the fully extended leaf width Lw1 of a corn leaf when it is naturally spread out is smaller than the width of the preceding and following segments in the point cloud of the leaf cross-section curve of a certain segment in Sample Set 1, then the segment is an abnormal case with missing leaf parts. The point cloud data of the leaf cross-section curve of that segment in Sample Set 1 is removed, and the corresponding segment of the leaf cross-section curve point cloud data of the corresponding leaf in Sample Set 2 is also removed.

[0066] In Step 5, calculate the leaf curling index LRI1 when the corn leaves are naturally extended, the leaf curling index LRI2 when the corn leaves are most curled, and the average actual leaf curling index. The formula is as follows:

[0067]

[0068]

[0069] In the formula,

[0070] LRI1 is the leaf curling index of a corn leaf when it is naturally unfolded, corresponding to the point cloud of a certain segment of the leaf cross-section curve.

[0071] LRI2 is the leaf curling index corresponding to the point cloud of a certain segment of the leaf cross-section curve when the corn leaf is most curled.

[0072] Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaves are naturally unfolded in the morning, and its unit is m.

[0073] Lw1 is the full flat width of a corn leaf when it is naturally spread out, which is the length of the cross-sectional curve of the leaf when it is naturally spread out in the morning, and its unit is m.

[0074] Ln2 is the width of the drought-stressed curled leaf when the corn leaf is most curled, which is the Euclidean distance between the farthest points of the cross-sectional curves of the corn leaf when it is most curled under drought stress, and its unit is m.

[0075] LRI r The actual leaf curling index caused by drought stress corresponds to the point cloud of a certain segment of the leaf cross-section curve.

[0076] The average actual leaf curl index;

[0077] q represents the total number of effective leaf curl indices under actual drought stress.

[0078] In Step 6, the drought stress severity grading evaluation criteria are as follows:

[0079] Average actual leaf curl index If the drought stress level is greater than 30%, the drought stress level is classified as extreme drought or severe drought.

[0080] Average actual leaf curl index Between 15% and 30%, the drought stress level is considered moderate drought.

[0081] Average actual leaf curl index Between 10% and 15%, the drought stress level is considered mild drought;

[0082] Average actual leaf curl index Below 10%, the drought stress level is considered non-drought.

[0083] This invention provides a system for measuring maize leaf curling index and evaluating drought stress, including a data acquisition module, a data processing module, and a drought stress evaluation module for implementing the above-mentioned methods for measuring maize leaf curling index and evaluating drought stress.

[0084] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0085] This invention uses maize point cloud data collected by ground-based lidar to calculate the average actual leaf curl index of maize leaves, and establishes a graded evaluation model for the degree of drought stress in maize based on this index. The point cloud processing methods, such as extracting leaf vein points and equidistantly segmenting leaf point clouds, developed independently throughout the process, are simple, fast, and accurate. The proposed evaluation index—the average actual leaf curl index of maize leaves—provides an objective and efficient graded evaluation of drought stress. In the process of maize growth, rapidly determining and quantifying the degree of drought stress through leaf curl morphology is of great significance for guiding smart irrigation and the breeding of superior drought-resistant varieties. Attached Figure Description

[0086] Figure 1 A flowchart illustrating a method for measuring maize leaf curling index and evaluating drought stress level provided by the present invention;

[0087] Figure 2 This is a schematic diagram of the location of the data acquisition scanning station provided by the present invention;

[0088] Figure 3 The flowchart for extracting leaf vein points based on point cloud cylindrical cutting provided by this invention;

[0089] Figure 4 This is a schematic diagram of leaf vein points provided by the present invention;

[0090] Figure 5 A flowchart for equidistant blade cutting provided by the present invention;

[0091] Figure 6 This is a schematic diagram illustrating the effect of equidistant blade cutting provided by the present invention;

[0092] Figure 7 The flowchart for calculating the blade curling index provided by this invention. Detailed Implementation

[0093] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0094] like Figure 1 The method shown here for measuring maize leaf curl index and evaluating drought stress includes the following steps:

[0095] Step 1: Use ground-based lidar to collect point cloud data of maize plants at two times of day, morning and noon, under different drought stress levels during the large trumpet stage;

[0096] The different drought stress levels include five levels: no drought, mild drought, moderate drought, severe drought, and extreme drought. Point cloud data of maize leaves were collected at two times, one in the morning and one at noon, when the maize leaves were naturally extended and the other when the maize leaves were most curled, and these were used as sample set 1 and sample set 2, respectively.

[0097] This invention uses ground-based lidar to collect point cloud data of maize leaves at 8:00 AM on a sunny day under different drought stress levels during the large trumpet stage, as sample set 1. Point cloud data of leaves at 2:00 PM on the same day, when the leaves are almost most curled, is collected as sample set 2.

[0098] Step 2: Threshold segmentation is performed on the population corn point clouds of the two collected sample sets to obtain the point cloud of a single corn plant. Then, manual stem and leaf segmentation is performed on each single corn plant point cloud to obtain the stem point cloud and the leaf point cloud. The leaf point cloud of at least one corn leaf at the top of the corn plant is selected and optimized by smoothing, filtering, upsampling and downsampling preprocessing to obtain the optimized smooth, uniform leaf point cloud without obvious noise.

[0099] Step 3: Use an increasingly larger cylinder to cut the leaf point cloud to obtain small segments of leaf point cloud. Find the geometric center of each small segment of leaf point cloud to obtain the leaf vein points.

[0100] Step 4: Based on the leaf vein points, the corn leaf is divided into equidistant plane segments from the leaf tip to the base to obtain multiple leaf cross-section curve point clouds.

[0101] Step 5: Calculate the Euclidean distance between the two furthest points in the leaf cross-section curve point clouds of each of the two sample sets. This yields the naturally curled leaf width Ln1 when the maize leaf is naturally extended and the drought-stressed curled leaf width Ln2 when the maize leaf is most curled, both distributed at equal intervals along the leaf vein direction. Using a greedy sorting algorithm based on the shortest distance and a cubic spline curve fitting algorithm, fit the leaf cross-section curve point cloud of sample set 1 to calculate the spline curve fitting length, thus obtaining the fully extended leaf width Lw1 when the maize leaf is naturally extended. Based on the naturally curled leaf width Ln1, the drought-stressed curled leaf width Ln2, and the fully extended leaf width Lw1, calculate the leaf curl index LRI1 and LRI2 corresponding to each segment of the leaf cross-section curve point cloud. Finally, calculate the average actual leaf curl index.

[0102] Because the tips of corn leaves are relatively small, the proportion of change with the width of the curled leaf can be very large, and the point cloud obtained from the tip is relatively small, which can easily lead to large errors. The part that is too close to the base of the leaf is more obscured by other leaves, resulting in more missing point clouds, which can also lead to large measurement errors. Therefore, in Step 5, the point cloud of the leaf cross-section curve between the 4th and 13th segments (that is, the leaf range of 18cm from 6cm from the tip of the leaf to 24cm from the tip of the leaf) is selected as the study area.

[0103] In Step 5, the fully flat leaf width of Sample Set 2 is calculated using the fully flat leaf width Lw1 of the corn leaf when it is naturally unfolded, as calculated in Sample Set 1. This is because the leaves in Sample Set 2 are more curled and have a more complex shape, which is not conducive to point cloud processing and fitting a better cross-sectional curve. Moreover, the variation in fully flat leaf width caused by leaf growth in the morning and afternoon of the same day is negligible. For the fully flat leaf width Lw1 of the corn leaf when it is naturally unfolded in a certain segment of the leaf cross-sectional curve point cloud in Sample Set 1, if it is smaller than both the preceding and following segments, then this segment is an abnormal case of missing leaf parts. The point cloud data of the leaf cross-sectional curve of this segment in Sample Set 1 is removed, and the corresponding segment of the leaf cross-sectional curve point cloud data of the corresponding leaf in Sample Set 2 is also removed.

[0104] In Step 5, calculate the leaf curling index LRI1 when the corn leaves are naturally extended, the leaf curling index LRI2 when the corn leaves are most curled, and the average actual leaf curling index. The formula is as follows:

[0105]

[0106] In the formula,

[0107] LRI1 is the leaf curling index of a corn leaf when it is naturally unfolded, corresponding to the point cloud of a certain segment of the leaf cross-section curve.

[0108] LRI2 is the leaf curling index corresponding to the point cloud of a certain segment of the leaf cross-section curve when the corn leaf is most curled.

[0109] Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaves are naturally unfolded in the morning, and its unit is m.

[0110] Lw1 is the full flat width of a corn leaf when it is naturally spread out, which is the length of the cross-sectional curve of the leaf when it is naturally spread out in the morning, and its unit is m.

[0111] Ln2 is the width of the drought-stressed curled leaf when the corn leaf is most curled, which is the Euclidean distance between the farthest points of the cross-sectional curves of the corn leaf when it is most curled under drought stress, and its unit is m.

[0112] LRI r The actual leaf curling index caused by drought stress corresponds to the point cloud of a certain segment of the leaf cross-section curve.

[0113] The average actual leaf curl index;

[0114] q represents the total number of effective leaf curl indices under actual drought stress.

[0115] Step 6: Assessment of drought stress severity.

[0116] The criteria for classifying and evaluating the severity of drought stress are as follows:

[0117] Average actual leaf curl index of leaves in the extreme drought group and the severe drought group Similar, both greater than 30%;

[0118] Moderate drought group Between 15% and 30%;

[0119] Mild drought group Between 10% and 15%;

[0120] Non-drought group Below 10%.

[0121] Example

[0122] Test materials

[0123] The selected variety is Jingnongke 728, a compact maize variety that is widely planted in the Beijing-Tianjin-Tangshan and Huang-Huai-Hai regions and is suitable for intensive planting and mechanized harvesting.

[0124] Experimental Design

[0125] The experiment was conducted outdoors using potted plants. A 2-3 cm layer of decomposed leaves was placed at the bottom of a styrofoam box (inner diameter: 34 cm long, 24 cm wide, 19 cm high; outer diameter: 39 cm long, 29 cm wide, 24 cm high) as an organic matter layer, followed by local natural soil, filling the box to 2 cm from the top edge. A total of 30 pots were used, each sown with three Jingnongke 728 corn seeds. When the corn seedlings reached the three-leaf stage, one strong plant was retained from each pot, and the other two excess seedlings were removed. The pots were spaced 1 meter apart to ensure the leaves did not overlap. The corn was planted on May 20, 2024, under the following conditions: outdoor sunlight, daytime high temperatures approximately 25℃–32℃, and nighttime low temperatures approximately 15℃–26℃.

[0126] Once the maize reached the large trumpet stage (the stage where the corn plants begin to tassel), it was divided into five treatment groups: no drought, mild drought, moderate drought, severe drought, and extreme drought, with six pots of maize in each treatment group. According to the Ministry of Agriculture standard NYT 2284-2012—Technical Specification for Field Investigation and Grading of Maize Disasters—the degree of drought stress on maize at the large trumpet stage can be divided into five levels, corresponding to different soil moisture and plant morphological characteristics. Based on surveys and experiments, the field water holding capacity of the soil in the potted plants in this experiment was approximately 36%. The relative soil moisture content multiplied by the field water holding capacity was the required controlled soil moisture content. The relevant soil moisture requirements and control conditions are shown in Table 1.

[0127] Table 1. Soil moisture requirements and control measures during the experiment.

[0128]

[0129] Step 1: Data Collection

[0130] Based on the above experimental design, different gradient irrigations were conducted for one week during the corn's large trumpet stage, placing it under different drought levels before data collection. Data was collected at 8:00 AM and 2:00 PM on July 7, 2024, with data collected at the morning when the corn leaves were naturally flat and at noon when they were curled, respectively. Because outdoor lidar data collection is significantly affected by wind, a simple windbreak was constructed using transparent plastic sheeting during the data collection experiment. The point cloud data acquisition equipment was the Trimble TX8 ground-based lidar, with a measurement range of 0.6m-120m, a field of view of 360°×317°, a scan density level of 2, and a measurement accuracy of 2mm. Scanning stations were distributed at the four corners of the experimental corn area, and four point cloud data were obtained for each experiment. The data was stored on the lidar's external USB drive. The scanning station locations and potted corn numbers are as follows: Figure 2 As shown, the positions of scanning stations 1 to 4 are respectivelyFigure 2 Table 2 shows the correspondence between the potted corn serial numbers and the set soil drought conditions.

[0131] Table 2. Correspondence between potted corn plant serial numbers and soil drought settings.

[0132]

[0133] Step 2: Data Preprocessing

[0134] Data preprocessing mainly includes target registration, data format conversion, point cloud filtering, single-plant segmentation, stem and leaf segmentation, and point cloud optimization. The specific steps are as follows:

[0135] Step 2.1: Import the data from the LiDAR USB drive into Trimble RealWorks software, use the self-extracting target registration, perform point cloud registration using the target sphere, and then complete the orientation of the registered point cloud.

[0136] Step 2.2: Extract point cloud data from the test area using a clipping box and perform spatial sampling. The sampling distance is set to 1 mm, and the exported point cloud data format is .las. In order to process the acquired point cloud using the PCL point cloud library, the corn plant point cloud data format is converted from .las to .pcd.

[0137] Step 2.3: Use pass-through filtering and statistical filtering to remove ground point clouds and noise points.

[0138] Step 2.4: Since the corn plants are planted in a regular pattern, thresholds are set in the X, Y, and Z axes to perform direct filtering segmentation and obtain point clouds of individual corn plants.

[0139] Step 2.5: Import the point cloud of a single corn plant into the cloudcompare software, manually segment the stem and leaves to obtain the point cloud of the stem and leaves of a single corn plant;

[0140] When dividing the stems and leaves of a single corn plant, select one to three leaves from the upper part of the corn plant (in order of leaf position from low to high, approximately the 7th to 9th leaves). The upper corn leaves are more tender, sensitive to moisture, and easier to process.

[0141] Step 2.6: Smooth the corn leaf point cloud using Moving Least Squares (MLS), then perform statistical filtering, then upsampling, and finally voxel filtering downsampling to obtain an optimized, smooth, uniform corn leaf point cloud with no obvious noise.

[0142] Step 3: Extract leaf vein points by gradually expanding cylindrical cutting.

[0143] The coordinates of a point on the cylindrical axis are obtained by using the spatial information of the stem and leaves. The direction of the cylindrical axis is obtained by the growth direction of the leaves. The cylinder is cut by setting a fixed cylindrical axis and increasing cylinder radii. Leaf vein points are extracted from the point cloud of the small annular leaf segments obtained by the cutting. The algorithm flow is as follows: Figure 3 As shown, the leaf vein point extraction effect is as follows: Figure 4 As shown, the specific steps are as follows:

[0144] Step 3.1: Read in the point cloud of corn leaves and stalks, and obtain the maximum and minimum values ​​of the X, Y, and Z axis coordinates of the stalk point cloud. smax y smax z smax x smin y smin z min Obtain the minimum Z-coordinate value z of the leaf point cloud. lmin L, the point with the largest Z coordinate zmax (x lmax ,y lmax ,z lmax ).

[0145] Step 3.2: The coordinates of a point on the axis of the cylinder being cut are C. cylinder (x cylinder ,y cylinder ,z cylinder ), calculate its coordinates according to formula (1).

[0146]

[0147] In the formula, x cylinder y cylinder z cylinder x represents the X-axis, Y-axis, and Z-axis coordinates of a point on the cylindrical axis; smax x smin y smax y smin These represent the maximum and minimum values ​​of the X-axis and Y-axis coordinates of the stem point cloud, respectively; z lmin The minimum Z-coordinate value of the leaf point cloud.

[0148] Step 3.3: Calculate the projection vector G(x) of the leaf growth direction onto the XOY plane. lmax -x cylinder ,y lmax -y cylinder ,0), the axial vector of the cylinder is the perpendicular vector of the growth direction vector K(1,k,0), where k is calculated according to formula (2).

[0149]

[0150] In the formula, x cylinder y cylinder x and y are the X and Y coordinates of a point on the cylindrical axis, respectively; x lmax y lmax , , are the X-axis and Y-axis coordinates of the point with the largest Z-coordinate in the leaf point cloud, respectively; k is the Y-axis coordinate of the perpendicular vector to be obtained.

[0151] Step 3.4: Using vector K and a point C on the cylinder axis... cylinder Determine the axis of the cylinder, set the cylinder radius to 0.02*j (in meters), obtain the blade point cloud within the cylinder, and save it as a blade annular point cloud PC. cj Other leaf point clouds are updated to new leaf point clouds, with j starting from 1, j = 1, 2, 3, ... until all leaf point clouds are divided by cylinders from smallest to largest.

[0152] Step 3.5: Read in all the leaf ring point clouds, calculate the three-dimensional coordinates of the geometric center point of the bounding box of all the leaf ring point clouds, and the three-dimensional coordinates of the geometric center point are the coordinates of the leaf vein point.

[0153] Step 3.6: Find the point L with the largest Z-coordinate in the leaf point cloud. zmax (x lmax ,y lmax ,z lmax The three-dimensional coordinates of the center points of the leaf rings obtained above are sorted in reverse order and stored sequentially in a new point cloud, forming all the leaf vein points C from the tip to the base of the leaf. gi , i = 0, 2, 3, ..., n-1, where n is the total number of leaf veins.

[0154] Step 4: Cut the blades at equal intervals

[0155] Based on the leaf vein points extracted from the cylindrical cutting process described above, the corn leaf point cloud is segmented into equidistant planes from the leaf tip to the base to obtain the leaf cross-sectional curve point cloud; the algorithm flow is as follows: Figure 5 As shown, the point cloud effect of the blade cross-section curve is as follows: Figure 6 As shown, the specific steps are as follows.

[0156] Step 4.1: Let the plane equation of the cross section be Ax + By + Cz + D = 0. The four parameters of the plane equation form four arrays A[], B[], C[], D[]. In the first set of parameters of the plane equation, A[0], B[0], and C[0] are the X-axis, Y-axis, and Z-axis coordinates of the first leaf vein vector V0(x0, y0, z0), respectively. The first point of the plane equation passes through the first leaf vein point, that is, the point L with the largest Z-coordinate in the leaf point cloud. zmax The initial A[0], B[0], C[0], D[0] are calculated according to formula (3), and the subsequent parameters are calculated from step 4.2 to step 4.8.

[0157]

[0158] Where:

[0159] A[0], B[0], C[0], D[0] are the parameters of the first set of plane equations;

[0160] x0, y0, z0 are the X-axis, Y-axis, and Z-axis coordinates of the first vein vector;

[0161] x g0 , y g0 , z g0 are the X-axis, Y-axis, and Z coordinates of the first vein point of the leaf, respectively.

[0162] Step4.2: Generate a vector V gi (x gi , y gi , z gi ) from the i-th vein point C gi+1 (x gi+1 , y gi+1 , z gi+1 ) to the (i + 1)-th vein point C i (x[[ID=4​​​​​​​​​​​​​​​​​​​​​​​​​​​​

[0168] In the formula:

[0169] A[m], B[m], C[m], D[m] are the parameters of the m-th set of plane equations to be solved;

[0170] x i ,y i ,z i Let X, Y, and Z be the coordinates of the i-th leaf vein vector.

[0171] x gi ,y gi ,z gi Let X, Y, and Z be the coordinates of the i-th leaf vein point;

[0172] H is the dividing interval, and its unit is m;

[0173] H0 is the initial segmentation distance, and its unit is meters.

[0174] l i The length of the leaf segment between the i-th and (i+1)-th leaf vein points is expressed in meters (m).

[0175] Step 4.7: Calculate the first segmentation plane after the i-th leaf vein point according to formula (4) and update the segmentation distance so that m++. Calculate the second segmentation plane after the i-th leaf vein point according to formula (5) and update the segmentation distance so that m++.

[0176]

[0177] In the formula: the meaning of each symbol is the same as that of formula (4).

[0178] Step 4.8: Increment i, calculate the segmentation plane after the next leaf vein point, and execute step 4.4.

[0179] Step 4.9: All segmentation planes have been calculated and stored in four parameter arrays. The obtained plane equations are used to extract the blade point cloud. For each set of parameters (i.e., m0 = 0, 1, 2...m, where m is the number of planes obtained in the previous step), the X-axis, Y-axis, and Z-axis coordinates of the blade point cloud are substituted into formula (6) in sequence. If the conditions are met, the result is exported and stored as a new point cloud, i.e., the blade cross-section curve point cloud.

[0180]

[0181] In the formula:

[0182] x, y, z refer to the X-axis, Y-axis, and Z-axis coordinates of all points on the blade section curve substituted into the inequality;

[0183] A[m0], B[m0], C[m0], D[m0] are the corresponding parameters of all the obtained plane equations, m0 = 0, 1, 2...m, where m is the number of planes obtained in the above steps.

[0184] Step 4.10: End the process of equidistantly cropping the leaf point cloud.

[0185] Step 5: Measurement of the average actual leaf curl index

[0186] The flowchart of the algorithm for measuring the average actual leaf curl index is as follows: Figure 7 As shown, the specific steps are as follows:

[0187] Step 5.1: Select the leaf cross-section curve point cloud between segment 4 (6cm from the leaf tip) and segment 13 (24cm from the leaf tip) as the study area; perform a global search to find the two points furthest apart, calculate their Euclidean distance, and obtain Ln1 and Ln2, which are equidistantly distributed along the leaf vein direction. Here, Ln1 represents the naturally curled leaf width when the maize leaf is naturally unfolded, and Ln2 represents the drought-stressed curled leaf width when the maize leaf is most curled. Taking the calculation results of segment 4 (6cm from the leaf tip) to segment 13 (24cm from the leaf tip) of five maize leaves in the moderate drought treatment group as an example, for ease of expression, the length unit of leaf width is converted to cm, as shown in Table 3:

[0188] Table 3 Calculation results for five leaf blades Ln1 (8 points) and Ln2 (14 points).

[0189]

[0190] Step 5.2: For sample set 1 (corn leaf point cloud at 8 am), process the point cloud of each segment of the leaf cross-section curve. Start from one point and sort using a greedy algorithm based on the shortest distance, ending at another point, so that the disordered point cloud becomes an ordered point cloud.

[0191] Step 5.3: Fit the cross-sectional curve using a cubic spline curve with piecewise parameters. Set the smoothing parameters for the pieces according to formula (7). Use 1000 interpolation points to interpolate the fitted curve. Calculate the distance between each pair of interpolation points and sum them to calculate the length of the spline curve fitting. That is, calculate the fully unfolded leaf width Lw1 of the maize leaf when it is naturally spread out. Taking the Lw1 calculation results of the 4th segment (6cm from the leaf tip) to the 13th segment (24cm from the leaf tip) of the maize leaf in the moderate drought treatment group as an example, for ease of expression, the length unit of the leaf width is converted to cm, as shown in Table 4:

[0192]

[0193] In the formula:

[0194] smooth_param is the smoothing parameter for fitting cubic spline curves;

[0195] Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaf is naturally unfolded in the morning, and its unit is m.

[0196] Table 4 Measurement results of Lw1 for five blades

[0197]

[0198] Note: The leaf number is composed of the corn plant number and which leaf is on the plant from the bottom up. For example, 1-3ye8 represents the 8th leaf of corn plant number 1-3. Step 5.4: If there is a segment of the leaf cross-section curve point cloud where the fully unfolded leaf width Lw1 is smaller than both the front and back, then there is an abnormality in the segment of the leaf cross-section curve point cloud where there is a local missing value. The calculation result is invalid. Record the position of the segment and remove it.

[0199] Step 5.5: For the corn leaves at 2 PM, there's no need to fit a cross-sectional curve to calculate the length. Instead, use the calculated length Lw1 from the point cloud cross-sectional curve of the corn leaves at 8 AM as the fully flat width of the leaf. This is because the leaves are more curled and have a more complex shape, which is not conducive to point cloud processing and fitting a better cross-sectional curve. Furthermore, the variation in fully flat leaf width caused by leaf growth in the morning and afternoon of the same day is negligible. Similarly, for the point cloud processing of the leaves at 2 PM, invalid values ​​corresponding to those at 8 AM should also be removed.

[0200] Step 5.6: Select the leaf section curve points from the cloud computing results after removing invalid values, and calculate the leaf curling index LRI2 (due to drought stress at noon) and the leaf curling index LRI1 (due to natural morning curling) at that location according to formula (8). The actual leaf curling index LRI due to drought stress is calculated. r The difference between the two is then calculated using the effective LRI. r Divide the sum by its number to get The calculation formula is shown in formula (9).

[0201]

[0202] In the formula,

[0203] LRI1 is the leaf curling index of a corn leaf when it is naturally unfolded, corresponding to the point cloud of a certain segment of the leaf cross-section curve.

[0204] LRI2 is the leaf curling index corresponding to the point cloud of a certain segment of the leaf cross-section curve when the corn leaf is most curled.

[0205] Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaves are naturally unfolded in the morning, and its unit is m.

[0206] Lw1 is the full flat width of a corn leaf when it is naturally spread out, which is the length of the cross-sectional curve of the leaf when it is naturally spread out in the morning, and its unit is m.

[0207] Ln2 is the width of corn leaves curled under drought stress when the leaves are most curled, that is, the width of corn leaves when they are most curled under drought stress, and its unit is m;

[0208] LRI r The actual leaf curling index caused by drought stress corresponds to the point cloud of a certain segment of the leaf cross-section curve.

[0209] The average actual leaf curl index;

[0210] q represents the total number of effective leaf curl indices under actual drought stress.

[0211] Table 5 LRI of 25 Corn Leaves r and

[0212]

[0213] Note: "-" indicates an abnormal situation and the measurement data has been removed. The last column is the sum of the previous valid values ​​divided by the number of valid values. The darker the background color in the table, the more water the group receives.

[0214] Step 6: Assessment of Drought Stress Level

[0215] Analyzing Table 5 of the experimental results, the average actual leaf curling index of the particularly drought group and the severe drought group is shown. Similar values, all greater than 30%, were combined for analysis. Moderate drought group. Between 15% and 30%, mild drought group Between 10% and 15% in the non-drought group Below 10%, as shown in Table 6:

[0216] Table 6. Comparison of Average Actual Leaf Curl Index with Drought Stress Level Classification

[0217]

[0218] This study quantified the degree of leaf curling in maize leaves that were not fully expanded during the large trumpet stage using the average actual leaf curling index under different drought stress conditions, and established a model for evaluating the degree of drought stress in maize based on the leaf curling index.

[0219] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring the leaf curling index of maize and evaluating the degree of drought stress, characterized in that, The method includes the following steps: Step 1: Use ground-based lidar to collect point cloud data of maize plants at two times of day, morning and noon, under different drought stress levels during the large trumpet stage; Among them, different drought stress levels include five drought stress levels: no drought, mild drought, moderate drought, severe drought, and extreme drought; point cloud data of maize leaves were collected at two times, one in the morning and one at noon, when the maize leaves were naturally extended and the other when the maize leaves were most curled, and these were used as sample set 1 and sample set 2, respectively. Step 2: Threshold segmentation is performed on the population corn point clouds of the two collected sample sets to obtain the point cloud of a single corn plant. Then, manual stem and leaf segmentation is performed on each single corn plant point cloud to obtain the stem point cloud and the leaf point cloud. The leaf point cloud of at least one corn leaf at the top of the corn plant is selected and the leaf point cloud is optimized by smoothing, filtering, upsampling and downsampling to obtain the optimized smooth, uniform leaf point cloud with no obvious noise. Step 3: Use an increasingly larger cylinder to cut the leaf point cloud to obtain small segments of leaf point cloud. Find the geometric center of each small segment of leaf point cloud to obtain the leaf vein points. Step 4: Based on the leaf vein points, the corn leaf is divided into equidistant plane segments from the leaf tip to the base to obtain multiple leaf cross-section curve point clouds. Step 5: Calculate the Euclidean distance between the two furthest points in the leaf cross-section curve point clouds of each of the two sample sets. This yields the naturally curled leaf width Ln1 when the maize leaf is naturally extended and the drought-stressed curled leaf width Ln2 when the maize leaf is most curled, both distributed at equal intervals along the leaf vein direction. Using a greedy sorting algorithm based on the shortest distance and a cubic spline curve fitting algorithm, fit the leaf cross-section curve point cloud of sample set 1 to calculate the spline curve fitting length, thus obtaining the fully extended leaf width Lw1 when the maize leaf is naturally extended. Based on the naturally curled leaf width Ln1, the drought-stressed curled leaf width Ln2, and the fully extended leaf width Lw1, calculate the leaf curl index LRI1 and LRI2 corresponding to each segment of the leaf cross-section curve point cloud. Finally, calculate the average actual leaf curl index. Step 6: Assessment of drought stress severity.

2. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, The specific steps for Step 3 are as follows: Step 3.1: Read in the point cloud of corn leaves and stalks, and obtain the maximum and minimum values ​​of the X, Y, and Z axis coordinates of the stalk point cloud. smax y smax z smax x smin y smin z min Obtain the minimum Z-coordinate value z of the leaf point cloud. lmin L, the point with the largest Z coordinate zmax (x lmax ,y lmax ,z lmax ); Step 3.2: The coordinates of a point on the axis of the cylinder being cut are C. cylinder (x cylinder ,y cylinder ,z cylinder ), calculate its coordinates according to formula (1); In the formula, x cylinder y cylinder z cylinder x represents the X-axis, Y-axis, and Z-axis coordinates of a point on the cylindrical axis; smax x smin y smax y smin These represent the maximum and minimum values ​​of the X-axis and Y-axis coordinates of the stem point cloud, respectively. z lmin The minimum Z-coordinate value of the leaf point cloud; Step 3.3: Calculate the projection vector G(x) of the leaf growth direction onto the XOY plane. lmax -x cylinder ,y lmax -y cylinder ,0), the cylindrical axial vector is the perpendicular vector of the growth direction vector K(1,k,0), where k is calculated according to formula (2); In the formula, x cylinder y cylinder x and y are the X and Y coordinates of a point on the cylindrical axis, respectively; x lmax y lmax , , are the X and Y coordinates of the point with the largest Z coordinate in the leaf point cloud, respectively; k is the Y coordinate of the perpendicular vector to be obtained; Step 3.4: Using vector K and a point C on the cylinder axis... cylinder Determine the axis of the cylinder, set the cylinder radius to 0.02*j (in meters), obtain the blade point cloud within the cylinder, and save it as a blade annular point cloud PC. cj Other leaf point clouds are updated to new leaf point clouds, j starts from 1, j = 1, 2, 3, ... until all leaf point clouds are divided by cylinders from smallest to largest; Step 3.5: Read in all the leaf ring point clouds, calculate the three-dimensional coordinates of the geometric center point of the bounding box of all the leaf ring point clouds, and the three-dimensional coordinates of the geometric center point are the coordinates of the leaf vein points. Step 3.6: Find the point L with the largest Z-coordinate in the leaf point cloud. zmax (x lmax ,y lmax ,z lmax The three-dimensional coordinates of the center points of the leaf rings obtained above are sorted in reverse order and stored sequentially in a new point cloud, forming all the leaf vein points C from the tip to the base of the leaf. gi , i = 0, 2, 3, ..., n-1, where n is the total number of leaf veins.

3. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, Step 4 involves the following specific steps: Step 4.1: Let the plane equation of the cross section be Ax + By + Cz + D = 0. The four parameters of the plane equation form four arrays A[], B[], C[], D[]; In the first set of parameters of the plane equation, A[0], B[0], and C[0] are the X-axis, Y-axis, and Z-axis coordinates of the first leaf vein vector V0(x0, y0, z0), respectively. The first point of the plane equation passes through the first leaf vein point, that is, the point L with the largest Z-coordinate in the leaf point cloud. zmax The initial A[0], B[0], C[0], D[0] are calculated according to formula (3), and the subsequent parameters are calculated from step 4.2 to step 4.

8. In the formula: A[0],B[0],C[0],D[0] are the parameters of the first set of plane equations; x0, y0, z0 are the X-axis, Y-axis, and Z-axis coordinates of the first leaf vein vector; x g0 ,y g0 ,z g0 These are the X-axis, Y-axis, and Z-coordinates of the first vein point on the leaf, respectively. Step 4.2: Generate the leaf vein from the i-th point C gi (x gi ,y gi ,z gi ) to the (i+1)th leaf vein point C gi+1 (x gi+1 ,y gi+1 ,z gi+1 The vector V i (x i ,y i ,z i () represents the leaf vein vector, and the distance between two adjacent leaf vein points constitutes a small segment of leaf length l. i , i = 0, 2, 3, ..., n-2, where n is the total number of leaf veins; Step 4.3: Set the segmentation interval H = 0.02, the initial segmentation distance H0 = 0.02, and the initial index m = 1 for the parameters of the plane equation to be determined; Step 4.4: Starting from i = 1, if i > n, execute step 4.9; otherwise, execute step 4.

5. Step4.5: If l i <H0, it means there is no segmentation plane behind this vein point. Update the segmentation distance H = H - l i , and then execute Step4.8; otherwise, execute step4.6; Step4.6: If l i < H + H0, it indicates that there is a segmentation plane behind this vein point. Calculate its parameters according to formula (4) and update the segmentation distance H, then execute Step4.

8. Otherwise, it indicates that there are two segmentation planes behind this vein point, and execute step4.7; In the formula: A[m], B[m], C[m], D[m] are the parameters of the m-th set of plane equations to be solved; x i ,y i ,z i Let X, Y, and Z be the coordinates of the i-th leaf vein vector. x gi ,y gi ,z gi Let X, Y, and Z be the coordinates of the i-th leaf vein point; H is the dividing interval, and its unit is m; H0 is the initial segmentation distance, and its unit is meters. l i The length of the leaf segment between the i-th and (i+1)-th leaf vein points is expressed in meters (m). Step 4.7: Calculate the first segmentation plane after the i-th leaf vein point according to formula (4) and update the segmentation distance so that m++. Calculate the second segmentation plane after the i-th leaf vein point according to formula (5) and update the segmentation distance so that m++. In the formula: the meaning of each symbol is the same as that in formula (4); Step 4.8: Increment i, calculate the segmentation plane after the next leaf vein point, and execute step 4.4; Step 4.9: All the dividing planes have been calculated and stored in four parameter arrays. Use the obtained plane equations to extract the blade point cloud. For each set of parameters, i.e. m0 = 0, 1, 2...m, m is the number of planes obtained in the above steps. Substitute the X-axis, Y-axis and Z-axis coordinates of the blade point cloud into formula (6) in sequence. If the conditions are met, export and store it as a new point cloud, i.e., the blade section curve point cloud. In the formula: x, y, z refer to the X-axis, Y-axis, and Z-axis coordinates of all points on the blade section curve substituted into the inequality; A[m0], B[m0], C[m0], D[m0] are the corresponding parameters of all the obtained plane equations, m0 = 0, 1, 2...m, where m is the number of planes obtained in the above steps; Step 4.10: End the process of equidistantly cropping the leaf point cloud.

4. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, In Step 5, the leaf cross-section curve point cloud between the 4th and 13th segments is selected, that is, the leaf range from 6cm to 24cm from the leaf tip, which is 18cm from the middle of the leaf, as the study area.

5. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, Step 5 uses a cubic spline curve with segmented parameters to fit the cross-sectional curve. The smoothing parameters of the segments are set according to formula (7). The fitted curve is represented by 1000 interpolation points. The distance between each pair of interpolation points is calculated and the spline curve fitting length is accumulated. That is, the fully flat leaf width Lw1 when the corn leaf is naturally spread out is calculated. In the formula: smooth_param is the smoothing parameter for fitting cubic spline curves; Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaf is naturally unfolded in the morning, and its unit is m.

6. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, In Step 5, the fully flat leaf width of sample set 2 is calculated using the fully flat leaf width Lw1 of the corn leaf when it is naturally spread out, as calculated in sample set 1.

7. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, In Step 5, if the fully extended leaf width Lw1 of a certain segment of the leaf cross-section curve point cloud in Sample Set 1 is smaller than both the preceding and following segments, then this segment is an abnormal case of missing leaf parts. The leaf cross-section curve point cloud data of this segment in Sample Set 1 is removed, and the corresponding segment of the leaf cross-section curve point cloud data of the corresponding leaf in Sample Set 2 is also removed.

8. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, In Step 5, calculate the leaf curling index LRI1 when the corn leaves are naturally extended, the leaf curling index LRI2 when the corn leaves are most curled, and the average actual leaf curling index. The formula is as follows: In the formula, LRI1 is the leaf curling index of a corn leaf when it is naturally unfolded, corresponding to the point cloud of a certain segment of the leaf cross-section curve. LRI2 is the leaf curling index corresponding to the point cloud of a certain segment of the leaf cross-section curve when the corn leaf is most curled. Ln1 is the width of the naturally curled leaf when the corn leaf is naturally unfolded, which is the Euclidean distance between the farthest points of the leaf cross-section curves when the leaves are naturally unfolded in the morning, and its unit is m. Lw1 is the full flat width of a corn leaf when it is naturally spread out, which is the length of the cross-sectional curve of the leaf when it is naturally spread out in the morning, and its unit is m. Ln2 is the width of the drought-stressed curled leaf when the corn leaf is most curled, which is the Euclidean distance between the farthest points of the cross-sectional curves of the corn leaf when it is most curled under drought stress, and its unit is m. LRI r The actual leaf curling index caused by drought stress corresponds to the point cloud of a certain segment of the leaf cross-section curve. The average actual leaf curl index; q represents the total number of effective leaf curl indices under actual drought stress.

9. The method for measuring maize leaf curling index and evaluating drought stress according to claim 1, characterized in that, In Step 6, the drought stress severity grading evaluation criteria are as follows: Average actual leaf curl index If the drought stress level is greater than 30%, the drought stress level is classified as extreme drought or severe drought. Average actual leaf curl index Between 15% and 30%, the drought stress level is considered moderate drought. Average actual leaf curl index Between 10% and 15%, the drought stress level is considered mild drought; Average actual leaf curl index Below 10%, the drought stress level is considered non-drought.

10. A system for measuring maize leaf curling index and evaluating drought stress, characterized in that, It includes a data acquisition module, a data processing module, and a drought stress assessment module for implementing the method for measuring maize leaf curl index and evaluating drought stress as described in any one of claims 1-9.

Images