Method for measuring leaf area index and measuring device thereof

Abstract

The application provides a leaf area index measurement method, comprising the following steps: step one, image binarization; step two, calculation of canopy gap fraction and contact constant; step three, extraction of measurement point scale G function; step four, extraction of sample plot scale G function; step five, calculation of measurement point scale effective leaf area index; step six, calculation of sample plot scale effective leaf area index; step seven, calculation of aggregation index; and step eight, calculation of sample plot real leaf area index. The application realizes measurement and calculation of vegetation canopy leaf area index by using an intelligent terminal, is convenient for field operation, fully excavates image information of measurement point scale and sample plot scale, can quickly acquire multi-angle gap fraction, solves the problem that the G function does not match the actual canopy situation in the existing method, and improves the precision of the narrow visual field angle terminal device in calculating the LAI.

Description

Technical Field

[0001] This invention relates to a method for low-cost measurement of leaf area index of vegetation, belonging to the field of leaf area index measurement technology, and particularly to a method for measuring leaf area index using a smart terminal. Background Technology

[0002] Leaf Area Index (LAI) is one of the most fundamental parameters characterizing vegetation canopy structure, defined as the sum of the surface areas of all leaves per unit area of ​​ground. LAI is an important indicator for monitoring vegetation growth and change, plant productivity, and health, and is often used as a key input parameter in climate and ecological models. Ground-based LAI measurement methods are divided into direct and indirect methods. Direct methods offer high accuracy but require destructive sampling of vegetation, disrupting the canopy structure and consuming significant manpower and resources. Indirect methods effectively avoid these problems, but the measuring instruments are mostly bulky, complex to operate, and expensive, making them unsuitable for long-term operation. With the improvement of imaging sensor performance and the expansion of software functions in smart terminals, using smart terminals to acquire LAI can improve the efficiency of field measurements of agricultural and forestry parameters and reduce the time and material costs of data acquisition. Currently, several application software programs based on smart terminals for LAI measurement exist, such as PocketLAI, LAISmart, and SmartfLAIr.

[0003] While smart terminals offer portability for field measurements, they also have limitations. Methods for acquiring LAI (Leaf Inclination Angle) using smart terminals still have significant room for improvement in accuracy. First, taking smartphones as an example, their lens field of view is only about 70°, making it impossible to capture vegetation structures within a hemispherical space. This narrow field of view limits the direct application of hemispherical imaging methods to smart terminals. Second, current algorithms for measuring LAI using smart terminals often differ from the actual vegetation structure in their assumptions about the leaf tilt angle (LAD) distribution within the canopy. For instance, LAISmart assumes a spherical canopy distribution and sets the G function to 0.5, while PocketLAI utilizes the fact that the G function is approximately equal to 0.5 when observing a zenith angle of 57°, requiring the smartphone to be tilted 57° to capture the canopy image. Third, existing methods often confuse the concepts of angular gap ratio and average gap ratio, leading to insufficient utilization of multi-angle gap ratios and introducing errors into LAI measurements. Although SmartfLAIr considers multi-angle gap ratios, it requires manual camera rotation to acquire multiple angles, resulting in a long operation time. Finally, the development of smart terminals in LAI measurement has made it possible to obtain multiple images within a single quadrat. However, most methods treat these multiple images as independent of each other and take the average of the multiple images as the quadrat LAI, ignoring the correlation between the multiple images. That is, LAI of the same vegetation type has both spatial differences and similarities in leaf inclination angle.

[0004] To address the aforementioned problems, this invention proposes an improved LAI measurement method based on smart terminals, enabling rapid and accurate extraction of vegetation LAI from images captured by smart terminals. Compared to existing measurement methods utilizing smart terminals, the measurement results of this invention are closer to the true values, better reflect vegetation canopy information, and the measurement process is simple, efficient, and easy to operate, significantly improving the efficiency and accuracy of LAI measurement. Summary of the Invention

[0005] A brief overview of this disclosure is provided below to offer a basic understanding of certain aspects of it. It should be understood that this overview is not an exhaustive summary of the disclosure. It is not intended to identify key or essential parts of the disclosure, nor is it intended to limit its scope. Its purpose is merely to present certain concepts in a simplified form as a prelude to the more detailed description that follows.

[0006] The purpose of this invention is to address the limitations of existing leaf area index (LAI) measurement methods, particularly those utilizing smart terminals. This invention fully leverages image information at both the point-scale (single image) and quadrat-scale (multiple images) perspectives. At the point-scale, it acquires multi-angle gap ratios from digital images captured with a narrow field of view and infers the canopy G-function. At the quadrat-scale, it calculates the quadrat LAI from multiple digital images. By improving the method for measuring canopy LAI using smart terminals, this invention ultimately enhances the accuracy of LAI measurement.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] A method for measuring leaf area index, characterized by comprising the following steps:

[0009] Step 1: Image binarization,

[0010] For the original RGB color image taken from above, the blue band B is extracted as a classification feature to classify the bright and dark pixels of the color image, resulting in a binarized image to distinguish between sky pixels and vegetation pixels.

[0011] Step 2: Calculate the canopy gap ratio and contact constant.

[0012] Based on the field of view of the camera lens of the smart terminal, for a single image, the recorded observed zenith angle is divided into z observed zenith angles at equal intervals on both sides of the principal optical axis.

[0013] (θ i (i = 1, 2, ..., z), based on z observed zenith angles, the binarized image is divided into z concentric circular regions (α...z). i(i = 1, 2, ..., z), for any region α i Calculate the canopy gap ratio P(θ) i ) and contact constant C(θ) i The formula is as follows:

[0014]

[0015] C(θ i )=-cosθ i ln P(θ i (2)

[0016] Where N(θ) i ) Total N(θ) is the total number of pixels within the region. i ) Sky The number of sky pixels;

[0017] Step 3: Extract the scale G function of the measurement point.

[0018] Since the LAI value is fixed at different viewing angles of the same image, and the contact constant and G function have similar curve shapes, the curve shape can be matched using the normalized corner distance matrix algorithm. This allows for the use of the multi-angle contact constant C(θ) in a single image. i The distribution pattern of the G function can be deduced;

[0019] The G-function is a function of the canopy extinction coefficient k(θ) and the observed zenith angle θ. The distribution curves of the G-function under different mean leaf tilt angles (MTA) can be obtained using the following formula:

[0020] G(θ)=k(θ)cosθ (3)

[0021]

[0022] MTA = 9.65(3+χ) -1.65 (5)

[0023] Where θ is the observed zenith angle, k(θ) is the canopy extinction coefficient, and χ is the ratio of the vertical axis length to the horizontal axis length of the ellipsoid of the leaf tilt angle ellipsoidal distribution function;

[0024] For any curve, suppose it has n corner points c1, c2, c3...c n The corresponding corner coordinates are (x1, y1), (x2, y2), (x3, y3)...(x n y n Define the normalized corner distance matrix D as follows:

[0025]

[0026] In the formula,

[0027]

[0028] d max =max(d i，j (8)

[0029]

[0030] Where i = 1, 2...n, j = 1, 2...n, d i，j Represents corner point c i and c j The Euclidean distance, φ i，j This indicates that the Euclidean distance has been normalized.

[0031] For both curve A and curve B, a normalized corner distance matrix D can be calculated. A and D B Then the similarity matrix Φ between curves A and B is... A，B for:

[0032]

[0033] Each element of the matrix D represents the normalized corner distance matrix. A and D B Divide corresponding elements;

[0034] The similarity between two curves is represented by the difference coefficient w. The closer w is to 0, the more similar the two curves are. The calculation formula is as follows:

[0035]

[0036] Curve A represents the contact constant C(θ) corresponding to different zenith angles. i B is the G function for different average leaf tilt angles. Calculate the difference coefficient w between curves A and B. The G function corresponding to the minimum w is the G function of the measuring point scale.

[0037] Step 4: Extract the quadrat scaling G function.

[0038] Assuming that the same vegetation has similar leaf tilt angle distribution patterns at a certain growth stage, the canopy LAI of different images within a quadrat will vary, but the average leaf tilt angle of different images is relatively stable. Based on the G function extracted from a single image in step three, the average leaf tilt angle (MTA) of a single image is calculated. p Then, the MTA of all images within the sample plots was calculated. p The mean value is the average leaf tilt angle (MTA) of the canopy for that quadrat. qIf there are N images of the quadrat, the formula for calculating the average leaf tilt angle of the vegetation observations in the quadrat is:

[0039]

[0040] in, This represents the average leaf tilt angle of N images.

[0041] Furthermore, the MTA q Substituting into formulas (3)-(5) in step three, the G function of the quadrat can be obtained;

[0042] Step 5: Calculate the effective leaf area index at the measuring point scale.

[0043] Without considering the aggregation effect of the canopy, the LAI calculated at this time is called the effective leaf area index (LAI). eff It is the true leaf area index (LAI). tru The effective leaf area index (LAI) of a single image is the product of the aggregation index Ω. eff (picture) Effective Leaf Area Index (LAI) of the z concentric circular regions divided in step two. eff (α i The average value is used to calculate the result, and the formula is as follows:

[0044]

[0045]

[0046] Wherein, G(θ) is the quadrat G function calculated in step four;

[0047] Step 6: Calculate the effective leaf area index at the quadrat scale.

[0048] Multiple images taken within the quadrat are spatially random samples; therefore, the effective leaf area index (LAI) of the entire quadrat is... eff (quadrat) consists of multiple single-image LAIs eff (picture) Average calculation yields:

[0049]

[0050] in, This represents the average effective leaf area index of multiple images taken in a quadrat.

[0051] Step 7: Calculate the aggregation index Ω.

[0052] The vegetation canopy is mostly non-randomly distributed and affected by aggregation effects. The LAI of the quadrat calculated in step six eff It will underestimate the true leaf area index (LAI) truThe aggregation index is a factor characterizing the degree of aggregation in the canopy and can be used to measure the aggregation level of the laminar flow index (LAI). eff Calibrate for LAI tru The aggregation index Ω is calculated using the finite length averaging method, as shown in the following formula:

[0053]

[0054] Wherein, P(θ) is the multi-angle gap ratio obtained in step two. This is the average of multiple gap rates. It is the average of the logarithms of multiple gap rates;

[0055] Step 8: Calculate the true leaf area index of the quadrat.

[0056] Quadrat Scale LAI tru It is calculated using the following formula:

[0057]

[0058] Among them, LAI eff Ω is the effective leaf area index of the quadrat calculated in step six, and Ω is the aggregation index calculated in step seven.

[0059] Furthermore, in step one, the original RGB color image is obtained by capturing it using a smart terminal.

[0060] Furthermore, the smart terminal is a mobile phone, tablet computer, or laptop computer.

[0061] Furthermore, in step one, the Otsu classification algorithm is used to classify the bright and dark pixels of the color image.

[0062] Furthermore, in step two, the field of view of the smart terminal's lens is 70°, with a 35° display space on each side of the principal optical axis. For a single image, the recorded observation zenith angle (0-35°) is divided into 7 observation zenith angles (θ) at 5° intervals. i (i = 1, 2, ..., 7).

[0063] The present invention also provides a measuring device corresponding to the above-described measuring method.

[0064] Compared with existing measurement methods, the advantages of this invention are:

[0065] 1. Using smart terminals to replace professional instruments and equipment to measure and calculate the leaf area index of vegetation canopy is characterized by low cost, high speed, and simple operation, making it convenient for field operations.

[0066] 2. By fully utilizing image information at both the measurement point scale and the quadrat scale, the gap ratio at multiple angles can be quickly obtained, solving the problem that the G function in existing methods does not match the actual canopy conditions and improving the accuracy of LAI calculation by terminal equipment with narrow field of view. Attached Figure Description

[0067] The specific details of this disclosure are described below with reference to the accompanying drawings, which will facilitate a more readily understanding of the above and other objects, features, and advantages of this disclosure. The drawings are merely for illustrating the principles of this disclosure. The dimensions and relative positions of the elements are not necessarily drawn to scale in the drawings.

[0068] Figure 1 This is a flowchart of the LAI calculation process of this invention.

[0069] Figure 2 This is a schematic diagram of the image after being divided from multiple angles.

[0070] Figure 3a This is a schematic diagram of the experimental area distribution.

[0071] Figure 3b This is a schematic diagram of the measurement route within the experimental area.

[0072] Figure 4 This is a comparison chart of the measurement results of the present invention and LI-3000. Detailed Implementation

[0073] Exemplary aspects of this disclosure will be described below with reference to the accompanying drawings. For clarity and brevity, not all features implementing this disclosure are described in the specification. However, it should be understood that many disclosure-specific decisions can be made in developing any such implementation of this disclosure to achieve the developer's specific goals, and these decisions may vary depending on the specific implementation of this disclosure.

[0074] It should also be noted that, in order to avoid obscuring the contents of this disclosure with unnecessary details, only the measurement schemes closely related to the schemes according to this disclosure are shown in the accompanying drawings, while other details that are not closely related to the contents of this disclosure are omitted.

[0075] It should be understood that this disclosure is not limited to the described embodiments by virtue of the following description with reference to the accompanying drawings. In this document, features may be substituted or borrowed between different embodiments where feasible, and one or more features may be omitted in one embodiment.

[0076] This invention provides a method for measuring the leaf area index (LAI) of vegetation canopy based on a smart terminal. The method involves using a smart terminal to collect images in the field, importing these images into the vegetation canopy LAI algorithm program of this invention, and then calculating the canopy LAI. Preferably, the smart terminal of this invention can be a smartphone, tablet computer, laptop computer, or other portable smart terminal device.

[0077] See Figure 1 To obtain the canopy LAI by processing images of vegetation canopy captured by a smart terminal, the following steps should be followed:

[0078] Step 1: Binarize the image to distinguish between sky pixels and vegetation pixels. Otsu's classification algorithm is preferred for binarization.

[0079] The smart terminal takes a vertically upward shot to obtain a color image (RGB format) of the vegetation canopy. The blue band B is extracted as a classification feature. The RGB image is classified into light and dark pixels according to the Otsu classification algorithm to distinguish between sky pixels (bright) and vegetation pixels (dark), resulting in a binarized image.

[0080] Step 2: For a single image, in order to extract multi-angle information, it is divided into multiple concentric circle regions, and the canopy gap ratio and contact constant are calculated for each region.

[0081] See Figure 2 Based on the field of view of the camera lens of the smart terminal, for a single image, the recorded observed zenith angle is divided into z observed zenith angles (θ) at equal intervals on both sides of the principal optical axis. i (i = 1, 2, ..., z), based on z observed zenith angles, the binarized image is divided into z concentric circular regions (α...z). i (i = 1, 2, ..., z), the preferred smart terminal is a mobile phone, whose lens field of view is approximately 70°, with a 35° display space on each side of the principal optical axis. To utilize multi-angle information, the observation zenith angle (0°-35°) of a single image captured by the mobile phone is divided into 7 observation zenith angles (θ) at 5° intervals, centered on the image center. i (i = 1, 2, ..., 7), based on 7 observed zenith angles, the binarized image is divided into 7 concentric circular regions (α). i (i = 1, 2, ..., 7), for any region α i Calculate the canopy gap ratio P(θ) i ) and contact constant C(θ) i The calculation formula is as follows:

[0082]

[0083] C(θ i )=-cosθi ln P(θ i (2)

[0084] Wherein, N(θ) i ) Total For region α i Total number of pixels, N(θ) i ) Sky For α i Number of pixels in the inner sky.

[0085] This step enables the rapid extraction of multi-angle gap ratios from a narrow field of view, solving the error problem caused by calculating LAI using a single angle in existing technologies.

[0086] Step 3: For a single image, use the normalized corner distance matrix algorithm to calculate the multi-angle contact constant C(θ). i Shape matching is performed with G functions of different average leaf tilt angles to find the optimal G function for each image.

[0087] The G-function is a function of the canopy extinction coefficient k(θ) and the observed zenith angle θ. The distribution curves of the G-function under different mean leaf tilt angles (MTA) can be obtained using the following formula:

[0088] G(θ)=k(θ)cosθ (3)

[0089]

[0090] MTA = 9.65(3+χ) -1.65 (5)

[0091] Where θ is the observed zenith angle, k(θ) is the canopy extinction coefficient, and χ is the ratio of the vertical axis length to the horizontal axis length of the ellipsoidal distribution function of the leaf tilt angle.

[0092] Since the LAI value is fixed at different observation angles for the same image, and the contact constant and G function have similar curve shapes, the curve shape can be matched using the normalized corner distance matrix algorithm. This allows for the use of the multi-angle contact constant C(θ) in a single image. i The distribution pattern of the G function can be deduced.

[0093] For any curve, suppose it has n corner points c1, c2...c n The corresponding corner coordinates are (x1, y1), (x2, y2)...(x n y n Define the normalized corner distance matrix D as follows:

[0094]

[0095] In the formula,

[0096]

[0097] d max =max(d i，j (8)

[0098]

[0099] Where i = 1, 2...n, j = 1, 2...n, d i，j Represents corner point c i and c j The Euclidean distance, φ i，j This indicates that the Euclidean distance has been normalized.

[0100] For both curve A and curve B, a normalized corner distance matrix D can be calculated. A and D B Then the similarity matrix Φ between curves A and B is... A，B for:

[0101]

[0102] Each element of the matrix D represents the normalized corner distance matrix. A and D B Divide corresponding elements.

[0103] The similarity between two curves is represented by the difference coefficient w. The closer w is to 0, the more similar the two curves are. The calculation formula is as follows:

[0104]

[0105] In this invention, curve A represents the contact constant C(θ) corresponding to different zenith angles. i Let B be the G function for different average leaf inclination angles. Calculate the difference coefficient w between curves A and B. The G function corresponding to the minimum w is the G function of the measuring point scale.

[0106] This invention uses a shape matching method to infer the distribution of the G function based on some known contact constants. Through this step, a G function with leaf tilt angle distribution and measurement point scale that is closer to the actual situation of the canopy is obtained, which solves the error caused by the assumptions made about the G function in previous studies.

[0107] Step 4: Utilizing the similarity of the average leaf tilt angle of the canopy in multiple images within the quadrat, average the leaf tilt angles of all images within the quadrat to calculate the average leaf tilt angle θ of the overall vegetation in the quadrat. q Furthermore, the G function of the sample plot was derived.

[0108] Assuming that similar vegetation exhibits similar leaf tilt angle distribution patterns at a given growth stage, the canopy LAI (Leaf Area Index) may differ across images within a quadrat, but the average leaf tilt angle (MTA) remains relatively stable across different images. To fully explore the correlation between multiple images within a quadrat, the average leaf tilt angle (MTA) of a single image is calculated based on the G function extracted in step three. p Then, the MTA of all images within the sample plots was calculated. p The mean value is the canopy mean leaf tilt angle (MTA) of that quadrat. q If there are N images of the quadrat, the formula for calculating the average leaf tilt angle of the vegetation observations in the quadrat is:

[0109]

[0110] in, This represents the average leaf tilt angle of N images.

[0111] Furthermore, the MTA q Substituting into formulas (3)-(5) in step three, the G function of the sample plot can be obtained.

[0112] This step fully explored the correlation between multiple images within the quadrat, obtaining a more stable distribution of vegetation leaf inclination angle and quadrat G function at the quadrat scale, thus solving the problem of insufficient correlation between multiple observation data in existing methods.

[0113] Step 5: Calculate the effective leaf area index (LAI) of the optimal z = 7 concentric circular regions determined in Step 2. eff (α i The average value of the leaf area index (LAI) at the measurement point scale is used to derive the effective leaf area index (LAI). eff (picture)).

[0114] Without considering the aggregation effect of the canopy, the LAI calculated at this time is called the effective leaf area index (LAI). eff It is the true leaf area index (LAI). tru The effective leaf area index (LAI) is the product of the aggregation index (Ω) and the aggregation index (Ω). eff (picture) The effective leaf area index (LAI) is obtained by dividing the area into 7 concentric circles in step two. eff (α i The average value is used to calculate the result, and the formula is as follows:

[0115]

[0116]

[0117] Wherein, G(θ) is the quadrat G function calculated in step four.

[0118] Step 6: Calculate the LAI of all images within the quadrat. eff The average value of the (picture) is used to derive the effective leaf area index (LAI) at the quadrat scale. eff (quadrat)).

[0119] Multiple images taken within the quadrat are spatially random samples; therefore, the effective leaf area index (LAI) of the entire quadrat is... eff (quadrat) can be composed of multiple single images LAI eff (picture) Average calculation yields:

[0120]

[0121] in, This represents the average effective leaf area index of multiple images taken in a quadrat.

[0122] Steps five and six use the G function, which better reflects the actual vegetation canopy distribution, to calculate the LAI at the measuring point scale and the quadrat scale. eff Compared to previous studies, it can significantly improve LAI. eff Precision.

[0123] Step 7: Calculate the aggregation index Ω using the effective length averaging method.

[0124] The vegetation canopy is mostly non-randomly distributed and affected by aggregation effects. The LAI of the quadrat calculated in step six eff It will underestimate the true leaf area index (LAI) tru The aggregation index is a factor characterizing the degree of aggregation in the canopy and can be used to measure the LAI (Laminated Area Index). eff Calibrate for LAI tru This invention calculates the aggregation index Ω using the finite-length averaging method, as shown in the following formula:

[0125]

[0126] Wherein, P(θ) is the multi-angle gap ratio obtained in step two. This is the average of multiple gap rates. It is the average of the logarithms of multiple gap rates.

[0127] Step 8: Calculate the true leaf area index of the quadrat.

[0128] Quadrat Scale LAI tru It can be calculated using the following formula:

[0129]

[0130] Among them, LAI effΩ is the effective leaf area index of the quadrat calculated in step six, and Ω is the aggregation index calculated in step seven.

[0131] The following is an example to further describe the specific implementation of the present invention.

[0132] Experimental area:

[0133] See Figure 3a Within the experimental area, three quadrats were selected for field sampling. Quagmire A was planted with corn, while quadrats B and C were planted with soybeans. The experiment lasted for two months, with sampling occurring weekly. The walking paths and measurement points within the quadrats, captured using smart terminals, are shown below. Figure 3b As shown. After processing using the method of this invention, the true leaf area index (LAI) can be obtained. tru Simultaneously, the LAI (LAI) measured by the LI-3000 instrument was acquired. LI-3000 As the true value, it is related to the LAI of this invention. tru Compare them.

[0134] Comparison of the measurement results of this method with those of LI-3000:

[0135] See Figure 4 The method showed good agreement with the destructive sampling results of LI-3000, with a regression slope of 0.7 and a root mean square error of 0.564. Table 1 shows the LAI values ​​obtained in the three sample areas. tru The LAI value is the improved value obtained based on this method. LI-3000 The measurement results from the LI-3000 show that the absolute deviations of the three sample areas are small, at 0.22, 0.15, and 0.11 respectively. The average deviation between the two methods is only 0.005, indicating that the method has good consistency with the true value.

[0136] Table 1 Comparison of LAI measurement results of this method and LI-3000

[0137]

[0138] The above specific embodiments are only for illustrating the technical concept and structural features of the present invention, and are intended to enable those skilled in the art to implement them. However, the above content does not limit the scope of protection of the present invention. Any equivalent changes or modifications made based on the technical features of the present invention should fall within the scope of protection of the present invention.

[0139] The foregoing description of this disclosure in conjunction with specific implementation schemes is exemplary and not intended to limit the scope of protection of this disclosure. Those skilled in the art can make various modifications and variations to this disclosure based on its spirit and principles, and such modifications and variations are also within the scope of this disclosure.

Claims

1. A method for measuring leaf area index, characterized in that, Includes the following steps: Step 1: Image binarization, For the original RGB color image taken from above, the blue band B is extracted as a classification feature to classify the bright and dark pixels of the color image, resulting in a binarized image to distinguish between sky pixels and vegetation pixels. Step 2: Calculate the canopy gap ratio and contact constant. Based on the field of view of the camera lens of the smart terminal, for a single image, the recorded observed zenith angle is divided into z observed zenith angles θ at equal intervals on both sides of the principal optical axis. i Let i = 1, 2, ..., z. Based on z observed zenith angles, the binarized image is divided into z concentric circular regions α. i Let i = 1, 2, ..., z, for any region α i Calculate the canopy gap ratio P(θ) i ) and contact constant C(θ) i The formula is as follows: C(θ i )=-cosθ i lnP(θ i ) (2) Where N(θ) i ) Total N(θ) is the total number of pixels within the region. i ) Sky The number of sky pixels; Step 3: Extract the scale G function of the measurement point. Since the LAI value is fixed at different viewing angles of the same image, and the contact constant and G function have similar curve shapes, the curve shape can be matched using the normalized corner distance matrix algorithm. This allows for the use of the multi-angle contact constant C(θ) in a single image. i The distribution pattern of the G function can be deduced; The G-function is a function of the canopy extinction coefficient k(θ) and the observed zenith angle θ. The distribution curves of the G-function under different mean leaf tilt angles (MTA) can be obtained using the following formula: G(θ)=k(θ)cosθ (3) MTA=9.65(3+χ) -1.65 (5) Where θ is the observed zenith angle, k(θ) is the canopy extinction coefficient, and χ is the ratio of the vertical axis length to the horizontal axis length of the ellipsoid of the leaf tilt angle ellipsoidal distribution function; For any curve, suppose it has n corner points c1, c2, c3...c n The corresponding corner coordinates are (x 1, y1), (x 2, y2), (x 3, y3)…(x n, y n Define the normalized corner distance matrix D as follows: In the formula, d max =max(d i,j ) (8) Where i = 1, 2, ..., n, j = 1, 2, ..., n, d i,j Represents corner point c i and c j The Euclidean distance, φ i,j This indicates that the Euclidean distance has been normalized. For both curve A and curve B, a normalized corner distance matrix D can be calculated. A and D B Then the similarity matrix Φ between curves A and B is... A,B for: Each element of this matrix represents the normalized corner distance matrix D. A and D B Divide corresponding elements; The similarity between two curves is represented by the difference coefficient w. The closer w is to 0, the more similar the two curves are. The calculation formula is as follows: Curve A represents the contact constant C(θ) corresponding to different zenith angles. i B is the G function for different average leaf tilt angles. Calculate the difference coefficient w between curves A and B. The G function corresponding to the minimum w is the G function of the measuring point scale. Step 4: Extract the quadrat scaling G function. Assuming that the same vegetation has similar leaf tilt angle distribution patterns at a certain growth stage, the canopy LAI of different images within a quadrat will vary, but the average leaf tilt angle of different images is relatively stable. Based on the G function extracted from a single image in step three, the average leaf tilt angle (MTA) of a single image is calculated. p Then, the MTA of all images within the sample plots was calculated. p The mean value is the average leaf tilt angle (MTA) of the canopy for that quadrat. q If there are N images of the quadrat, the formula for calculating the average leaf tilt angle of the vegetation observations in the quadrat is: in, This represents the average leaf tilt angle of N images; Furthermore, the MTA q Substituting into formulas (3)-(5) in step three, the G function of the quadrat can be obtained; Step 5: Calculate the effective leaf area index at the measuring point scale. Without considering the aggregation effect of the canopy, the LAI calculated at this time is called the effective leaf area index (LAI). eff It is the true leaf area index (LAI). tru The effective leaf area index (LAI) of a single image is the product of the aggregation index Ω. eff (picture) Effective Leaf Area Index (LAI) of the z concentric circular regions divided in step two. eff (α i The average value is used to calculate the result, and the formula is as follows: Wherein, G(θ) is the quadrat G function calculated in step four; Step 6: Calculate the effective leaf area index at the quadrat scale. Multiple images taken within the quadrat are spatially random samples; therefore, the effective leaf area index (LAI) of the entire quadrat is... eff (quadrat) consists of multiple single-image LAIs eff (picture) Average calculation yields: in, This represents the average effective leaf area index of multiple images taken in a quadrat. Step 7: Calculate the aggregation index Ω. The vegetation canopy is mostly non-randomly distributed and affected by aggregation effects. The LAI of the quadrat calculated in step six eff It will underestimate the true leaf area index (LAI) tru The aggregation index is a factor characterizing the degree of aggregation in the canopy and can be used to measure the aggregation level of the laminar flow index (LAI). eff Calibrate for LAI tru The aggregation index Ω is calculated using the finite length averaging method, as shown in the following formula: Wherein, P(θ) is the multi-angle gap ratio obtained in step two. This is the average of multiple gap rates. It is the average of the logarithms of multiple gap rates; Step 8: Calculate the true leaf area index of the quadrat. True Leaf Area Index (LAI) tru It is calculated using the following formula: Among them, LAI eff Ω is the effective leaf area index of the quadrat calculated in step six, and Ω is the aggregation index calculated in step seven.

2. The leaf area index measurement method according to claim 1, characterized in that, In step one, the original RGB color image is obtained by capturing it using a smart terminal.

3. The leaf area index measurement method according to claim 2, characterized in that, The smart terminal is a mobile phone, tablet computer, or laptop computer.

4. The leaf area index measurement method according to claim 1, characterized in that, In step one, the Otsu classification algorithm is used to classify the bright and dark pixels of the color image.

5. The leaf area index measurement method according to claim 1, characterized in that, In step two, the field of view of the smart terminal's lens is 70°, with a 35° display space on each side of the principal optical axis. For a single image, the recorded 0-35° observation zenith angle is divided into 7 observation zenith angles θ at 5° intervals. i , i = 1, 2, ..., 7.

6. A leaf area index measuring device, characterized in that, Includes the following modules: (1) Image binarization module, For the original RGB color image taken from above, the blue band B is extracted as a classification feature to classify the bright and dark pixels of the color image, resulting in a binarized image to distinguish between sky pixels and vegetation pixels. (2) Module for calculating canopy gap ratio and contact constant. Based on the field of view of the camera lens of the smart terminal, for a single image, the recorded observed zenith angle is divided into z observed zenith angles θ at equal intervals on both sides of the principal optical axis. i Let i = 1, 2, ..., z. Based on z observed zenith angles, the binarized image is divided into z concentric circular regions α. i Let i = 1, 2, ..., z, for any region α i Calculate the canopy gap ratio P(θ) i ) and contact constant C(θ) i The formula is as follows: C(θ i )=-cosθ i lnP(θ i ) (2) Where N(θ) i ) Total N(θ) is the total number of pixels within the region. i ) Sky The number of sky pixels; (3) Measurement point scale G function extraction module Since the LAI value is fixed at different viewing angles of the same image, and the contact constant and G function have similar curve shapes, the curve shape can be matched using the normalized corner distance matrix algorithm. This allows for the use of the multi-angle contact constant C(θ) in a single image. i The distribution pattern of the G function can be deduced; The G-function is a function of the canopy extinction coefficient k(θ) and the observed zenith angle θ. The distribution curves of the G-function under different mean leaf tilt angles (MTA) can be obtained using the following formula: G(θ)=k(θ)cosθ (3) MTA=9.65(3+χ) -1.65 (5) Where θ is the observed zenith angle, k(θ) is the canopy extinction coefficient, and χ is the leaf tilt angle ellipsoid. The ratio of the vertical axis length to the horizontal axis length of the ellipsoid of the distribution function; For any curve, suppose it has n corner points c1, c2, c3...c n Its corresponding corner point Coordinates are (x 1, y1), (x 2, y2), (x 3, y3)…(x n, y n Define the normalized corner distance matrix D as follows: In the formula, d max =max(d i,j ) (8) Where i = 1, 2, ..., n, j = 1, 2, ..., n, d i,j Represents corner point c i and c j The Euclidean distance, φ i,j This indicates that the Euclidean distance has been normalized. For both curve A and curve B, a normalized corner distance matrix D can be calculated. A and D B Then the similarity matrix Φ between curves A and B is... A,B for: Each element of this matrix represents the normalized corner distance matrix D. A and D B Divide corresponding elements; The similarity between two curves is represented by the difference coefficient w. The closer w is to 0, the more similar the two curves are. The calculation formula is as follows: Curve A represents the contact constant C(θ) corresponding to different zenith angles. i B is the G function for different average leaf tilt angles. Calculate the difference coefficient w between curves A and B. The G function corresponding to the minimum w is the G function of the measuring point scale. (4) G-function extraction module for quadrat scale Assuming that similar vegetation exhibits similar leaf tilt angle distribution patterns at a given growth stage, the canopy LAI (Leaf Area Inclination) of different images within a quadrat will vary, but the average leaf tilt angle (MTA) of different images will remain relatively stable. Based on the extracted G function of a single image, the average leaf tilt angle (MTA) of that single image is calculated. p Then, the MTA of all images within the sample plots was calculated. p The mean value is the average leaf tilt angle (MTA) of the canopy for that quadrat. q If there are N images of the quadrat, the formula for calculating the average leaf tilt angle of the vegetation observations in the quadrat is: in, This represents the average leaf tilt angle of N images; Furthermore, the MTA q Substituting into formulas (3)-(5), the G function of the sample plot can be obtained; (5) Effective leaf area index calculation module at measuring point scale Without considering the aggregation effect of the canopy, the LAI calculated at this time is called the effective leaf area index (LAI). eff It is the true leaf area index (LAI). tru The effective leaf area index (LAI) of a single image is the product of the aggregation index Ω. eff (picture) Effective Leaf Area Index (LAI) is obtained by dividing the region into z concentric circles. eff (α i The average value is used to calculate the result, and the formula is as follows: Where G(θ) is the quadrat G function; (6) Module for calculating effective leaf area index at the quadrat scale. Multiple images taken within the quadrat are spatially random samples; therefore, the effective leaf area index (LAI) of the entire quadrat is... eff (quadrat) consists of multiple single-image LAIs eff (picture) Average calculation yields: in, This represents the average effective leaf area index of multiple images taken in a quadrat. (7) Aggregation index Ω calculation module, The vegetation canopy is mostly non-randomly distributed and affected by aggregation effects, resulting in calculated LAI (Local Area of ​​Intake) of quadrats. eff It will underestimate the true leaf area index (LAI) tru The aggregation index is a factor characterizing the degree of aggregation in the canopy and can be used to measure the aggregation level of the laminar flow index (LAI). eff Calibrate for LAI tru The aggregation index Ω is calculated using the finite length averaging method, as shown in the following formula: Where P(θ) is the multi-angle gap ratio, This is the average of multiple gap rates. It is the average of the logarithms of multiple gap rates; (8) Module for calculating the true leaf area index of quadrats. True Leaf Area Index (LAI) tru It is calculated using the following formula: Among them, LAI eff Ω represents the effective leaf area index of the quadrat, and Ω represents the aggregation index.

7. The leaf area index measuring device according to claim 6, characterized in that, The original RGB color image was captured using a smart device.

8. The leaf area index measuring device according to claim 7, characterized in that, The smart terminal is a mobile phone, tablet computer, or laptop computer.

9. The leaf area index measuring device according to claim 6, characterized in that, The Otsu classification algorithm is used to classify the bright and dark pixels in a color image.

10. The leaf area index measuring device according to claim 6, characterized in that, The smart terminal's lens has a field of view of 70°, with 35° of display space on each side of the principal optical axis. For a single image, the recorded 0-35° observation zenith angle is divided into 7 observation zenith angles θ at 5° intervals. i , i = 1, 2, ..., 7.

Images