A directional component proportion calculation method for a complex vegetation scene

Abstract

The present application belongs to the technical field of thermal infrared vegetation observation, and particularly relates to a directional component proportion calculation method for complex vegetation scene. The present application firstly uses high-resolution remote sensing image to obtain classification data; then combines vegetation structure parameters in the target area to establish a simplified three-dimensional scene model by taking a cube as a representative of vegetation. Through computer simulation of the three-dimensional model, the overlapping area of the two projections along the light and observation directions on the ground is obtained. With the help of vegetation structure parameters and the simplified three-dimensional model, the average area, effective projection area and light and shadow proportion of the projection of each type of vegetation on the ground are derived through geometric relationship. The average projection area is brought into the calculation of the Boolean model to obtain the porosity. Finally, the porosity, light and shadow proportion and overlapping area are used to calculate the proportion of the component. The present application can promote the researches on the directional radiation of sparse vegetation scene, satellite remote sensing ground temperature verification and ecological environment monitoring.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of thermal infrared vegetation observation, and particularly relates to a directional component proportion calculation method for a complex vegetation scene. BACKGROUND

[0002] The land surface temperature is an important parameter for representing the radiation characteristics of the land surface and the energy balance process between the land surface and the atmosphere. The land surface temperature is widely used in the research fields of climate, environment, agriculture and the like. Satellite remote sensing can quickly obtain the land surface temperature of a large range due to its own wide observation range. Domestic and foreign research institutions have published a variety of land surface temperature products. At present, these satellite remote sensing land surface temperature products have achieved good application effects in multiple research fields. However, the satellite remote sensing land surface temperature still faces problems such as difficulty in verifying the accuracy of the complex vegetation surface and difficulty in coordinating the multi-source land surface temperature. One of the important reasons leading to these problems is that the land surface temperature has directionality under the heterogeneous surface. When the satellite sensor observes the same complex vegetation region from different observation directions, the component proportions in the field of view may be obviously different, and thus the observation results may be obviously different. Therefore, it is particularly important to obtain the component proportions in the field of view of the satellite in the process of studying the directionality of the satellite land surface temperature.

[0003] In existing researches, the vegetation surface is usually simplified as a simple geometric model with random distribution or a homogeneous vertical layered model to estimate the visible component proportions in the field of view of the sensor. Since the 1970s, in order to study the directional radiation energy or reflectivity of the scene with three-dimensional structure, researchers began to use simple geometric bodies to replace the vegetation in the real scene to establish a geometric model, and then calculated the proportions of various components in the field of view according to the light and shadow conditions. For dense vegetation, researchers regard the vegetation canopy temperature as a simplified model with vertical direction layering, horizontal direction equality and certain porosity of each layer of the canopy, and then calculate the visible component proportions by using the Boolean model. The geometric optical method using the simple geometric model usually ignores the actual distribution of the trees, and brings more uncertainty to the calculation results of the component proportions. The method using the homogeneous vertical layered model is usually difficult to apply to the unevenly distributed sparse vegetation surface. Although researchers try to introduce the aggregation index to improve the applicability of these models to the sparse vegetation surface, it is still difficult to accurately describe the influence of the plant distribution position on the component proportions. SUMMARY

[0004] In order to accurately obtain the component proportions of the complex vegetation surface in the satellite observation field of view, the directional component proportion calculation method for the complex vegetation scene is provided. The method uses a simplified three-dimensional scene model based on high-resolution classification data and vegetation structure parameters to calculate the projection related parameters, and then realizes the calculation by combining the Boolean model.

[0005] The technical scheme of the application is:

[0006] A directional component ratio calculation method for complex vegetation scenes, comprising the following steps:

[0007] S1, establishing a three-dimensional scene model:

[0008] Through high-resolution remote sensing images to obtain classification data, and then combining the vegetation structure parameters in the target area, a three-dimensional scene model is established by taking a cube as a representative of the vegetation;

[0009] S2, through computer simulation of the three-dimensional model, the overlapping area of the two projections along the light and observation directions on the ground is obtained; at the same time, with the help of the vegetation structure parameters and the three-dimensional model, the average projection area, the effective projection area and the light and shadow ratio of each type of vegetation on the ground are derived through geometric relationship;

[0010] S3, the average projection area is brought into the Boolean model to obtain the porosity;

[0011] S4, using the porosity, the light and shadow ratio and the overlapping area to calculate the component ratio of the light vegetation, the shadow vegetation, the light soil, the shadow soil, the light tree trunk and the shadow tree trunk.

[0012] Further, in S1, the vegetation in the target area is defined to include Tamarix and Populus euphratica, in order to ensure the modeling accuracy as much as possible, the classification data of the target scene with a spatial resolution much higher than the plant size is selected as the basis for modeling when the three-dimensional model is simplified, and a cube with a bottom side length of the spatial resolution of the classification data and a height of the vegetation height is used to establish the vegetation model, and then the entire scene model is reconstructed, in the classification image, the relative height of the soil pixel is 0; Tamarix is a cube with a bottom height of 0 and a top height of h; the tree trunk is a cube with a bottom of 0 and a top height of h1, and the crown is a cube with a bottom height of h1 and a top height of h1+h2.

[0013] Further, the calculation method of the average projection area in S2 is:

[0014] Define the pixel side length as l1, the number of Tamarix inside as N1, the number of Populus euphratica as N2, the side length of a single Tamarix as l2, the crown side length of a single Populus euphratica as l3, the tree trunk side length as l4, the observation zenith angle and azimuth angle as The solar zenith angle and azimuth angle are The included angle between the observation azimuth angle and the solar azimuth angle is In the calculation process, when α>90°, take α=180°-α. Then the projection area of each Tamarix, Populus euphratica crown or tree trunk on the ground is :

[0015]

[0016] Wherein, subscript CL represents Tamarix, subscript HY represents Populus euphratica, and subscript SG represents tree trunk.

[0017] Further, the calculation method of the effective projection area in S2 is as follows:

[0018] If the vegetation is distributed in clusters, the effective area ratio R is:

[0019]

[0020] If the vegetation is evenly distributed, the effective area ratio of the Tamarix and Populus euphratica canopy projections is:

[0021]

[0022] It is defined that when the length of the projection of the plant side on the ground is less than or equal to the average spacing of the plants, the projection of the plant on the ground does not overlap, that is, the effective area ratio is equal to 1.

[0023] Further, the calculation method of the light and shadow ratio in S2 is as follows:

[0024] It is defined that in the three-dimensional model, the cube always has only one side facing the sun, and therefore the light and shadow ratio Q in the projection of the same vegetation in the observation direction needs to consider the angle between the observation azimuth angle and the solar azimuth angle. When the angle is less than 90°, the side light component needs to be considered, and when the angle is greater than or equal to 90°, only the proportion of the top light part needs to be considered. It is defined that the angle less than 90° corresponds to the first quadrant, and the calculation formula is as follows:

[0025]

[0026] The calculation formula of other quadrants is adjusted based on the calculation formula of the first quadrant.

[0027] Further, the calculation method of the porosity in S3 is as follows:

[0028] It is defined that the porosity P of each type of ground object is as follows:

[0029]

[0030] Since there is mutual shielding between Tamarix, tree canopy and tree trunk, the calculation formula is defined as follows:

[0031]

[0032] Wherein, P BG1 is the background porosity of the tree canopy and tree trunk, P BG2 is the background porosity of the Tamarix and tree canopy, and P BG3 is the background porosity of the Tamarix, tree canopy and tree trunk.

[0033] Further, the calculation method of the component ratio in S4 is as follows:

[0034] The light illumination vegetation and the shadow vegetation include light illumination tamarix, shadow tamarix, light illumination populus euphratica, and shadow populus euphratica, and the definition calculation formula is:

[0035]

[0036] Wherein, F gCL , F sCL , F gHY , F sHY , F gSoil , F sSoil , F gSG , F sSG are light illumination tamarix, shadow tamarix, light illumination populus euphratica, shadow populus euphratica, light illumination soil, shadow soil, light illumination tree trunk and shadow tree trunk, A O represents the overlapping area of the observation direction and the light direction on the ground projection.

[0037] The beneficial effects of the present application are: the original directional component ratio calculation method often ignores the actual distribution of vegetation or is difficult to apply to unevenly distributed sparse vegetation surface, and is difficult to meet the directional component ratio calculation demand of complex vegetation surface. Using the method of the present application, the calculation accuracy and simplicity can be considered in the case of considering the vegetation distribution and structure parameters, and then the research of sparse vegetation scene directional radiation, satellite remote sensing surface temperature verification and ecological environment monitoring is promoted. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 It is the overall flowchart of the present application.

[0039] Figure 2 It is a high-resolution classification data example.

[0040] Figure 3 It is a simplified three-dimensional scene model schematic diagram.

[0041] Figure 4 It is a directional component ratio example without distinguishing the light condition. DETAILED DESCRIPTION

[0042] The technical scheme of the present application will be described in detail below in combination with the drawings.

[0043] The flow of the present application is as follows Figure 1The classification data were first obtained using high-resolution remote sensing images; then the vegetation structure parameters in the target area were combined to establish a simplified three-dimensional scene model with a cube representing the vegetation. Through computer simulation of the three-dimensional model, the overlapping area on the ground of the two projections along the light and observation directions was obtained. With the help of the vegetation structure parameters and the simplified three-dimensional model, the average area, effective projection area and light and shadow ratio of the projection of each type of vegetation on the ground were derived through geometric relationships. The average projection area was brought into the Boolean model calculation to obtain the porosity. Finally, the porosity, light and shadow ratio and overlapping area were used to calculate the proportions of eight components, i.e. lighted tamarisk, shadowed tamarisk, lighted poplar, shadowed poplar, lighted soil, shadowed soil, lighted tree trunk and shadowed tree trunk.

[0044] The method for establishing a simplified three-dimensional scene model is as follows:

[0045] To ensure the modeling accuracy as much as possible, the classification data of the target scene with a spatial resolution much higher than the plant size were selected as the basis for modeling when the three-dimensional model was simplified. In the modeling process, a cube with a bottom side length of the spatial resolution of the classification data and a height of the vegetation height was used to establish the vegetation model, and then the entire scene model was reconstructed. In the classification image, the relative height of the soil pixel is 0; the tamarisk is a cube with a bottom height of 0 and a top height of h; the tree trunk is a cube with a bottom of 0 and a top height of h1, and the crown is a cube with a bottom height of h1 and a top height of h1+h2, as shown in Fig. 1. Figure 3

[0046] The method for calculating the average projection area is as follows:

[0047] The pixel side length is l1, the number of tamarisks inside is N1, the number of poplars is N2, the side length of a single tamarisk is l2, the crown side length of a single poplar is l3, the tree trunk side length is l4, the observation zenith angle and azimuth angle are The sun zenith angle and azimuth angle are The included angle between the observation azimuth angle and the sun azimuth angle is In the calculation process, when a>90°, a=180°-a. The projection area of each tamarisk (CL), poplar (HY) crown or tree trunk (SG) on the ground can be represented by formula (1) as follows:

[0048]

[0049] The method for calculating the effective projection area is as follows:

[0050] ​Assuming that the vegetation is distributed in a square area, the plants are close together but do not overlap each other, due to the mutual overlap of the projection, the effective projection area of the actual single plant vegetation is less than the actual projection area, the effective area ratio of the projection of the Tamarix, the tree crown and the tree trunk can be expressed by formula (2), wherein the tree trunk is located directly below the center of the tree crown, so the tree trunk cannot be adjacent to other tree trunks, in addition, the crown width of the tree crown is much larger than the diameter of the tree trunk, and the tree trunk is considered to be always in a non-overlapping state during calculation.

[0051]

[0052] When there are gaps between the plants and the gaps are uniformly distributed, the effective area ratio R of the projection of the Tamarix and the Populus euphratica crown can be expressed by formula (3), with the change of the zenith angle and the azimuth angle, when the length of the plant side on the ground is less than or equal to the average spacing of the plants, the plant does not overlap on the ground, that is, the effective area ratio is equal to 1.

[0053]

[0054] The method for calculating the light and shadow ratio is as follows:

[0055] In the simplified three-dimensional model, the cube always keeps only one side facing the sun, so only the top surface and the side can be illuminated. When calculating the light and shadow ratio Q in the projection of the same vegetation in the observation direction, the included angle between the observation azimuth angle and the solar azimuth angle needs to be considered, when the included angle is less than 90°, the side illumination component needs to be considered, and when the included angle is greater than or equal to 90°, only the proportion of the top surface illumination part needs to be considered. In the case where the included angle is less than 90°, the light and shadow area ratio of the projection of the same vegetation can be calculated by formula (4), when the included angle is greater than 90°, formula (4) needs to be adjusted according to the quadrant in which the angle is located, because the formula expression is similar, only the formula in the first quadrant is listed here:

[0056]

[0057] The method for calculating the porosity is as follows:

[0058] The Boolean model is widely used in vegetation porosity calculation, and the present application uses the model in combination with the simplified three-dimensional model to express the porosity P of each type of ground object alone as follows:

[0059]

[0060] Because there is mutual shielding between the Tamarix, the tree crown and the tree trunk, the background porosity P BG1 of the tree crown and the tree trunk, the background porosity P BG2 of the Tamarix and the tree crown, and the background porosity P BG3 of the Tamarix, the tree crown and the tree trunk can be expressed as shown in formula (6):

[0061]

[0062] The method for calculating the directional component ratio is:

[0063] In the sparse vegetation scene of shrubs and trees mixed, considering the lighted tamarisk, shadow tamarisk, lighted poplar, shadow poplar, lighted soil, shadow soil, lighted tree trunk and shadow tree trunk eight components, the corresponding component ratio F gCL , F sCL , F gHY , F sHY , F gSoil , F sSoil , F gSG , F sSG The calculation method is shown in formula (7), wherein A O represents the overlapping area of the observation direction and the ground projection of the light direction, and the tamarisk and the tree trunk are possibly blocked by the higher tree crown in addition to the mutual blocking.

[0064]

[0065] So far, the method for calculating the directional component ratio based on the classified data and the vegetation structure parameters and the simplified three-dimensional model is established.

[0066] The specific execution steps of the present application include:

[0067] Step 1. Obtain the high-resolution remote sensing image of the target area, and then perform classification Figure 2 ). The ground cover type should include the main vegetation in the target area.

[0068] Step 2. Obtain the vegetation height, crown width and other structure parameters in the target area by sampling measurement and other means.

[0069] Step 3. Use the classified data and the vegetation structure parameters to establish a simplified three-dimensional model. The vegetation in the model is represented by a cube. In order to facilitate parameter derivation, keep the normal vector of a side of the cube and the light vector in the same plane.

[0070] Step 4. Use computer simulation method to calculate the overlapping area of the projection on the ground along the light and observation directions from the simplified three-dimensional model.

[0071] Step 5. Derive the average projection area, effective projection area and light and shadow ratio of each type of vegetation single plant on the ground from the three-dimensional model according to the geometric relationship.

[0072] Step 6. Use the average projection area to substitute into the Boolean model to calculate the porosity of each type of ground object.

[0073] Step 7. Calculate the proportion of each component by combining porosity, the ratio of light and shadow, and the overlapping area Figure 4 The scene in the case mainly contains two kinds of vegetation, shrubs (tamarisk) and trees (poplar), so the proportions of eight components, light tamarisk, shadow tamarisk, light poplar, shadow poplar, light soil, shadow soil, light trunk, and shadow trunk, are calculated as shown in Table 1:

[0074] Table 1. Example of the proportion of directional components

[0075]

Claims

1. A method for calculating the proportion of directional components for a complex vegetation scene, characterized in that The method comprises the following steps: S1, establishing a three-dimensional scene model: Through high-resolution remote sensing image to obtain classification data, and then combining the vegetation structure parameters in the target area, a three-dimensional scene model is established by taking a cube as a representative of the vegetation; the vegetation in the target area is defined to include Tamarix and Populus euphratica, and in the modeling process, a vegetation model is established by using a cube with a bottom side length of the spatial resolution of the classification data and a height of the vegetation height, and then the entire scene model is reconstructed; in the classification image, the relative height of the soil pixel is 0; the Tamarix is a cube with a bottom height of 0 and a top height of h; the tree trunk is a cube with a bottom of 0 and a top height of h1; and the tree crown is a cube with a bottom height of h1 and a top height of h1+h2; S2, obtaining the overlapping area on the ground in two projections along the light and observation directions through computer simulation of the three-dimensional model; and simultaneously, the average projection area, the effective projection area and the light and shadow ratio of each type of vegetation on the ground are derived through geometric relationship by means of the vegetation structure parameters and the three-dimensional model; The calculation method of the average projection area is: The definition of the pixel edge length is l1, the number of interior of the Mallotus philippinensis is N1, the number of Populus euphratica is N2, the single Mallotus philippinensis edge length is l2, the single Populus euphratica crown edge length is l3, the trunk edge length is l4, the observation zenith angle and azimuth angle are (θ v , φ v ), the sun zenith angle and azimuth angle are (θ i , φ i ), the included angle of the observation azimuth angle and the sun azimuth angle is α = |φ v - φ i |, in the calculation process, when α > 90°, α = 180°- α is taken, then the projection area Ā of each Mallotus philippinensis, Populus euphratica crown or trunk on the ground is: ， Wherein, the subscript CL represents the Tamarix, the subscript HY represents the Populus euphratica, and the subscript SG represents the tree trunk; The calculation method of the effective projection area is: If the vegetation is gathered and distributed, the effective area ratio R is: ， If the vegetation is uniformly distributed, the effective area ratio of the Tamarix and the Populus euphratica crown projection is: ， It is defined that when the side surface of the plant in the projection on the ground is less than or equal to the average interval of the plant, the projection of the plant on the ground does not overlap, that is, the effective area ratio is equal to 1; The calculation method of the light and shadow ratio is: It is defined that in the three-dimensional model, the cube always keeps only one side surface facing the sun, and then the light and shadow ratio Q in the projection of the same type of vegetation in the observation direction needs to consider the included angle between the observation azimuth angle and the solar azimuth angle; when the included angle is less than 90°, the side surface light component needs to be considered, and when the included angle is greater than or equal to 90°, only the proportion of the top surface light component is considered; it is defined that the case where the included angle is less than 90° corresponds to the first quadrant, and the calculation formula is: ， The calculation formula of other quadrants is adjusted based on the calculation formula of the first quadrant; S3, the average projection area is brought into the Boolean model to obtain the porosity; the calculation method of the porosity is: It is defined that the void ratio P of each type of ground object is: ， Because there is mutual shielding between the Tamarix, the tree crown and the tree trunk, the calculation formula is defined as: ， where P BG1 is the background porosity of the crown and the trunk, P BG2 is the background porosity of the crown of the tamarisk, P BG3 is the background porosity of the crown and the trunk of the tamarisk; S4, the porosity, the light and shadow ratio and the overlapping area are used to calculate the component ratio of the light vegetation, the shadow vegetation, the light soil, the shadow soil, the light tree trunk and the shadow tree trunk, and the calculation method is: The light vegetation and the shadow vegetation include the light Tamarix, the shadow Tamarix, the light Populus euphratica and the shadow Populus euphratica, and the calculation formula is defined as: ， where F gCL , F sCL , F gHY , F sHY , F gSoil , F sSoil , F gSG , F sSG are illuminated Dracaena, shadow Dracaena, illuminated Populus euphratica, shadow Populus euphratica, illuminated soil, shadow soil, illuminated tree trunk and shadow tree trunk, respectively, represents the overlapping area of the observation direction and the light direction on the ground projection.

Images