A method for confirming similar stands based on multi-source data

Abstract

The application discloses a kind of based on the confirmation method of similar stand of multi-source data, comprising the following steps: S1: establish spatial grid: according to the area size of the similar stand of demand in target search area Suitable spatial grid is established, and the unique ID number of each spatial grid is coded;S2: information spatial grid: the stand factor in spatial grid is valued to spatial grid;S3: search target spatial grid: determine stand factor target range, using database management tool to query the spatial grid that meets target;S4: optimization query result: various stand factors are concentrated in one or several spatial grid, according to spatial grid selection traffic convenient distance shortest route, planning investigation scheme.The application can quickly, accurately confirm similar stand position based on multi-source data, field investigation intensity is small, fast, accuracy is high.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of forest resource management, and particularly relates to a method for confirming similar forest stands based on multi-source data. BACKGROUND

[0002] A forest stand refers to a piece of forest with generally consistent internal characteristics and obvious differences from adjacent sections. That is, the forest section has generally similar tree species composition, forest origin, forest layer or forest aspect, forest type, forest age, site class, timber yield, and other factors, and has obvious differences from adjacent sections. The forests in a forest area can be divided into different forest stands according to the composition of tree species, forest origin, forest aspect, forest age, density, forest type, and other factors. Similar forest stand confirmation refers to finding forest stands with the same tree species, diameter at breast height (ground diameter), and height as the target area under the same site conditions.

[0003] Similar forest stands are widely used in forest resource management: due to ecological destruction behaviors such as illegal logging, deforestation and reclamation, illegal mining and sand and soil extraction, illegal construction, illegal transplantation, harmful biological damage, human-caused fires, illegal tourism development, and environmental pollution behaviors such as pollutant discharge and dumping, resulting in forest loss and the need to use similar forest stands to assess on-site losses of forest resources; similar forest stands are used to identify the volume of timber after logging in forest area acceptance.

[0004] The document "Establishing Similar Forest Stand Sample Tree Factor Regression Model to Calculate Harvested Wood Volume" (Wu Riheng, Hunan Forestry Science and Technology, Vol. 41, No. 1, January 2014) discloses setting up a sample plot in the similar forest stand of Cunninghamia lanceolata plantations in Anhua County, measuring the ground diameter, breast height, and tree height of 26 trees on the diagonal of the sample plot, establishing regression models of ground diameter and breast height, and breast height and tree height using Excel software, measuring the ground diameter of harvested wood, predicting the breast height and tree height of harvested wood using the regression model, and calculating the volume of harvested wood. At the same time, the volume of timber is calculated using the ground diameter monomial stand volume table method, and compared with the volume of timber calculated using the binary stand volume table method of the harvesting operation design. The results show that the method of calculating the volume of harvested wood using the similar forest stand sample tree factor regression model is close to the volume of harvested wood calculated using the harvesting operation design. The document mainly introduces the application of similar forest stands, and does not disclose how to find similar forest stands.

[0005] The paper "Influence of multi-source data on forest dynamic prediction and uncertainty analysis" (Tian Xianglin et al., Forest Science, Vol. 57, No. 3, March 2021) takes the example of whole forest model updating to demonstrate the data-model logical framework of the Bayesian method of presenting information in the form of probability distribution, which is constantly circulating, updating and fusing. The purpose is to compare the influence of multi-source data on forest dynamic prediction, analyze the variation law of model parameters and prediction uncertainty, evaluate the model from the accuracy and reliability, and obtain the data requirements of the improved model to provide suggestions for data collection strategies in forest survey. The method is as follows: collect the 3-period survey (1990, 2005 and 2012) and 4 types of information (temporary sample plots, fixed sample plots, sample trees and multi-source data) modeling data of Pinus tabulaeformis forest in Qinling Mountains, design a set of variable density whole forest models with low data information requirements, and analyze the relationship between traditional forest survey data and growth yield model based on the dynamic fusion framework of Bayesian information. The joint posterior distribution of parameters obtained by MCMC sampling technology is used to quantify the uncertainty of forest dynamic simulation: on the one hand, compare the probability distribution change process of model parameters and prediction after the same type of multi-period forest survey data constantly trains the model; on the other hand, compare the influence of model prediction by using 4 types of data respectively. The data and model updating cycle is realized by the method of constantly transforming prior information and posterior information, that is, the joint posterior distribution of parameters obtained by the previous fitting is used as the prior of the next data. The integration of different data types is realized according to the independent likelihood structure designed according to the sampling and observation error of the data itself. In order to avoid the influence of rough data or outliers on the model, the likelihood function describing the error distribution adopts the heavy-tailed normal distribution. The heteroscedasticity of observation error is controlled by automatically adjusting the variance of likelihood function in iteration. The paper mainly compares, analyzes and discusses the influence of multi-source data on forest dynamic prediction and the variation law of model parameters and prediction uncertainty, while in actual forest survey, not only the accuracy and reliability of the survey need to be met, but also the executability of the survey work from the aspects of personnel safety and work efficiency needs to be considered. The method of finding similar forest is not disclosed.

[0006] How to find similar forest, especially from various data to obtain effective information still faces great difficulties: 1. Data is diverse, and standardization and consistency work is heavy. Especially the standardization and consistency work of electronic data format is heavy, especially the work of obtaining effective information from raster data format; 2. Spatial information is insufficient. Due to the lack of spatial information in the early data, the positioning is not accurate, and the work efficiency of finding similar forest is low; 3. The workload of survey is large. The indicators are relatively independent and cannot be optimized and combined. Therefore, the information of finding similar forest is inaccurate, the efficiency is low, the workload of survey is large, and the advantages of information technology cannot be fully played. Therefore, a simple, fast, accurate and low-cost method for finding similar forest is explored. SUMMARY

[0007] The application aims to provide a similar forest stand confirmation method based on multi-source data, which can solve the technical problems of time-consuming, laborious and difficult operation in the prior art in searching for similar forest stands, and can quickly and accurately confirm similar forest stands, reduce labor intensity, and provide help for forest resource assessment.

[0008] To achieve the above-mentioned purpose, the following scheme is proposed:

[0009] A similar forest stand confirmation method based on multi-source data, comprising the following steps:

[0010] S1: Establishing a spatial grid: according to the size of the required similar forest stand area, a suitable spatial grid is established in the target search area, and each spatial grid is coded with a unique ID number;

[0011] S2: Informationized spatial grid: assigning forest stand factors in the spatial grid to the spatial grid;

[0012] S3: Searching and querying the target spatial grid: determining the target range of forest stand factors, and using a database management tool to query the spatial grid that meets the target;

[0013] S4: Optimizing the query result: concentrating various forest stand factors in one or several spatial grids, and planning the investigation scheme according to the shortest route of the spatial grid selected.

[0014] Preferably, in step S1, when high accuracy is required, a small spatial grid is established, otherwise, a large spatial grid is established.

[0015] Preferably, in step S1, when the similarity area of the forest stand factors such as site factors and vegetation factors in the target search area is large, a large spatial grid is established, otherwise, the area of the similar forest stand is divided into small blocks, and a smaller spatial grid is established.

[0016] Preferably, in step S1, the spatial grid is established by using the fishing net tool in the toolbox of ArcGIS.

[0017] Preferably, the forest stand factors include site factors and vegetation factors.

[0018] Preferably, the site factors include the average elevation, slope, and slope direction of the sample plot.

[0019] Preferably, the vegetation factors include similar vegetation coverage and vegetation height.

[0020] Preferably, in step S3, the target range of forest stand factors is within ±5% of the average value and standard deviation of each spatial information index.

[0021] Preferably, in step S3, the Python language programming is completed by using ArcPy, numpy package to write program code.

[0022] Compared with the prior art, the present application has the following beneficial effects:

[0023] (1) The present application quickly and accurately identifies similar forest stands by obtaining forest site conditions, vegetation structure before damage, vegetation cover factors and soil factors necessary for confirming similar forest stands from DEM, remote sensing images, soil data and the like, thereby providing help for forest resource assessment.

[0024] (2) The present application is clear in positioning, has spatial information, easy to identify similar forest stand location, convenient for field line planning, easy to select the optimal result; if multiple results can be selected, the results can be optimized, reducing the field work load, reducing the field work load, improving the field efficiency, especially when multiple stand factor surveys are performed, the effect is more significant; high repeatability, some factors can be created once and used multiple times, such as elevation, slope, slope direction and the like in site factors. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a flowchart of the present application;

[0026] Figure 2 is a statistical table in the form of a spatial grid slope of Example 1 of the present application;

[0027] Figure 3 is a statistical table in the form of a spatial grid of each factor of Example 1 of the present application;

[0028] Figure 4 is a field survey map of Example 1 of the present application;

[0029] Figure 5 is a range map of Guangxi Zhuang Autonomous Region Gaofeng State-owned Forest Farm of Example 2 of the present application;

[0030] Figure 6 is a query result schematic diagram of step S4 of Example 2 of the present application. DETAILED DESCRIPTION

[0031] Before further describing the specific embodiments of the present application, it should be understood that the scope of protection of the present application is not limited to the following specific embodiments; it should also be understood that the terms used in the embodiments of the present application are for describing the specific embodiments, but not for limiting the scope of protection of the present application.

[0032] Example 1

[0033] A similar stand identification method based on multi-source data, as shown in Figure 1 the following steps are included:

[0034] S1: Establish a spatial grid: Take Wuming District of Nanning City as the target area for searching similar stands. This time, the fishing net tool in the toolbox of ArcGIS is used to establish a spatial grid of 25.82 x 25.82 m, and the area of each grid is 666.7 m 2 , which is 1 mu, a total of 9537624 spatial grids, and each spatial grid is coded with ID number from 1 to 9537624. The principle is to find the spatial grid of similar stands with an area of 1 mu in Wuming District of Nanning City based on the set sample plot, so as to obtain the spatial information and site factors, vegetation factors, etc. of similar stands.

[0035] S2: Informationized spatial grid: Informationized spatial grid is to further concentrate the stand factor information such as site factors and vegetation factors in the spatial grid for management and query. This time, the site factors and vegetation factors are verified: the site factors mainly include the average elevation, slope and aspect of the sample plot. This time, the 12.5 m DEM informationized spatial grid is used, that is, the average elevation, slope and aspect of each spatial grid range are assigned to the spatial grid extraction. The DEM data comes from the official website of NASA (https: / / search.asf.alaska.edu / # / ); the vegetation factors mainly obtain similar vegetation coverage and vegetation height through DEM, DSM and high-resolution satellite data. The DEM and DSM data come from the laser radar data of Guangxi 2019 second-class survey, and the high-resolution satellite data is the image data used for change detection of Guangxi Zhuang Autonomous Region. This time, the regional attribute statistics table of each factor for each spatial grid region is generated by using the regional analysis function of ArcGIS toolbox, including the spatial information, average value, standard deviation, etc. of each grid for each factor, such as the spatial grid slope statistics table (see Figure 2 ). Because the remote sensing image used this time is a single scene data that has not been spliced, in order to facilitate the verification of the results, the image block statistics work is completed by using the ArcPy and GDAL packages to write program code through Python language, and the vector data (shp format) of the image block that meets the data screening conditions is exported, which is convenient for the query result optimization of the 4th step.

[0036] S3: Search for target spatial grid: When querying, first determine the specific screening main indicators according to the purpose, this time mainly refer to the average value and standard deviation of each spatial information index in the region; secondly, determine the control range of the query, this time use the ±5% range of the average value and standard deviation of each spatial information index, as shown in Table 1; then, use the ArcPy package to program the query of the spatial grid that meets the conditions of each factor through Python language and export the queried spatial grid.

[0037] Table 1 query target space grid characteristic value query condition

[0038]

[0039] Table 2 space information grid query result

[0040]

[0041] S4: optimization query result: according to the space grid queried by each factor, and generate each factor space grid statistics table( Figure 3 ), calculate the distance from the field survey starting point to all grid and sort the distance; at the same time, superimpose the space grid of each factor, as far as possible to have more than two forest factors in one space grid, and refer to the distance of the field survey starting point and the traffic condition, confirm the specific space grid position that needs to be investigated( Figure 4 ), facilitate field investigation, thereby reducing the field workload and improving work efficiency.

[0042] Example 2

[0043] In the "Guangxi Zhuang Autonomous Region State-owned Gaofeng Forest Farm Forest Carbon Sink Pilot Construction Project", multiple sample plots need to be laid out, and the feasibility of the present application is further verified by using these laid-out sample plots.

[0044] S1: establish space grid: the size of the sample plot of the Guangxi Zhuang Autonomous Region State-owned Gaofeng Forest Farm Forest Carbon Sink is 25.82x25.82m, therefore, the same size of space grid layer is established using the fishnet tool in the toolbox of ArcGIS, and 611914 grids are established in the Guangxi Zhuang Autonomous Region State-owned Gaofeng Forest Farm, such as Figure 5 .

[0045] S2: information space grid: the data used this time still comes from the official website of NASA (https: / / search.asf.alaska.edu / # / ); the vegetation data comes from the Guangxi 2019 second-class survey laser radar data. The Guangxi Zhuang Autonomous Region State-owned Gaofeng Forest Farm Forest Carbon Sink Pilot mainly does sample plot investigation in red camphor forest, therefore, the vector layer of annual update management of the Guangxi Zhuang Autonomous Region State-owned Gaofeng Forest Farm is also used as a control layer in this query, so as to make the queried space grid fall in the red camphor forest as far as possible, and further reduce the query range. The image used is the image data used for change plot detection in Guangxi Zhuang Autonomous Region.

[0046] S3: search query target space grid: query conditions should be based on the target set query conditions. This search target, the query main factor is vegetation cover and vegetation height, and the slope, aspect and other site factors are used as reference factors, and the vegetation height and image space grid query conditions are as shown in Table 3.

[0047] Table 3: Vegetation height and image space grid characteristic value query condition

[0048]

[0049] S4: Optimize query results: superimpose the red camphor forest spatial layer of Guangxi Zhuang Autonomous Region State-owned Gao Feng Forest Farm with the searched vegetation height spatial grid, image spatial grid and slope, aspect and elevation spatial grid, and select 78 spatial grids in total, and the query result schematic diagram is as shown in Figure 6 .

[0050] Using the optimized query spatial grid and combining with the traffic conditions, further filter the field investigation sample site location, and plan the field investigation work. In the red camphor forest investigation sample site of Guangxi Zhuang Autonomous Region State-owned Gao Feng Forest Farm, the application can avoid the work of screening the sample site location in advance, and can also directly navigate to the field according to the spatial grid location to quickly determine the sample site. In the field investigation test, the time for determining the sample point is at least more than 1 hour compared with the method of using only paper and electronic data to find similar stands.

[0051] The above-described embodiments are only preferred modes of the application and do not limit the scope of the application. Without departing from the design spirit of the application, various modifications and improvements to the technical solutions of the application made by those skilled in the art shall fall within the protection scope of the application as defined by the claims.

Claims

1. A method for confirming a similar stand based on multi-source data, characterized in that, The method comprises the following steps: S1: Establishing a spatial grid: according to the size of the similar stand area in the target search area, a suitable spatial grid is established, and each spatial grid is coded with a unique ID number; S2: Informationized spatial grid: the stand factors in the spatial grid are assigned to the spatial grid; S3: Searching and querying the target spatial grid: determine the target range of stand factors, and use the database management tool to query the spatial grid that meets the target; In step S1, when the target search area has a large similar area of site factors, vegetation factors and stand factors, a large spatial grid is established; Otherwise, the similar stand area is divided into small blocks, and a small spatial grid is established; The stand factors include site factors and vegetation factors; The site factors include the average elevation, slope and slope direction of the sample plot; The vegetation factors include similar vegetation coverage and vegetation height; In step S3, the target range of stand factors is within the range of ±5% of the average value and standard deviation of each spatial information index; S4: Optimizing the query result: according to the spatial grid queried by each factor, a statistical table of each factor spatial grid is generated, the distance from the field investigation starting point to all the grids of the query result is calculated, and the grids are sorted according to the distance; Various stand factors are concentrated in one or several spatial grids, the shortest route is selected according to the spatial grid and convenient transportation, and the investigation plan is programmed.

2. The method of claim 1, wherein: In step S1, when the precision is high, a small spatial grid is established, otherwise, a large spatial grid is established.

3. The method of claim 1, wherein: In step S1, the fishnet tool in the toolbox of ArcGIS is used to establish the spatial grid.

4. The method of claim 1, wherein: In step S3, the program code is written by using ArcPy and numpy package to complete the Python language programming.

Images