A dead-end road location and importance classification method based on remote sensing data

Through the vector road network and multi-spectral data processing of remote sensing data, combined with surface attribute analysis, the efficiency and accuracy of de-head road identification and importance grading are solved, and the automated positioning and importance evaluation of de-head roads are realized, which is suitable for urban construction. and transportation planning.

CN113643242BInactive Publication Date: 2024-11-19SOUTHEAST UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110824000.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-07-21
Publication Date
2024-11-19
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The identification and application of existing technology in broken-end roads is less researched, and it is highly targeted, low efficiency and poor applicability. It is not possible to fully utilize the important target clue guiding role of broken-end roads, making it difficult to achieve fast and accurate broken-end road positioning. and importance rating.

Method used

Using a method based on remote sensing data, the vector road network data and multi-spectral remote sensing data are registered and cropped, the eight-neighborhood method is used to detect the position of the broken-head road, calculate the surface attribute parameters, determine the broken-head road related area, and based on the terrain information We will conduct importance ratings and scenario estimation with the proportion of land objects.

Benefits of technology

It realizes the automated positioning and importance grading of broken-end roads, improves detection efficiency and accuracy, and meets the application needs of urban construction and transportation planning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113643242B_ABST
    Figure CN113643242B_ABST
Patent Text Reader

Abstract

The present invention discloses a dead-end road positioning and importance grading method based on remote sensing data, which fully considers the important target clue guiding role of the dead-end road. After accurately locating the dead-end road position using a vector road network, the dead-end road position is used as a potential area of ​​important target, and surface parameters of its associated area are extracted and analyzed to achieve importance grading of dead-end roads and estimation of associated area scenes, which meets the application needs of urban construction, traffic planning and other departments, and has important practical significance.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of remote sensing, and particularly relates to a dead-end road positioning and importance grading method. Background Art

[0002] Dead-end roads are a special type of road with special road location characteristics and important target clues. The fundamental purpose of remote sensing is to obtain information. With the advent of the information age, how to quickly and accurately identify artificial objects is of great significance to the construction of spatial geographic information databases. Typical artificial objects include bridges, dams, airports, ports, tunnels and other important targets, and these typical artificial objects are inseparable from the special road form of dead-end roads. So far, many scholars at home and abroad have proposed many road detection algorithms, but there are few studies on the identification and application of dead-end roads.

[0003] At present, the research on dead-end roads is limited to treating airport shelters as dead-end roads, extracting the complete vector road network of the airport area, and using the eight-neighborhood method to determine the location of the shelter. However, the important target clue guidance role of dead-end roads has not been taken seriously. From the existing research, key technologies such as image segmentation, region of interest detection, and feature extraction are mainly used to extract the main features of the target to be identified from large images. Through the comprehensive analysis of multiple features, the recognition and detection of important targets are realized. However, this type of target recognition algorithm is highly targeted, low in efficiency, and poor in applicability.

[0004] Therefore, it is of great practical significance to fully consider the important target clue guiding role of dead-end roads and realize the positioning and importance classification of dead-end roads, which meets the application needs of urban construction, transportation planning and other departments. Summary of the invention

[0005] In order to solve the technical problems mentioned in the above background technology, the present invention proposes a method for locating and grading the importance of dead-end roads based on remote sensing data, which fully considers the important target clue guiding role of dead-end roads, uses vector road network to accurately locate the position of dead-end roads, takes the position of dead-end roads as potential areas of important targets, extracts and analyzes surface parameters of related areas, and realizes the importance grading of dead-end roads and the estimation of related area scenes.

[0006] In order to achieve the above technical objectives, the technical solution of the present invention is:

[0007] A dead-end road location and importance classification method based on remote sensing data comprises the following steps:

[0008] (1) Collecting vector road network data, multispectral remote sensing data, and digital elevation model data; first, using the vector road network data as a benchmark, aligning the multispectral remote sensing data; then, using the coverage of the multispectral remote sensing data as a working area, clipping the vector road network data to obtain the vector road network in the working area;

[0009] (2) Rasterizing the vector road network in the work area to obtain a rasterized road network; using the eight-neighborhood method to perform global detection on the rasterized road network to determine the locations of all dead-end roads;

[0010] (3) Calculating surface attribute parameters, including remote sensing index information of various types of objects and terrain information, wherein the terrain information includes slope and elevation;

[0011] (4) determining the road direction of the dead-end road obtained in step (2), determining the associated area of ​​the dead-end road according to the location of the dead-end road and the road direction of the dead-end road, and counting the surface attribute parameters in the associated area of ​​each dead-end road;

[0012] (5) Determine the landform and slope type of the area based on the terrain information of the dead-end road associated area; determine the land use type of the area based on the proportion of different land features in the dead-end road associated area; then, classify the importance of the dead-end roads and estimate the scene type of the dead-end road associated area.

[0013] Furthermore, the specific process of step (2) is as follows:

[0014] (2.1) Obtain the number of rows and columns of the multispectral remote sensing image data, and create a single-band blank image with a pixel value of 0 and the same number of rows and columns as the multispectral remote sensing image data;

[0015] (2.2) Overlay the vector road network data with the created single-band blank image; if there is a road network in the overlapping area, set the pixel value in the corresponding blank image to 255, otherwise it remains 0; after traversing the entire single-band blank image, obtain the rasterized road network image P;

[0016] (2.3) Construct a 3×3 operator, traverse the grid road network image P, and calculate the sum of the pixel values ​​of the 8 adjacent pixels with the point to be detected as the center; if the sum of the pixel values ​​of the 8 adjacent pixels is 255 and the point is not on the boundary of the image P, then the center point is a dead end road, and the dead end road coordinate point set R is constructed.

[0017] Furthermore, in step (3), the water index, vegetation index, building land index, bare soil index and impervious surface index are calculated using multispectral remote sensing data to obtain a corresponding remote sensing index result map; the slope is calculated using digital elevation model data to obtain a corresponding slope map.

[0018] Furthermore, the method for obtaining the remote sensing index result map is as follows:

[0019] (3.1) Read multispectral remote sensing image data and store the six bands as blue band image b1, green band image b2, red band image b3, near infrared band image b4, mid-infrared band 1 image b5, and mid-infrared band 2 image b6;

[0020] (3.2) Use the normalized water index (b2-b4) / (b2+b4) to calculate the multispectral remote sensing image data and obtain the water index result map;

[0021] The normalized vegetation index (b4-b3) / (b3+b4) is used to calculate the multispectral remote sensing image data to obtain the vegetation index result map;

[0022] The ratio building land index kt1 / kt2 is used to calculate the multispectral remote sensing image data to obtain the building land index result map, where:

[0023] kt1=0.326*b1+0.509*b2+0.56*b3+0.567*b4,

[0024] kt2=-0.311*b1-0.356*b2-0.325*b3+0.819*b4;

[0025] The impervious surface index ((b4+b5+b6)-(b1+b2+b3)) / ((b4+b5+b6)+(b1+b2+b3)) is used to calculate the multispectral remote sensing image data and obtain the impervious surface index result map;

[0026] The normalized bare soil index (b4-b5) / (b4+b5) is used to calculate the multispectral remote sensing image data to obtain the bare soil index result map.

[0027] Furthermore, the method to obtain the slope map is as follows:

[0028] (3a) obtaining projection information of multispectral remote sensing image data, performing projection conversion on digital elevation model data, and obtaining converted digital elevation model data DEM1;

[0029] (3b) Establish a 3×3 operator, traverse all pixels of DEM1, and calculate the east-west change rate S of the central pixel according to formula (1) we , calculate the north-south change rate S of the central pixel according to formula (2) sn , calculate the slope value S of the central pixel according to formula (3) and obtain the slope map:

[0030]

[0031]

[0032]

[0033] Among them, e1 and e3 represent the pixel values ​​of the left and right adjacent pixels of the central pixel in DEM1, and e2 and e4 represent the pixel values ​​of the upper and lower adjacent pixels of the central pixel in DEM1;

[0034] (3c) The obtained slope map and digital elevation model data are resampled to the same size as the multispectral remote sensing image data.

[0035] Furthermore, in step (4), the method for determining the road direction of the dead-end road is as follows:

[0036] Using the rasterized road network obtained in step (2), with the dead-end road obtained in step (2) as the center point, detect the pixel point A whose pixel value is not 0 on the square boundary 5 pixels away from the center point; according to the pixel point A and the dead-end road position coordinates, calculate the azimuth β of the line connecting the two points, with (45°, 135°] as the bottom, (135°, 255°] as the right, (225°, 315°] as the top and (0°, 45°] as the left, and divide the road direction of the dead-end road into four directions in the image: up, down, left, and right.

[0037] Furthermore, in step (4), the method for determining the associated area of ​​the dead-end road is as follows:

[0038] If the dead-end road goes upward, the dead-end point is used as the center point of the lower side of the square, and a square window with a side length of n is set; if the dead-end road goes downward, the dead-end point is used as the center point of the upper side of the square, and a square window with a side length of n is set; if the dead-end road goes to the left, the dead-end point is used as the center point of the right side of the square, and a square window with a side length of n is set; if the dead-end road goes to the right, the dead-end point is used as the center point of the left side of the square, and a square window with a side length of n is set.

[0039] Furthermore, in step (4), the process of calculating the surface attribute parameters in the associated area of ​​the dead-end road is as follows:

[0040] (4.1) Set the threshold T of the water index result map W , vegetation index result map threshold T V , set the upper and lower thresholds T for the impervious surface index result map NS1 、T NS2 , building land index result map threshold T R , bare soil index result map setting threshold T S ;

[0041] (4.2) Set the number of pixels in the square window in step (4) to n 2; The pixel value in the water index result map within the statistical window is greater than T W The number of pixels N W , the pixel value in the vegetation index result map is greater than T V The number of pixels N V , the pixel value in the impervious surface index result map is greater than T NS1 and less than T NS2 The number of pixels N ns , the pixel value in the building land index result map is less than T R The number of pixels N r , the pixel value in the bare soil index result map is less than T S The number of pixels N s ;

[0042] (4.3) Calculate the ratio of each feature in the window i :

[0043]

[0044] Among them, N i represents the number of pixels belonging to feature i, i = W,V,ns,r,s;

[0045] (4.4) Count the maximum, minimum and mean values ​​of elevation and slope in the area associated with the dead-end road.

[0046] Furthermore, the specific process of step (5) is as follows:

[0047] (5.1) Based on the maximum, minimum and mean values ​​of elevation in the dead-end road associated area, the dead-end road scene is classified into landform types, and the landform types of the scene are determined to be plain, hilly and mountainous areas; based on the maximum, minimum and mean values ​​of slope in the dead-end road associated area, the dead-end road scene is classified into slope types, and the slope types of the scene are determined to be flat slope, gentle slope and inclined slope;

[0048] (5.2) According to the proportion of each type of land feature in the dead-end road associated area, the scene land use type is determined as agricultural land or ecological land mainly covered by vegetation or water, construction land covered in whole or in part by construction area, or unused land covered by large areas of bare soil;

[0049] (5.3) The importance of dead-end roads is graded according to the landform type, slope type and land use type of the areas associated with the dead-end roads, and the scenario type of the areas associated with the dead-end roads is estimated.

[0050] The beneficial effects brought by adopting the above technical solution are:

[0051] (1) The present invention uses vector road network data to detect and locate the position of dead-end roads; uses multispectral remote sensing images to calculate corresponding index result maps according to different ground object remote sensing index calculation formulas, and uses DEM data to calculate elevation and slope information; uses the important target clue guidance role of dead-end roads, conducts a comprehensive analysis of the surface information of the areas associated with the dead-end roads, classifies the importance of the dead-end roads, and estimates the scenes of the areas associated with the dead-end roads, which basically realizes automation, has good stability and important practical significance.

[0052] (2) The present invention uses vector road network data to detect and locate dead-end roads. The positioning result is only affected by the vector road network data and the result is accurate. The present invention uses the rasterized road network image to determine the associated area of ​​the dead-end road and conducts a comprehensive analysis of the surface parameters of the associated area of ​​the dead-end road, thereby realizing the importance classification of the dead-end road and the scene estimation of the associated area. There is no need to perform global feature recognition analysis, which improves the detection efficiency and shortens the time. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flow chart of the method of the present invention;

[0054] Figure 2 is a remote sensing image of the area to be detected in the embodiment;

[0055] Figure 3 is a grid road network diagram in the embodiment;

[0056] Figure 4 : is a diagram of the dead-end road position detection result in the embodiment;

[0057] Figure 5 It is a water body index, vegetation index, normalized bare soil index, impervious surface index, ratio building land index, and slope calculation result diagram in the embodiment;

[0058] Figure 6 This is a diagram of the dead-end road associated area scene estimation result in the embodiment. DETAILED DESCRIPTION

[0059] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0060] The present invention designs a dead-end road location and importance classification method based on remote sensing data, such as Figure 1 As shown, the steps are as follows:

[0061] Step 1. Data preparation: Collect vector road network data, multispectral remote sensing data (including red, green, blue bands, near infrared bands and two mid-infrared bands), and digital elevation model (DEM) data.

[0062] Step 2: Dead-end road detection and positioning based on vector road network. The specific implementation steps are as follows:

[0063] Step 2.1: rasterize the vector road network data. Figure 2 ) covers the working area, clips the vector road network data to obtain the vector road network in the working area, and then rasterizes the vector road network in the working area to obtain the raster road network result map P, as shown in Figure 3 As shown;

[0064] Step 2.2, dead-end road detection and location. Use the eight-neighborhood method to perform pixel-by-pixel calculations on the grid road network P, detect and locate the location of the dead-end road, and obtain the dead-end road location detection result, such as Figure 4 As shown;

[0065] Step 3: Calculate the surface attribute parameters. The specific implementation steps are as follows:

[0066] Surface attribute parameters include remote sensing index information of various types of land objects, slope, elevation and other terrain information.

[0067] Step 3.1, calculate the remote sensing index result map based on multispectral remote sensing image data. Calculate the relevant ground object remote sensing index according to the normalized water index, normalized vegetation index, ratio building land index, normalized bare soil index, and impervious surface index to obtain the corresponding ground object remote sensing index result map, such as Figure 5 As shown in (a)-(e);

[0068] Step 3.2, calculate the slope map based on DEM data. Use DEM data to calculate the elevation change rate of each pixel in the east-west and north-south directions, so as to obtain the slope value of each pixel. After generating the slope map, resample it to obtain the slope map, such as Figure 5 As shown in (f) in .

[0069] Step 4: Determine the associated area of ​​the dead-end road and count the surface parameters in the associated area. The specific implementation steps are as follows:

[0070] Step 4.1, determine the direction of the dead-end road: take the dead-end road as the center point, detect the pixel point A whose pixel value is not 0 on the square boundary 5 pixels away from the center point in the grid road network P, and calculate the azimuth β between the center point and point A; according to the calculated azimuth β, (45°, 135°] is down, (135°, 225°] is right, (225°, 315°] is up, (315°, 360°] and [0°, 45°] are left, and the direction of the dead-end road is divided into four directions in the image: up, down, left, and right;

[0071] Step 4.2, determine the associated area of ​​the dead-end road: take the location of the dead-end point as the starting point, set a square window with a side length of n in the extension direction of the dead-end road according to the direction of the road, which is the associated area of ​​the dead-end road;

[0072] Step 4.3, calculate the proportion of different land features in the dead-end road associated area: count the number of pixels N with pixel values ​​greater than 0 in the water index result map in the square window W , the number of pixels N with pixel values ​​greater than 0.3 in the vegetation index result map V , the number of pixels N with pixel values ​​greater than 0 and less than 0.6 in the impervious surface index result map ns , the number of pixels N with pixel values ​​less than -5 in the building land index result map r , the number of pixels N with pixel values ​​less than 0 in the bare soil index result map s , calculate the ratio of each feature in the square window based on the number of pixels of different features obtained by statistics i ;

[0073] Step 4.4, calculate the terrain information of the dead-end road associated area: according to the resampled DEM data, the maximum, minimum and mean elevation in the dead-end road associated area are counted; according to the resampled slope map, the maximum, minimum and mean slope in the dead-end road associated area are counted.

[0074] Step 5: Dead-end road importance classification and associated area scenario estimation. The specific implementation steps are as follows:

[0075] Step 5.1, set the elevation threshold T DEM1 =200m, T DEM2 =500m to classify the landform types of the scene in the area associated with the dead-end road. The average elevation in the associated area is less than T DEM1 The topography is plain, and the average elevation is greater than T DEM1 and less than T DEM2 The topography is hilly, and the average elevation is greater than T DEM2 The terrain is mountainous; set the slope threshold T Slope1 =5°, T Slope2 =15° to divide the slope of the dead-end road related area scene, and the average slope in the related area is less than T Slope1 The slope is flat and the average slope is greater than T Slope1 and less than T Slope2 The slope is gentle and the average slope is greater than T Slope2 is a slope;

[0076] Step 5.2: According to the ratio of each type of land feature in the dead-end road associated area i, determine the scene land use type. The dead-end road-related areas with a construction land ratio of more than 20% or an impervious surface ratio of more than 50% are determined as construction land; the dead-end road-related areas with a vegetation ratio of more than 80% are determined as agricultural land or ecological land in a large vegetation coverage area; the dead-end road-related areas with a water body ratio of more than 80% are determined as agricultural land or ecological land in a large water coverage area; the dead-end road-related areas with a bare soil ratio of more than 80% are determined as unused land;

[0077] Step 5.3, classify the importance of the dead-end road according to the landform type, slope type and land use type of the dead-end road associated area, and estimate the scene type of the dead-end road associated area. For example, the importance of building land with a bare soil and vegetation ratio greater than 15% and located in a flat slope area is low, and it is estimated to be a residential area; the importance of building land with a water body ratio greater than 30% is high, and it is estimated to be a port; the importance of a large area of ​​vegetation coverage with an impervious surface ratio of about 10% and located in a flat slope area is relatively high, and it is estimated to be a tunnel entrance, and so on. The dead-end road importance classification and associated area scene estimation results are obtained, such as Figure 6 shown.

[0078] The embodiments are only for illustrating the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A dead-end road location and importance classification method based on remote sensing data, characterized in that: The following steps are involved: (1) Collect vector road network data, multispectral remote sensing data, and digital elevation model data; first, use the vector road network data as a benchmark to align the multispectral remote sensing data; then, use the coverage of the multispectral remote sensing data as the working area to crop the vector road network data to obtain the vector road network within the working area; (2) Rasterize the vector road network in the work area to obtain a rasterized road network; use the eight-neighborhood method to perform global detection on the rasterized road network to determine the locations of all dead-end roads; (3) Calculating surface attribute parameters, including remote sensing index information of various land objects and terrain information, including slope and elevation; (4) determining the road direction of the dead-end road obtained in step (2), determining the associated area of ​​the dead-end road according to the location and road direction of the dead-end road, and calculating the surface attribute parameters within the associated area of ​​each dead-end road; (5) Determine the landform and slope type of the area based on the topographic information of the area associated with the dead-end road; Based on the proportion of different land features in the dead-end road associated area, the land use type of the area is determined; then, the importance of the dead-end roads is graded and the scene type of the dead-end road associated area is estimated; (2.1) Obtain the number of rows and columns of the multispectral remote sensing image data, and create a single-band blank image with a pixel value of 0 and the same number of rows and columns as the multispectral remote sensing image data; (2.2) Overlay the vector road network data with the created single-band blank image; if a road network exists in the overlapping area, the pixel value in the corresponding blank image is set to 255, otherwise it remains 0; after traversing the entire single-band blank image, the rasterized road network image P is obtained; (2.3) Construct a 3×3 operator, traverse the grid road network image P, take the point to be detected as the center point, and calculate the sum of the pixel values ​​of the adjacent 8 pixels; If the sum of the pixel values ​​of the adjacent 8 pixels is 255 and the point is not on the boundary of the image P, then the center point is a dead end, and the dead end coordinate point set R is constructed; In step (4), the method for determining the direction of the dead-end road is as follows: Using the rasterized road network obtained in step (2), with the dead-end road obtained in step (2) as the center point, detect the pixel point A on the square boundary 5 pixels away from the center point, whose pixel value is not 0; based on the pixel point A and the dead-end road position coordinates, calculate the azimuth angle β of the line connecting the two points, with (45°, 135°] as the bottom, (135°, 255°] as the right, (225°, 315°] as the top and (0°, 45°] as the left, and divide the road direction of the dead-end road into four directions in the image: up, down, left and right; In step (4), the method for determining the associated area of ​​the dead-end road is as follows: If the dead-end road is heading upward, the dead-end point is used as the center point of the lower edge of the square, and a square window with a side length of n is set; if the dead-end road is heading downward, the dead-end point is used as the center point of the upper edge of the square, and a square window with a side length of n is set; if the dead-end road is heading left, the dead-end point is used as the center point of the right edge of the square, and a square window with a side length of n is set; if the dead-end road is heading right, the dead-end point is used as the center point of the left edge of the square, and a square window with a side length of n is set; In step (4), the process of calculating the surface attribute parameters in the associated area of ​​the dead-end road is as follows: (4.1) Set the threshold T of the water index result map W , vegetation index result map threshold T V , set the upper and lower thresholds T in the impervious surface index result map NS1 、T NS2 , building land index result map threshold T R , the bare soil index result map sets the threshold T S ; (4.2) Set the number of pixels in the square window in step (4) to n 2 ; The pixel value in the water index result map within the statistical window is greater than T W The number of pixels N W , the pixel value in the vegetation index result map is greater than T V The number of pixels N V , the pixel value in the impervious surface index result map is greater than T NS1 and less than T NS2 The number of pixels N ns , the pixel value in the building land index result map is less than T R The number of pixels N r , the pixel value in the bare soil index result map is less than T S The number of pixels N s ; (4.3) Calculate the ratio of each feature in the window i : Among them, N i represents the number of pixels belonging to feature i, i = W, V, ns, r, s; (4.4) Calculate the maximum, minimum and mean values ​​of elevation and slope in the area associated with the dead-end road.

2. The dead-end road location and importance classification method based on remote sensing data according to claim 1 is characterized in that: In step (3), the water index, vegetation index, building land index, bare soil index and impervious surface index are calculated using multispectral remote sensing data to obtain the corresponding remote sensing index result map; the slope is calculated using digital elevation model data to obtain the corresponding slope map.

3. The dead-end road location and importance classification method based on remote sensing data according to claim 2 is characterized in that: The method to obtain the remote sensing index result map is as follows: (3.1) Read multispectral remote sensing image data and store the six bands as blue band image b1, green band image b2, red band image b3, near infrared band image b4, mid-infrared band 1 image b5, and mid-infrared band 2 image b6; (3.2) Calculate the multispectral remote sensing image data using the normalized water index (b2-b4) / (b2+b4) to obtain the water index result map; The normalized vegetation index (b4-b3) / (b3+b4) is used to calculate the multispectral remote sensing image data to obtain the vegetation index result map; The ratio of building land index kt1 / kt2 is used to calculate the multispectral remote sensing image data to obtain the building land index result map, where: kt1=0.326*b1+0.509*b2+0.56*b3+0.567*b4, kt2=-0.311*b1-0.356*b2-0.325*b3+0.819*b4; The impervious surface index ((b4+b5+b6)-(b1+b2+b3)) / ((b4+b5+b6)+(b1+b2+b3)) is used to calculate the multispectral remote sensing image data to obtain the impervious surface index result map; The normalized bare soil index (b4-b5) / (b4+b5) was used to calculate the multispectral remote sensing image data to obtain the bare soil index result map.

4. The dead-end road location and importance classification method based on remote sensing data according to claim 2 is characterized in that: The method to obtain the slope map is as follows: (3a) obtaining projection information of multispectral remote sensing image data, performing projection conversion on digital elevation model data, and obtaining converted digital elevation model data DEM1; (3b) Establish a 3×3 operator, traverse all pixels of DEM1, and calculate the east-west change rate S of the central pixel according to formula (1) we , calculate the north-south change rate S of the central pixel according to formula (2) sn , calculate the slope value S of the central pixel according to formula (3) and obtain the slope map: Among them, e1 and e3 represent the pixel values ​​of the left and right adjacent pixels of the central pixel in DEM1, and e2 and e4 represent the pixel values ​​of the upper and lower adjacent pixels of the central pixel in DEM1; (3c) The obtained slope map and digital elevation model data are resampled to the same size as the multispectral remote sensing image data.

5. The dead-end road location and importance classification method based on remote sensing data according to claim 1 is characterized in that: The specific process of step (5) is as follows: (5.1) Based on the maximum, minimum, and mean elevation values ​​within the dead-end road's associated area, the dead-end road scene was classified into landform types, including plain, hilly, and mountainous areas. Based on the maximum, minimum, and mean slope values ​​within the dead-end road's associated area, the dead-end road scene was classified into slope types, including flat slope, gentle slope, and inclined slope. (5.2) Based on the proportion of each type of land feature in the dead-end road associated area, determine the scene land use type as agricultural land or ecological land covered mainly by vegetation or water, construction land covered entirely or partially by built-up areas, or unused land covered by large areas of bare soil; (5.3) The importance of dead-end roads is graded according to the landform type, slope type and land use type of the areas associated with dead-end roads, and the scenario type of the areas associated with dead-end roads is estimated.