CN115909025A - Terrain vision autonomous detection and identification method for small celestial body surface sampling point - Google Patents

Abstract

The invention discloses a terrain vision autonomous detection and identification method for sampling points on the surface of a small celestial body, which belongs to the technical field of small celestial body surface detection, wherein a mechanical arm system is provided with a binocular stereo vision camera to realize vision imaging of a sampling area and identification and measurement of the terrain of the sampling points, and noise in an image is filtered by a preprocessing algorithm; then, distortion correction is carried out on the image, and imaging errors caused by an optical system of the camera are reduced; then, after binocular epipolar line correction is carried out on the images, matching is carried out on the left eye image and the right eye image, and a dense disparity map is obtained; calculating the three-dimensional coordinates of the two-dimensional disparity map in a coordinate system of a left eye camera point by point so as to obtain a three-dimensional point cloud map; sensing the overall terrain by using the point cloud data; and finally, carrying out meshing division on the plane area of the ground, calculating the area information in the sliding window, and judging the terrain of the area information, so that the operable position of the sampling device on the mechanical arm can be screened out.

Description

A method for autonomous visual detection and recognition of terrain at sampling points on the surface of small celestial bodies

Technical Field

The present invention belongs to the technical field of small celestial body surface detection, and in particular relates to a method for autonomous visual detection and recognition of terrain at sampling points on the surface of a small celestial body.

Background Art

By collecting rock samples from the surface of small celestial bodies and analyzing their composition, we can explore clues to the formation of the solar system and even the origin of life. It is an important means for small celestial body probes to carry robotic arms to conduct autonomous surface sampling and return. Due to limitations in energy, propulsion, etc., the probe's attachment mission time on small celestial bodies is relatively short. However, small celestial bodies are very far away from the earth, and the communication time is extended, making it impossible to carry out real-time measurement and control. Therefore, during the robotic arm sampling process, it is necessary to rely on on-orbit fully autonomous visual image processing to realize the detection and identification of the sampling point terrain in order to complete the robotic arm sampling operation.

The detection and recognition of non-cooperative targets is one of the difficult problems in the field of space vision. The surface rocks of small celestial bodies as sampling objects have the characteristics of weak texture and high similarity, which makes their image processing more difficult. The image of the sample rock has very small difference in pixel value, and it is very difficult to extract features. Therefore, it is very challenging to identify and measure such unstructured high-similarity rock targets. The literature (Wang Yalin et al., Analysis and Simulation Verification of the Surface Topography of asteroids with rubble pile structure, Journal of Deep Space Exploration, 2019, 6(5)) studied the surface terrain characteristics of asteroids with rubble pile structure characteristics, proposed a method for generating a simulation model of asteroid surface terrain, and experimentally simulated the power law distribution of stones that affect the terrain structure, but did not provide a method for identifying and detecting the sampling points of the robotic arm. Invention patent CN111721302A discloses a method for identifying and perceiving complex terrain features on irregular asteroid surfaces. It uses optical images taken by deep space probes to detect and distinguish craters and rock features based on the geometric characteristics of the asteroid surface terrain features. This method is mainly used for on-orbit navigation and obstacle avoidance of deep space probes, and is not suitable for detailed terrain identification of sampling points based on close-range imaging after landing.

The present invention conducts relevant research on the selection of sampling points of the robotic arm in the surface exploration mission of small celestial bodies, and proposes an autonomous detection and recognition method of the sampling point terrain based on binocular stereo vision, which has broad application prospects in the surface exploration mission of small celestial bodies.

Summary of the invention

The technical problem solved by the present invention is: in view of the weak texture high similarity characteristics of the rocks on the surface of small celestial bodies as sampling objects, a method for visual autonomous detection and identification of the terrain of sampling points on the surface of small celestial bodies based on binocular stereo vision is provided, which solves the problem of autonomous detection and identification of the terrain of surface sampling points by the sampling robot arm in a type of small celestial body surface detection mission.

The purpose of the present invention is achieved through the following technical solutions:

The robotic arm system is equipped with binocular stereo vision cameras. The common field of view of the two cameras covers the sample collection area to achieve visual imaging of the sampling area and identification and measurement of the terrain of the sampling point. The robotic arm system has been widely used. This is a prior art and will not be elaborated on here. The binocular camera is equipped with an active lighting source to ensure a good imaging effect. The two cameras synchronously image the sampling area. First, the noise in the image is filtered out by a preprocessing algorithm; then the image is distorted to reduce the imaging error caused by the camera optical system. Then, after the image is binocular epipolar line correction, the left eye image and the right eye image are matched to obtain a dense disparity map. The three-dimensional coordinates of the two-dimensional disparity map in the left camera coordinate system are calculated point by point to obtain a three-dimensional point cloud map. The point cloud data is used to perceive the overall terrain situation. Finally, the plane area where the ground is located is gridded, the area information in the sliding window is calculated, and the terrain is judged, so that the operable position of the sampling device on the robotic arm can be screened out.

A method for visual autonomous detection and recognition of sampling points on the surface of a small celestial body mainly comprises the following steps:

(1) Use a binocular stereo vision camera to image the sampling area; ensure that the public field of view of the left camera and the right camera can cover the sampling area;

(2) Preprocess the original images collected by the left and right cameras and reduce the influence of noise through appropriate filtering algorithms;

(3) performing distortion correction on the denoised image after filtering obtained in step (2) to reduce the error caused by the distortion of the camera optical system;

(4) performing binocular epipolar line correction on the distortion-corrected image obtained in step (3) to generate a left-eye epipolar line corrected image and a right-eye epipolar line corrected image;

(5) Searching for matching points of the left-eye image in the right-eye image to calculate the disparity, thereby obtaining a dense disparity map;

(6) calculating the three-dimensional coordinates of the two-dimensional disparity map obtained in step (5) in the left camera coordinate system point by point, thereby obtaining a three-dimensional point cloud map;

(7) Use point cloud data to perform ground detection on the visible area to determine the overall terrain situation within the sampling area;

(8) Grid division is performed in the plane area where the ground is located, and the regional information within the sliding window is calculated to determine its terrain.

Furthermore, the preprocessing in step (2) is performed by performing median filtering on the left-eye image and the right-eye image, the filter window size is m×m, and the calculation method includes the following steps:

A1) The width and height of the image are W and H respectively. The image boundary is expanded, and the image width and height become W+2×[m/2] and H+2×[m/2]. The expanded image pixels are set to 0;

A2) For a point (u _ori ,v _ori ) on the original image, its gray value is I(u _ori ,v _ori ), and the median filter calculation formula is as follows:

Where G(u _f ,v _f ) is the grayscale value of the pixel after filtering, W is the m×m filtering template, and i and j represent the coordinates of the pixel points on the template W.

Furthermore, the distortion correction in step (3) includes the following calculation steps:

The two-dimensional pixel homogeneous coordinates of the k-th pixel in the image before distortion correction:

The two-dimensional physical homogeneous coordinates of the k-th pixel in the image before distortion correction:

The matrix A is the intrinsic parameter matrix of the camera, and A ^-1 represents the inverse of the matrix A.

Lens distortion horizontal component value:

The vertical component of lens distortion:

k₁、k₂、k₃为一、二、三阶径向畸变，p₁、p₂为一、二阶切向畸变，in

k ₁ , k ₂ , k ₃ are the first, second, and third order radial distortions, p ₁ , p ₂ are the first and second order tangential distortions,

Then the two-dimensional physical homogeneous coordinates of the k-th pixel after distortion correction are:

Convert to two-dimensional pixel homogeneous coordinates:

Furthermore, in the step (4), the correction formula for the epipolar line correction is as follows:

Among them, [u _c v _c 1] ^T is the homogeneous pixel coordinate of the spatial point in the image before epipolar correction, [uv 1] ^T is the homogeneous pixel coordinate of the spatial point in the image after epipolar correction, M is the intrinsic parameter matrix of the camera, R is the rotation matrix of the camera coordinate system, M′ is the intrinsic parameter matrix of the camera after correction, R _rec is the rotation matrix of the camera coordinate system after correction, and λ≠0 is a constant.

Furthermore, in step (5), the block matching method is used to search for matching points of the left-eye image in the right-eye image to calculate the disparity, which includes the following steps:

B1) Calculate the difference between two pixels, that is, measure the grayscale similarity under different disparities:

e(u,v,d)＝|G _L (u,v)-G _R (ud,v)| ⑦

Where G(u,v) is the grayscale value of the pixel with coordinates (u,v) in the pixel coordinate system, and d is the disparity.

B2) Select a window around the matching point as the similarity measurement area, where the corresponding pixel is the center point of the window. In the selected window, the matching cost of the corresponding pixel is superimposed, and the result is used as the matching similarity measurement value of the point:

Where S is the similarity measurement area, which is generally an n×n rectangular area.

B3) Select the point corresponding to the smallest matching cost superposition value within the search range as the final matching point.

B4) Calculate the disparity of the entire image pixel by pixel, and finally obtain a dense disparity map.

Furthermore, in step (6), for any point (u, v) in the two-dimensional disparity map obtained by stereo matching, the three-dimensional coordinates (X, Y, Z) in the left camera coordinate system are calculated as follows:

Among them, f _x and f _y are the equivalent focal lengths of the camera, u ₀ and v ₀ are the pixel coordinates of the principal point, and B is the baseline length.

Furthermore, in step (7), the visible area ground detection includes the following steps:

C1) Randomly select K points to fit the plane, and let the plane equation be Z = AX + BY + C, then we have

Where (X _l ,Y _l ,Z _l ) are the three-dimensional space coordinates of the lth point. C2) Use the remaining points to verify the effect of plane fitting and calculate the distance from each point to the plane:

Set the flatness threshold T, count the number N _c of points whose distance to the plane is less than the threshold T under a set of plane parameters, perform multiple cycles to obtain the maximum value of N _c , and take the corresponding plane as the visible area ground.

Furthermore, in step (8), the process of determining the terrain by region includes the following steps:

S1) Determine the grid area with reference to the operation area of the sampling device, and divide the point cloud set into p×q grid areas.

S2) In each grid area, retrieve the distance value D _i from the point in the area to the ground, and calculate the roughness of the current window. Use a histogram descriptor to represent it. The histogram H is divided into several equal parts from 0 to the maximum distance D _max , and the distance is counted.

S3) setting thresholds and judgment rules, and judging whether the terrain in the area is suitable for executing the task based on the roughness histogram.

S4) Setting a sliding window and a sliding step size, performing the above roughness calculation on the sliding window, and determining whether the sliding window can execute the task.

S5) Mark the grid area that meets the task requirements, record the terrain parameters inside the area, and form an information list of the operable area.

Compared with the prior art, the present invention has the following beneficial effects:

1. Aiming at the problem of terrain detection and identification of sampling points on the surface of small celestial bodies, the present invention proposes an autonomous detection and identification method based on binocular stereo vision, which effectively solves the problem of autonomous detection and identification of rocks on the surface of small celestial bodies with weak texture and high similarity;

Second, the method for detecting and identifying terrain at sampling points on the surface of a small celestial body provided by the present invention can realize fully autonomous data processing on orbit, effectively solving the problem of extended communication time and inability of the ground system to perform real-time processing in the small celestial body detection mission;

3. The visual detection and recognition method adopted by the present invention is simple, efficient and reliable, and is suitable for space environments where computing resources are relatively scarce.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the configuration structure of a binocular stereo camera according to the present invention;

FIG2 is a flow chart of a method for autonomously detecting and identifying terrain at sampling points based on binocular stereo vision according to the present invention;

FIG. 3 is a schematic diagram of three-dimensional plane grid area division according to the present invention.

DETAILED DESCRIPTION

In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

Example

As shown in Figure 1, the binocular stereo camera is configured on the ground observation surface of the small celestial body detector to image the sampling area to detect and identify the terrain in the sampling area. The binocular stereo camera ensures the best observation range by designing the position and posture of the camera in the whole device. The binocular stereo camera is equipped with an active light source so that the scenery in its public field of view can obtain a better imaging effect. The binocular stereo camera ensures that the left camera and the right camera are imaged at the same time through synchronous acquisition to realize stereo vision calculation.

As shown in Figure 2, the sampling point terrain autonomous detection and recognition method based on binocular stereo vision mainly includes the following steps:

(1) Preprocess the binocular camera image to reduce the impact of noise. According to the imaging characteristics of the sample rock, the image preprocessing performs median filtering on the left and right images, which can smooth the data and retain tiny sharp details. The filter window size is 3×3, and the specific calculation process is as follows:

A1) The width and height of the image are W and H respectively. The image boundary is expanded, and the image width and height become W+2 and H+2, and the expanded image pixels are set to 0.

A2) For a point (u _ori ,v _ori ) on the original image, its gray value is I(u _ori ,v _ori ), and the 3×3 median filter calculation formula is as follows:

Where G(u _f ,v _f ) is the grayscale value of the pixel after filtering, W is a 3×3 filtering template, so i and j are both integers in the interval [-1,1].

(2) Perform distortion correction on the left-eye filtered image and the right-eye filtered image. The calculation process is as follows (taking the image of one camera as an example):

The two-dimensional pixel homogeneous coordinates of the k-th pixel in the image before distortion correction:

The two-dimensional physical homogeneous coordinates of the k-th pixel in the image before distortion correction:

Wherein matrix A is the intrinsic parameter matrix of the camera, and A ^-1 represents the inverse of matrix A.

Lens distortion horizontal component value:

The vertical component of lens distortion:

k₁、k₂、k₃为一、二、三阶径向畸变，p₁、p₂为一、二阶切向畸变。in

k ₁ , k ₂ , k ₃ are first, second, and third order radial distortions, and p ₁ , p ₂ are first and second order tangential distortions.

Then the two-dimensional physical homogeneous coordinates of the k-th pixel after distortion correction are:

Convert to two-dimensional pixel homogeneous coordinates:

(3) Perform epipolar correction on the distortion-corrected image. The calculation process is as follows (taking the image of one of the cameras as an example):

Among them, [u _c v _c 1] ^T is the homogeneous pixel coordinate of the spatial point in the image before epipolar correction, [uv 1] ^T is the homogeneous pixel coordinate of the spatial point in the image after epipolar correction, M is the intrinsic parameter matrix of the camera, R is the rotation matrix of the camera coordinate system, M′ is the intrinsic parameter matrix of the camera after correction, R _rec is the rotation matrix of the camera coordinate system after correction, and λ≠0 is a constant.

(4) Search and match the points with the same name in the left and right images. The calculation process is as follows:

B1) First, calculate the difference between two pixels based on the pixel points:

e(u,v,d)＝|G _L (u,v)-G _R (ud,v)| ⑦

Among them, G(u,v) is the grayscale value of the pixel with coordinates (u,v) in the pixel coordinate system.

B2) Select a window around the matching point as the similarity measurement area, where the corresponding pixel is the center point of the window. In the selected window, the matching cost of the corresponding pixel is superimposed, and the result is used as the matching similarity measurement value of the point:

Where S is the similarity measurement area, which is generally a rectangular area of n×n. In this example, n=21 is determined according to the resolution of the test image.

B3) Selecting the point corresponding to the smallest value of the matching cost superposition value within the search range as the final matching point;

B4) Calculate the disparity of the entire image pixel by pixel, and finally obtain a dense disparity map.

(5) For any point (u, v) in the two-dimensional disparity map obtained by stereo matching, calculate its three-dimensional coordinates (X, Y, Z) in the left camera coordinate system. The calculation formula is:

Among them, f _x and f _y are the equivalent focal lengths of the camera, u ₀ and v ₀ are the pixel coordinates of the principal point, and B is the baseline length.

(6) The visible area is detected using point cloud data. The calculation process is as follows:

C1) Randomly select K points to fit the plane, and let the plane equation be Z = AX + BY + C, then we have

Among them, (X _l ,Y _l ,Z _l ) are the three-dimensional space coordinates of the lth point.

C2) Use the remaining points to verify the effect of plane fitting and calculate the distance from each point to the plane:

Set the flatness threshold T, count the number N _c of points whose distance to the plane is less than the threshold T under a set of plane parameters, perform multiple cycles to obtain the maximum value of N _c , and take the corresponding plane as the visible area ground.

(7) Grid division is performed in the plane area where the ground is located, and its terrain is judged. The calculation process is as follows:

S1) Perform area division and sliding window preset in the detected plane, as shown in Figure 3. The grid size can be determined according to the single-point operation area of the sampling device. For example, if the single-point operation area of the sampling device is 100mm×100mm, and the common field of view of the binocular stereo camera is 1.2m×1.1m, 12×11 grid areas can be set, and the size of each grid is set to 100mm×100mm.

S2) Retrieve the distance value D _i from the point in each grid area to the ground, calculate the roughness of the current window, and represent it with a histogram descriptor. The histogram H is divided into 15 equal parts from 0 to the maximum distance D _max , and the distance is counted.

S3) setting a histogram interval threshold and corresponding judgment rules, and judging whether the area can perform a certain type of task according to the roughness histogram.

S4) Setting the sliding window and sliding step size, the reference grid size can be set to 100 mm × 100 mm, the sliding step size 50 mm. The above roughness calculation is performed on the sliding window, and it is determined whether the sliding window can execute the task.

S5) Mark the grid area that meets the task requirements, record the terrain parameters inside the area, and form an information list of the operable area.

In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "setting" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention are understood according to the specific circumstances. In the description of this specification, the description of the reference terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in combination with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

The above are only preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to conventional means in the art unless otherwise specified and limited.

Claims (8)

1. A terrain vision autonomous detection and identification method for small celestial body surface sampling points comprises the following steps:

(1) Imaging the sampling area by using a binocular stereo vision camera; ensuring that the common view field of the left eye camera and the right eye camera can cover the sampling area;

(2) Preprocessing original images acquired by a left eye camera and a right eye camera, and reducing the influence of noise through a filtering algorithm;

(3) Distortion correction is carried out on the denoised image obtained in the step (2) after filtering processing, and errors caused by distortion of an optical system of a camera are reduced;

(4) Performing binocular epipolar line correction on the image subjected to distortion correction obtained in the step (3) to generate a left eye epipolar line corrected image and a right eye epipolar line corrected image;

(5) Searching matching points of the left eye image and the right eye image to calculate parallax, and obtaining a dense parallax image;

(6) Calculating the three-dimensional coordinates of the two-dimensional disparity map obtained in the step (5) in a coordinate system of a left eye camera point by point, and thus obtaining a three-dimensional point cloud map;

(7) Carrying out ground detection on the visible area by using the point cloud data, and determining the overall terrain situation in the sampling area;

(8) And carrying out grid division in a plane area where the ground is positioned, and calculating area information in a sliding window so as to judge the terrain.

2. The method for terrain visual autonomous detection and identification of small celestial surface sampling points according to claim 1, characterized in that: the preprocessing in the step (2) is to perform median filtering processing on the left eye image and the right eye image, the size of a filtering window is m multiplied by m, and the calculation method comprises the following steps:

a1 The width and height of the image are W, H respectively, the image boundary is expanded, the image width and height are changed into W +2 x [ m/2] and H +2 x [ m/2], and the expanded image pixel is set as 0;

a2 For a point (u) on the original image _ori ,v _ori ) The gray value thereof is I (u) _ori ,v _ori ) The median filter calculation is as follows:

wherein G (u) _f ,v _f ) Is the gray value of the pixel after filtering, W is the filtering template of m multiplied by m, i, j represent the coordinate of the pixel point on the template W.

3. The method for terrain visual autonomous detection and identification of small celestial surface sampling points according to claim 2, characterized in that: the distortion correction in the step (3) includes the following calculation steps:

the homogeneous coordinate of the two-dimensional pixel of the kth pixel point of the image before distortion correction is as follows:

before distortion correctionThe k-th pixel point of the image has two-dimensional physical homogeneous coordinates:

wherein the matrix A is an internal parameter matrix of the camera, A ^-1 The inverse of the matrix a is represented by,

lens distortion level component value:

lens distortion vertical component value:

wherein

k ₁ 、k ₂ 、k ₃ Is a radial distortion of one, two or three orders, p ₁ 、p ₂ Is a first order and a second order tangential distortion,

then the distortion corrected two-dimensional physical homogeneous coordinate of the kth pixel point:

conversion to two-dimensional pixel homogeneous coordinates:

。

4. the method for automatically detecting and identifying the terrain of the sampling point on the surface of the small celestial body according to claim 3, which is characterized in that: in the step (4), the correction formula of the epipolar line correction is as follows:

wherein [ u ] _c v _c 1] ^T For homogeneous pixel coordinates of spatial points in the pre-epi-polar corrected image, [ u v 1] ^T Is the homogeneous pixel coordinate of the space point in the epipolar line corrected image, M is the internal parameter matrix of the camera, R is the rotation matrix of the camera coordinate system, M' is the corrected internal parameter matrix of the camera, R _rec λ ≠ 0 is a constant for the camera coordinate system rotation matrix after calibration.

5. The method for terrain visual autonomous detection and identification of small celestial surface sampling points according to claim 4, characterized in that: in the step (5), searching for a matching point of the left eye image in the right eye image by using a block matching method to calculate the parallax, and the method comprises the following steps:

b1 Compute the difference between two pixels, i.e. for gray level similarity measurements at different disparities:

e(u,v,d)＝|G _L (u,v)-G _R (u-d,v)| ⑦

wherein G (u, v) is the gray value of the pixel point with the coordinate (u, v) under the pixel coordinate system, and d is the parallax.

B2 A window surrounding the matching point is selected as a similarity measurement area, wherein the corresponding pixel is the central point of the window, in the selected window, the matching cost of the corresponding pixel is subjected to superposition operation, and the obtained result is used as the matching similarity measurement value of the point:

where S is a similarity measurement region, which is generally an n × n rectangular region.

B3 Point corresponding to the value with the minimum matching cost superposition value is selected in the search range to serve as a final matching point;

b4 The parallax is calculated pixel by pixel for the whole image, and finally a dense parallax image is obtained.

6. The method for terrain visual autonomous detection and identification of small celestial surface sampling points according to claim 5, characterized in that: in the step (6), for any point (u, v) in the two-dimensional disparity map obtained by stereo matching calculation, a three-dimensional coordinate (X, Y, Z) in the left eye camera coordinate system is calculated according to the following formula:

wherein f is _x 、f _y Is the equivalent focal length of the camera, u ₀ 、v ₀ Is the principal point pixel coordinate and B is the baseline length.

7. The method for terrain visual autonomous detection and identification of small celestial surface sampling points according to claim 6, characterized in that: in the step (7), the ground detection of the visible area comprises the following steps:

c1 ) randomly selecting K point fitting planes, and if the plane equation is Z = AX + BY + C, then there are

Wherein (X) _l ,Y _l ,Z _l ) Is the three-dimensional space coordinate of the ith point.

C2 The remaining points are used to verify the effect of the plane fit, and the distance of each point to the plane is calculated:

setting a flatness threshold T, and counting the number N of points with the distance to the plane less than the threshold T under a group of plane parameters _c Performing multiple cycles to obtain N _c Taking the corresponding plane as the visible areaAnd (4) the ground.

8. The method for automatically detecting and identifying the terrain of the sampling point on the surface of the small celestial body according to claim 1, which is characterized in that: in the step (8), the process of judging the terrain by regions comprises the following steps:

s1) determining a grid division area by referring to the operation area of a sampling device and the like, and dividing a point cloud set into p multiplied by q grid areas;

s2) searching the distance value D between the point in each grid area and the ground in the area _i Calculating the roughness of the current window, and using histogram descriptor to represent the current window, wherein the histogram H is from 0 to the maximum distance D _max Equally dividing the distance into a plurality of equal parts, and counting the distance;

s3) setting a threshold and a judgment rule, and judging whether the regional terrain can execute a task according to the roughness histogram;

s4) setting a sliding window and a sliding step length, performing the roughness calculation on the sliding window, and judging whether the sliding window can execute a task or not;

and S5) marking a grid area meeting the task requirement, recording the internal terrain parameters of the area, and forming an operable area information list.

Images