Degradation environment map construction method and device based on adaptive pose estimation

Abstract

The invention relates to a computer vision technology, and discloses a degraded environment map construction method based on adaptive pose estimation, which comprises the following steps: acquiring an IMU (Inertial Measurement Unit) measurement value, and calculating an IMU pre-integration factor; acquiring laser radar data, and performing point cloud distortion removal processing to obtain standard laser radar data; extracting feature points in the standard laser radar data, and performing degradation environment identification to obtain a degradation environment identification result; calculating a degradation factor, and adjusting a feature weight corresponding to the feature point; carrying out adaptive pose updating according to a degradation environment identification result, and carrying out point cloud matching to obtain a laser radar factor; and integrating the IMU pre-integration factor and the laser radar factor into a factor graph to carry out factor graph optimization, outputting pose information, and obtaining a degraded environment map according to the pose information and the adaptive pose. The invention further provides a degradation environment map construction device and equipment based on self-adaptive pose estimation and a medium. The method can improve the success rate of degraded environment mapping.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a method and device for constructing a degraded environment map based on adaptive pose estimation. Background Art

[0002] Laser SLAM (Simultaneous Localization and Mapping) systems primarily consist of two components: a laser radar (LiDAR) and an algorithm. The LiDAR emits a laser beam and receives reflected light, acquiring distance information about the surrounding environment and generating point cloud data. The algorithm uses this point cloud data to perform robot positioning and map construction.

[0003] However, reality presents many geometrically degraded environments, including flat ground, stairs, subway tunnels, off-road terrain, inland waterways, urban tunnels, and bridges. Traditional laser SLAM algorithms struggle to construct accurate point cloud maps in these degraded environments. Most current methods lack robust and effective degradation detection mechanisms. This lack of a degradation detection mechanism can lead to poor system performance in degraded environments, making it impossible to adjust algorithm parameters or take appropriate measures to address degradation, thus impacting the performance and stability of the SLAM system. This can lead to serious errors in the mapping process and ultimately to map failure. Summary of the Invention

[0004] The present invention provides a method and device for constructing a degraded environment map based on adaptive pose estimation, which can improve the success rate of mapping in a degraded environment.

[0005] To achieve the above objectives, the present invention provides a method for constructing a degraded environment map based on adaptive pose estimation, comprising:

[0006] Obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values;

[0007] Obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data;

[0008] Extract feature points from standard lidar data, perform degraded environment recognition based on the feature points, and obtain degraded environment recognition results;

[0009] Calculate the degradation factor based on the degradation environment identification result, and adjust the feature weight corresponding to the feature point according to the degradation factor;

[0010] Adaptive pose update is performed based on the degraded environment recognition results, and point cloud matching is performed with standard radar data in the adaptive pose to obtain the lidar factor;

[0011] The IMU pre-integration factor and lidar factor are integrated into the factor graph for factor graph optimization, and the pose information is output according to the optimized factor graph. The degraded environment map is obtained based on the pose information and the adaptive pose.

[0012] Optionally, calculating an IMU pre-integration factor according to an IMU measurement value includes:

[0013]

[0014] Where Δp ij is the translation of the IMU state from time i to time j, p i is the translation of the IMU state at time i, p j is the translation of the IMU state at time j, Δv ij is the linear velocity change of the IMU state from time i to time j, v i is the linear velocity of the IMU state at time i, g is the gravity vector, ΔR ij is the rotation change of the IMU state from time i to time j, is the transpose of the rotation matrix of the IMU state at time i, R j is the rotation matrix of the IMU state at time j, Δt ij is the time difference from time i to time j, δp ij is the first noise, δv ij is the second noise, δφ ij The third noise.

[0015] Optionally, performing point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data includes:

[0016] Calculate the rotation pose of the point cloud frame in the lidar data in the most recent time frame from the current moment according to the IMU pre-integration factor;

[0017] The relative rotation amount is calculated based on the rotation pose, and the relative rotation amount is used to correct the pose of the point cloud in the lidar data at the current moment to complete the point cloud distortion processing.

[0018] Optionally, extracting feature points from standard lidar data includes:

[0019] Calculate the curvature of the point cloud set in the standard lidar data, and identify the edge points and plane points in the standard lidar data according to the curvature value;

[0020] Calculate the angle between the point cloud scan line and the point cloud scan point in the standard lidar data, and identify the ground points in the standard lidar data according to the preset angle threshold.

[0021] Optionally, after extracting the feature points from the standard lidar data, the method further includes scoring the importance of the feature points and evaluating the effect of feature point extraction based on the importance scores.

[0022] Optionally, adjusting the feature weight corresponding to the feature point according to the degradation factor includes:

[0023] Weighting function w of ground points g , weighted function w of edge points e They are defined as:

[0024]

[0025] in, is the initial weight of the ground point, is the initial weight of the edge point, α is the first weight magnification coefficient, β is the second weight magnification coefficient, s is the ground score, and D is the final degradation factor.

[0026] Optionally, the performing of degraded environment identification according to the feature points to obtain a degraded environment identification result includes:

[0027] Extract the non-ground point cloud set from the feature points, and perform eigenvalue decomposition on the non-ground point cloud set to obtain the eigenvalue size;

[0028] Identify whether the current scene is a single feature degraded environment based on the eigenvalue size and the degree of discreteness of the point cloud in the direction of the corresponding eigenvector;

[0029] Identify whether the current scene is a feature-deficient degraded environment based on the number of feature points and plane points.

[0030] In order to solve the above problems, the present invention also provides a device for constructing a degraded environment map based on adaptive pose estimation, the device comprising:

[0031] A factor calculation module is used to obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values; obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data based on the IMU pre-integration factors to obtain standard lidar data; extract feature points in the standard lidar data and perform degraded environment identification based on the feature points to obtain degraded environment identification results; calculate degradation factors based on the degraded environment identification results and adjust the feature weights corresponding to the feature points based on the degradation factors; perform adaptive pose update based on the degraded environment identification results and perform point cloud matching with the standard lidar data in the adaptive pose to obtain the lidar factor;

[0032] The degraded environment map construction module is used to integrate the IMU pre-integration factor and the lidar factor into the factor graph for factor graph optimization, output the pose information according to the optimized factor graph, and obtain the degraded environment map based on the pose information and the adaptive pose.

[0033] In order to solve the above problem, the present invention further provides an electronic device, comprising:

[0034] at least one processor; and,

[0035] a memory communicatively connected to the at least one processor; wherein,

[0036] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the above-mentioned method for constructing a degraded environment map based on adaptive pose estimation.

[0037] In order to solve the above problems, the present invention also provides a computer-readable storage medium, which stores at least one computer program. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned method for constructing a degraded environment map based on adaptive pose estimation.

[0038] The present invention performs point cloud dedistortion on the point cloud data in the lidar data based on the IMU pre-integration factor, which can ensure the spatiotemporal consistency of the lidar point cloud data. In addition, the feature points in the standard lidar data are extracted and the degraded environment is identified based on the feature points, which can ensure the accuracy of the degraded environment detection. In addition, through IMU pre-integration and adaptive pose update, the impact of local errors on the generation of degraded environment maps can be reduced, thereby ensuring the success rate of mapping. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a flow chart of a method for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0040] Figure 2 A schematic diagram of an algorithm framework of a method for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0041] Figure 3 A schematic diagram of motion distortion in a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention;

[0042] Figure 4 A schematic diagram of point cloud dedistortion in a method for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0043] Figure 5 A schematic diagram of ground point determination in a method for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0044] Figure 6 A schematic diagram of a degraded environment according to a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention;

[0045] Figure 7 A schematic diagram of an adaptive pose estimation method for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0046] Figure 8 A schematic diagram comparing mapping effects of a method for constructing a map in a degraded environment based on adaptive pose estimation provided by an embodiment of the present invention;

[0047] Figure 9 A functional module diagram of a device for constructing a degraded environment map based on adaptive pose estimation provided by one embodiment of the present invention;

[0048] Figure 10 A schematic structural diagram of an electronic device for implementing a method for constructing a degraded environment map based on adaptive pose estimation provided in one embodiment of the present invention.

[0049] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0050] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] An embodiment of the present application provides a method for constructing a degraded environment map based on adaptive pose estimation. The execution subject of the method for constructing a degraded environment map based on adaptive pose estimation includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for constructing a degraded environment map based on adaptive pose estimation can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0052] Reference Figure 1 FIG2 is a flow chart of a method for constructing a degraded environment map based on adaptive pose estimation according to an embodiment of the present invention. In this embodiment, the method for constructing a degraded environment map based on adaptive pose estimation includes:

[0053] S1. Obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values.

[0054] In the embodiments of the present invention, IMU measurement values refer to a set of data obtained by measuring the inertial measurement unit of a moving target. Here, IMU (Inertial Measurement Unit) refers to an inertial measurement unit, which is a sensor device that measures the motion state of an object based on the principle of inertia. It is used in fields such as navigation, attitude control, and motion tracking. Its core function is to obtain the acceleration, angular velocity, and magnetic field information of the moving target in real time through multi-sensor fusion, and then infer parameter information such as position, velocity, and attitude.

[0055] As an embodiment of the present invention, calculating an IMU pre-integration factor according to an IMU measurement value includes:

[0056]

[0057] Where Δp ij is the translation of the IMU state from time i to time j, p i is the translation of the IMU state at time i, p j is the translation of the IMU state at time j, Δv ij is the linear velocity change of the IMU state from time i to time j, v i is the linear velocity of the IMU state at time i, g is the gravity vector, ΔR ij is the rotation change of the IMU state from time i to time j, is the transpose of the rotation matrix of the IMU state at time i, R j is the rotation matrix of the IMU state at time j, Δt ij is the time difference from time i to time j, δp ij is the first noise, δv ij is the second noise, δφ ij The third noise.

[0058] In the embodiment of the present invention, the first noise, the second noise and the third noise obey Gaussian distribution.

[0059] S2. Obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data.

[0060] In the embodiments of the present invention, LiDAR (Light Detection and Ranging) data refers to three-dimensional spatial information obtained by scanning the environment through a moving target's LiDAR sensor, usually presented in the form of a point cloud. LiDAR data records the geometric features of the surfaces of objects in the environment.

[0061] As an embodiment of the present invention, point cloud dedistortion processing is performed on point cloud data in lidar data according to an IMU pre-integration factor to obtain standard lidar data, including:

[0062] Calculate the rotation pose of the point cloud frame in the lidar data in the most recent time frame from the current moment according to the IMU pre-integration factor;

[0063] The relative rotation amount is calculated based on the rotation pose, and the relative rotation amount is used to correct the pose of the point cloud in the lidar data at the current moment to complete the point cloud distortion processing.

[0064] For example, the point cloud data in the lidar data is dedistorted according to the IMU pre-integration factor to obtain the standard lidar data, and the following implementation steps can be adopted: wherein the pose transformation from time i to time j is T j is the pose matrix at moment j, and then the point cloud at moment i is projected to moment j according to the pose transformation of these two moments. According to this principle, the point cloud of each frame can be projected to the initial moment of each frame.

[0065] S3. Extract feature points from the standard lidar data, and perform degraded environment recognition based on the feature points to obtain a degraded environment recognition result.

[0066] In an embodiment of the present invention, feature points in standard lidar data refer to key points with significant geometric characteristics, which are used to improve data processing efficiency and environment modeling accuracy, and are not limited to edge points, plane points, and ground points.

[0067] In an embodiment of the present invention, extracting feature points from standard laser radar data includes:

[0068] Calculate the curvature of the point cloud set in the standard lidar data, and identify the edge points and plane points in the standard lidar data according to the curvature value;

[0069] Calculate the angle between the point cloud scan line and the point cloud scan point in the standard lidar data, and identify the ground points in the standard lidar data according to the preset angle threshold.

[0070] For example, the following steps may be used to extract feature points from standard lidar data:

[0071] Step 1: The laser radar used is a rotating laser radar. In order to ensure uniform constraints in a 360-degree circle, each laser line is divided into 6 areas. The point cloud set in each area is represented by N. The curvature of the point cloud in each area is calculated separately. The curvature formula is as follows:

[0072]

[0073] Among them, S is the set of points on the same laser radar scan line of point i, |S| represents the number of points included in the set S, and P i,k represents the coordinates of the i-th point in the point cloud set at the k-th frame time, ||P i,k || is the coordinate norm of point i in the LiDAR coordinate system, that is, the Euclidean distance between point i and the LiDAR center, k represents the kth frame, and point j is the neighboring point of point i.

[0074] Step 2: Arrange the points according to the calculated curvature c, and take the 20 points with the largest curvature as edge points. Then, among the remaining points, the points with curvature less than the threshold c are sorted. th The point set is downsampled as plane points.

[0075] Step 3. Next is the extraction of ground feature points. First, the scanning range of the laser radar is [-15°, -15°], and the ground points should appear on [-15°, -1°]. First, assume that the ground points only appear on the downward scanning line. If the angle between the same index points on two adjacent scanning lines in the vertical direction satisfies θ<θ th The point is considered to be a ground point. The calculation formula for the θ angle is:

[0076]

[0077] Where r represents the index of the scan line and m represents the index of the scan point.

[0078] Furthermore, after extracting feature points from the standard lidar data, the method also includes scoring the importance of the feature points and evaluating the feature point extraction effect based on the importance score.

[0079] For example, the importance of feature points is scored, and the feature point extraction effect is evaluated based on the importance score. The following implementation steps can be used:

[0080] Define the current frame ground point cloud set as Local ground point cloud set P l g , the number of point clouds is The covariance matrix of the local ground point cloud is calculated as:

[0081]

[0082] in P l g The i-th ground point in , P l g The mean of .

[0083] right Perform eigenvalue decomposition, and the eigenvector corresponding to the minimum eigenvalue is the normal vector of the local ground Determine the equation of the fitted plane.

[0084] To evaluate the effect of ground features, statistics The distance from the plane is greater than d th Number of point clouds Calculate the cosine similarity as:

[0085]

[0086] The comprehensive score of the ground is:

[0087]

[0088] The value range of s is [0,1]. The closer it is to 1, the higher the ground reliability.

[0089] As an embodiment of the present invention, degraded environment recognition is performed based on feature points to obtain a degraded environment recognition result, including:

[0090] Extract the non-ground point cloud set from the feature points, and perform eigenvalue decomposition on the non-ground point cloud set to obtain the eigenvalue size;

[0091] Identify whether the current scene is a single feature degraded environment based on the eigenvalue size and the degree of discreteness of the point cloud in the direction of the corresponding eigenvector;

[0092] Identify whether the current scene is a feature-deficient degraded environment based on the number of feature points and plane points.

[0093] S4. Calculate the degradation factor according to the degradation environment identification result, and adjust the feature weight corresponding to the feature point according to the degradation factor.

[0094] In the embodiment of the present invention, the degradation factor refers to an indicator used in the SLAM (Simultaneous Localization and Mapping) system to quantify the degree of deficiency of the geometric features of the environment. It mainly solves the positioning drift problem caused by the failure of feature matching in degraded environments (such as tunnels, corridors, and open scenes) of sensors.

[0095] As an embodiment of the present invention, calculating the degradation factor according to the degradation environment identification result includes:

[0096] When the current scene is a single feature degradation environment, the following formula is used to calculate the first degradation factor D t :

[0097]

[0098] Among them, λ k represents the obtained eigenvalue, represents the largest eigenvalue.

[0099] When the current scene is a feature-less degradation environment, the second degradation factor D is calculated using the following formula: b :

[0100]

[0101] in, is the number of edge feature points in the current frame, is the number of plane feature points in the current frame, is the number of edge feature points in the local map, is the number of plane feature points in the local map.

[0102] The first degradation factor and the second degradation factor are integrated to obtain the final degradation factor D, which is integrated using the following formula:

[0103]

[0104] Among them, D th is the degradation factor threshold, the typical value is 0.1,

[0105] Exemplarily, the feature weights corresponding to the feature points are adjusted according to the degradation factors, using the following implementation steps:

[0106] When the environment is degraded, the edge points in the feature points are lacking in features and their reliability is reduced. In order to improve the stability and accuracy of the system, the feature weights of the edge points and ground points are adjusted according to the degree of degradation.

[0107] Weighting function w of ground points g , weighted function w of edge points e They are defined as:

[0108]

[0109] in, is the initial weight of the ground point, is the initial weight of the edge point, α is the first weight magnification coefficient, β is the second weight magnification coefficient, s is the ground score, and D is the final degradation factor.

[0110] S5. Update the adaptive pose according to the degraded environment recognition result, and perform point cloud matching with the standard radar data in the adaptive pose to obtain the lidar factor.

[0111] In the embodiment of the present invention, posture refers to the general term for the position (Position) and orientation (Orientation) of a robot or vehicle in three-dimensional space. Among them, posture describes the rotation state of an object, and is usually represented by Euler angles (Roll, Pitch, Yaw) or quaternions. Euler angles uniquely determine the spatial orientation of an object through the rotation angles around three orthogonal axes (x-axis, y-axis, z-axis); Roll (roll angle): rotation around the x-axis, reflecting the left and right tilt of the object (such as the angle change when the vehicle rolls over); Pitch (pitch angle): rotation around the y-axis, reflecting the front and back pitch of the object (such as the head-up angle of the vehicle when going uphill); Yaw (yaw angle): rotation around the z-axis, reflecting the heading direction of the object (such as the forward direction of the vehicle).

[0112] As an embodiment of the present invention, an adaptive posture update is performed according to the degraded environment recognition result, including:

[0113] When the degraded environment identification result is not degraded, the IMU pre-integration factor is used to update the initial pose of the lidar device in each frame;

[0114] When the degraded environment identification result is degraded, the initial pose of the lidar device in each frame is updated by weighted fusion based on the historical poses.

[0115] For example, when the degraded environment identification result is degraded, weighted fusion is performed based on historical poses to update the initial pose of the lidar device in each frame, using the following implementation steps:

[0116] When the state change between two frames is abnormal, a weighted update is performed based on the historical state information. The position update formula is as follows:

[0117]

[0118] in represents the displacement between two frames, x′ k , y′ k 、z′ k Represents the anomaly estimate to be updated.

[0119] The update formulas for roll and pitch are as follows:

[0120]

[0121] where roll′ k is the estimated value of the roll angle anomaly to be updated, pitch′ kIndicates the pitch angle anomaly estimate to be updated.

[0122] In the embodiment of the present invention, for yaw (yaw angle), according to The eigenvalue decomposition results show that the direction of the eigenvector corresponding to the maximum eigenvalue is the vehicle's heading, which is used to determine yaw. The physical significance of this determination method lies in determining the vehicle's heading by capturing the dominant direction of the spatial distribution of point cloud data, thus avoiding misjudgments of direction due to other interference factors in complex environments.

[0123] As an embodiment of the present invention, performing point cloud matching with standard radar data in an adaptive pose includes:

[0124] Calculate the distance between the point cloud of the current frame and the edge point on the standard radar data as the first residual;

[0125] Calculate the distance from the point cloud of the current frame to the nearest straight line on the standard radar data as the second residual;

[0126] Calculate the distance from the point cloud of the current frame to the plane on the standard radar data as the third residual;

[0127] The optimal transformation is constructed based on the first residual, the second residual and the third residual to complete the point cloud matching.

[0128] S6. Integrate the IMU pre-integration factor and the lidar factor into the factor graph to perform factor graph optimization, output the pose information according to the optimized factor graph, and obtain the degraded environment map according to the pose information and the adaptive pose.

[0129] In this embodiment of the present invention, factor graph optimization refers to a nonlinear optimization method based on a graphical model. This method models multi-sensor data as factors and constructs a global optimization problem to solve the robot's pose and environmental map. In the context of multi-sensor fusion involving LiDAR and IMU sensors, the core concept is to model the observation constraints of each sensor as factor nodes, express the dependencies between variables through a graph structure, and achieve high-precision state estimation.

[0130] In the embodiment of the present invention, Global Navigation Satellite System (GNSS) data may also be introduced into the factor graph as a factor.

[0131] Furthermore, the global navigation satellite system (GNSS) data conversion factor in this patent application adopts the following implementation steps:

[0132] The positioning of the embodiment of the present invention adopts the local northeast sky ENU coordinate system, while the GNSS data is based on the LLA coordinate system, so it needs to be converted from the LLA coordinate system to the ENU coordinate system before fusion.

[0133] Assume that the GNSS data in the LLA coordinate system are longitude λ, latitude φ and altitude h. First, they need to be converted into the three-axis coordinates (x, y, z) in the LLA coordinate system. L , the calculation formula is as follows:

[0134]

[0135] Where RN is the radius of curvature and e is the eccentricity of the ellipse.

[0136] If the LLA coordinate corresponding to the starting point of the local ENU coordinate system is (x0, y0, z0) L . Get new LLA coordinates (x,y,z) L Then, through the rotation matrix Convert to ENU coordinates (x, y, z) E :

[0137]

[0138] Through the above steps, the conversion of GNSS data from LLA to local ENU is completed.

[0139] In the embodiment of the present invention, the Global Navigation Satellite System (GNSS) refers to a space radio navigation system that provides position, velocity, and time information to global or regional users through a satellite constellation.

[0140] The present invention performs point cloud dedistortion on the point cloud data in the lidar data based on the IMU pre-integration factor, which can ensure the spatiotemporal consistency of the lidar point cloud data. In addition, the feature points in the standard lidar data are extracted and the degraded environment is identified based on the feature points, which can ensure the accuracy of the degraded environment detection. In addition, through IMU pre-integration and adaptive pose update, the impact of local errors on the generation of degraded environment maps can be reduced, thereby ensuring the success rate of mapping.

[0141] like Figure 2 , which is a schematic diagram of an algorithm framework of a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0142] like Figure 3 , which is a schematic diagram of motion distortion of a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0143] like Figure 4 , which is a schematic diagram of point cloud dedistortion of a degraded environment map construction method based on adaptive pose estimation provided by an embodiment of the present invention.

[0144] like Figure 5, which is a schematic diagram of ground point judgment in a degraded environment map construction method based on adaptive pose estimation provided by an embodiment of the present invention.

[0145] like Figure 6 , which is a schematic diagram of a degraded environment according to a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0146] like Figure 7 , which is a schematic diagram of an adaptive pose estimation method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0147] like Figure 8 , which is a schematic diagram comparing mapping effects of a method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0148] like Figure 9 , which is a functional module diagram of a degraded environment map construction device based on adaptive pose estimation provided by one embodiment of the present invention.

[0149] The apparatus 100 for constructing a degraded environment map based on adaptive pose estimation of the present invention can be installed in an electronic device. According to the functions to be implemented, the apparatus 100 for constructing a degraded environment map based on adaptive pose estimation can include a factor calculation module 101 and a degraded environment map construction module 102.

[0150] The module described in the present invention may also be referred to as a unit, which refers to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, and is stored in a memory of the electronic device.

[0151] In this embodiment, the functions of each module / unit are as follows:

[0152] The factor calculation module 101 is used to obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values; obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data based on the IMU pre-integration factors to obtain standard lidar data; extract feature points in the standard lidar data and perform degraded environment identification based on the feature points to obtain degraded environment identification results; calculate degradation factors based on the degraded environment identification results and adjust feature weights corresponding to the feature points based on the degradation factors; perform adaptive pose update based on the degraded environment identification results and perform point cloud matching with the standard radar data in the adaptive pose to obtain the lidar factor.

[0153] In the embodiment of the present invention, the IMU measurement value refers to a set of data obtained by measuring using an inertial measurement unit, where IMU (Inertial Measurement Unit) refers to an inertial measurement unit, which is a sensor device that measures the motion state of an object based on the principle of inertia. It is used in fields such as navigation, attitude control, and motion tracking. Its core is to obtain the acceleration, angular velocity, and magnetic field information of an object in real time through multi-sensor fusion, and then infer parameter information such as position, velocity, and attitude.

[0154] As an embodiment of the present invention, calculating an IMU pre-integration factor according to an IMU measurement value includes:

[0155]

[0156]

[0157] Where Δp ij is the translation of the IMU state from time i to time j, p i is the translation of the IMU state at time i, p j is the translation of the IMU state at time j, Δv ij is the linear velocity change of the IMU state from time i to time j, v i is the linear velocity of the IMU state at time i, g is the gravity vector, ΔR ij is the rotation change of the IMU state from time i to time j, is the transpose of the rotation matrix of the IMU state at time i, R j is the rotation matrix of the IMU state at time j, Δt ij is the time difference from time i to time j, δp ij is the first noise, δv ij is the second noise, δφ ij The third noise.

[0158] In the embodiment of the present invention, the first noise, the second noise and the third noise obey Gaussian distribution.

[0159] In the embodiments of the present invention, LiDAR (Light Detection and Ranging) data refers to three-dimensional spatial information obtained by scanning the environment through a moving target's LiDAR sensor, usually presented in the form of a point cloud. LiDAR data records the geometric features of the surfaces of objects in the environment.

[0160] As an embodiment of the present invention, point cloud dedistortion processing is performed on point cloud data in lidar data according to an IMU pre-integration factor to obtain standard lidar data, including:

[0161] Calculate the rotation pose of the point cloud frame in the lidar data in the most recent time frame from the current moment according to the IMU pre-integration factor;

[0162] The relative rotation amount is calculated based on the rotation pose, and the relative rotation amount is used to correct the pose of the point cloud in the lidar data at the current moment to complete the point cloud distortion processing.

[0163] For example, the point cloud data in the lidar data is dedistorted according to the IMU pre-integration factor to obtain the standard lidar data, and the following implementation steps can be adopted: wherein the pose transformation from time i to time j is T j is the pose matrix at moment j, and then the point cloud at moment i is projected to moment j according to the pose transformation of these two moments. According to this principle, the point cloud of each frame can be projected to the initial moment of each frame.

[0164] In an embodiment of the present invention, feature points in standard lidar data refer to key points with significant geometric characteristics, which are used to improve data processing efficiency and environment modeling accuracy, and are not limited to edge points, plane points, and ground points.

[0165] In an embodiment of the present invention, extracting feature points from standard laser radar data includes:

[0166] Calculate the curvature of the point cloud set in the standard lidar data, and identify the edge points and plane points in the standard lidar data according to the curvature value;

[0167] Calculate the angle between the point cloud scan line and the point cloud scan point in the standard lidar data, and identify the ground points in the standard lidar data according to the preset angle threshold.

[0168] For example, the following steps may be used to extract feature points from standard lidar data:

[0169] Step 1: The laser radar used is a rotating laser radar. In order to ensure uniform constraints in a 360-degree circle, each laser line is divided into 6 areas. The point cloud set in each area is represented by N. The curvature of the point cloud in each area is calculated separately. The curvature formula is as follows:

[0170]

[0171] Among them, S is the set of points on the same laser radar scan line of point i, |S| represents the number of points included in the set S, Pi, k represents the coordinates of the i-th point in the point cloud set at the k-th frame time, ||P i,k || is the coordinate norm of point i in the LiDAR coordinate system, that is, the Euclidean distance between point i and the LiDAR center, k represents the kth frame, and point j is the neighboring point of point i.

[0172] Step 2: Arrange the points according to the calculated curvature c, and take the 20 points with the largest curvature as edge points. Then, among the remaining points, the points with curvature less than the threshold c are sorted. th The point set is downsampled as plane points.

[0173] Step 3. Next is the extraction of ground feature points. First, the scanning range of the laser radar is [-15°, -15°], and the ground points should appear on [-15°, -1°]. First, assume that the ground points only appear on the downward scanning line. If the angle between the same index points on two adjacent scanning lines in the vertical direction satisfies θ<θ th The point is considered to be a ground point. The calculation formula for the θ angle is:

[0174]

[0175] Where r represents the index of the scan line and m represents the index of the scan point.

[0176] Furthermore, after extracting feature points from the standard lidar data, the method also includes scoring the importance of the feature points and evaluating the feature point extraction effect based on the importance score.

[0177] For example, the importance of feature points is scored, and the feature point extraction effect is evaluated based on the importance score. The following implementation steps can be used:

[0178] Define the current frame ground point cloud set as Local ground point cloud set P l g , the number of point clouds is The covariance matrix of the local ground point cloud is calculated as:

[0179]

[0180] in P l g The i-th ground point in , P l g The mean of .

[0181] right Perform eigenvalue decomposition, and the eigenvector corresponding to the minimum eigenvalue is the normal vector of the local ground Determine the equation of the fitted plane.

[0182] To evaluate the effect of ground features, statistics The distance from the plane is greater than d th Number of point clouds Calculate the cosine similarity as:

[0183]

[0184] The comprehensive score of the ground is:

[0185]

[0186] The value range of s is [0,1]. The closer it is to 1, the higher the ground reliability.

[0187] As an embodiment of the present invention, degraded environment recognition is performed based on feature points to obtain a degraded environment recognition result, including:

[0188] Extract the non-ground point cloud set from the feature points, and perform eigenvalue decomposition on the non-ground point cloud set to obtain the eigenvalue size;

[0189] Identify whether the current scene is a single feature degraded environment based on the eigenvalue size and the degree of discreteness of the point cloud in the direction of the corresponding eigenvector;

[0190] Identify whether the current scene is a feature-deficient degraded environment based on the number of feature points and plane points.

[0191] In the embodiment of the present invention, the degradation factor refers to an indicator used in the SLAM (Simultaneous Localization and Mapping) system to quantify the degree of deficiency of the geometric features of the environment. It mainly solves the positioning drift problem caused by the failure of feature matching in degraded environments (such as tunnels, corridors, and open scenes) of sensors.

[0192] As an embodiment of the present invention, calculating the degradation factor according to the degradation environment identification result includes:

[0193] When the current scene is a single feature degradation environment, the following formula is used to calculate the first degradation factor D t :

[0194]

[0195] Among them, λ k represents the obtained eigenvalue, represents the largest eigenvalue.

[0196] When the current scene is a feature-less degradation environment, the second degradation factor D is calculated using the following formula: b :

[0197]

[0198] in, is the number of edge feature points in the current frame, is the number of plane feature points in the current frame, is the number of edge feature points in the local map, is the number of plane feature points in the local map.

[0199] The first degradation factor and the second degradation factor are integrated to obtain the final degradation factor D, which is integrated using the following formula:

[0200]

[0201] Among them, D th is the degradation factor threshold, the typical value is 0.1,

[0202] Exemplarily, the feature weights corresponding to the feature points are adjusted according to the degradation factors, using the following implementation steps:

[0203] When the environment is degraded, the edge points in the feature points are lacking in features and their reliability is reduced. In order to improve the stability and accuracy of the system, the feature weights of the edge points and ground points are adjusted according to the degree of degradation.

[0204] Weighting function w of ground points g , edge point w e The weighting functions are defined as:

[0205]

[0206] in, is the initial weight of the ground point, is the initial weight of the edge point, α is the first weight amplification coefficient, β is the second weight amplification coefficient, and s is the ground score.

[0207] In the embodiment of the present invention, posture refers to the general term for the position (Position) and orientation (Orientation) of a robot or vehicle in three-dimensional space. Among them, posture describes the rotation state of an object, and is usually represented by Euler angles (Roll, Pitch, Yaw) or quaternions. Euler angles uniquely determine the spatial orientation of an object through the rotation angles around three orthogonal axes (x-axis, y-axis, z-axis); Roll (roll angle): rotation around the x-axis, reflecting the left and right tilt of the object (such as the angle change when the vehicle rolls over); Pitch (pitch angle): rotation around the y-axis, reflecting the front and back pitch of the object (such as the head-up angle of the vehicle when going uphill); Yaw (yaw angle): rotation around the z-axis, reflecting the heading direction of the object (such as the forward direction of the vehicle).

[0208] As an embodiment of the present invention, an adaptive posture update is performed according to the degraded environment recognition result, including:

[0209] When the degraded environment identification result is not degraded, the IMU pre-integration factor is used to update the initial pose of the lidar device in each frame;

[0210] When the degraded environment identification result is degraded, the initial pose of the lidar device in each frame is updated by weighted fusion based on the historical poses.

[0211] For example, when the degraded environment identification result is degraded, weighted fusion is performed based on historical poses to update the initial pose of the lidar device in each frame, using the following implementation steps:

[0212] When the state change between two frames is abnormal, a weighted update is performed based on the historical state information. The position update formula is as follows:

[0213]

[0214] in represents the displacement between two frames, x′ k , y′ k 、z′ k Represents the anomaly estimate to be updated.

[0215] The update formulas for roll and pitch are as follows:

[0216]

[0217] where roll′ k is the estimated value of the roll angle anomaly to be updated, pitch′ k Indicates the pitch angle anomaly estimate to be updated.

[0218] In the embodiment of the present invention, for yaw (yaw angle), according to The eigenvalue decomposition results show that the direction of the eigenvector corresponding to the maximum eigenvalue is the vehicle's heading, which is used to determine yaw. The physical significance of this determination method lies in determining the vehicle's heading by capturing the dominant direction of the spatial distribution of point cloud data, thus avoiding misjudgments of direction due to other interference factors in complex environments.

[0219] As an embodiment of the present invention, performing point cloud matching with standard radar data in an adaptive pose includes:

[0220] Calculate the distance between the point cloud of the current frame and the edge point on the standard radar data as the first residual;

[0221] Calculate the distance from the point cloud of the current frame to the nearest straight line on the standard radar data as the second residual;

[0222] Calculate the distance from the point cloud of the current frame to the plane on the standard radar data as the third residual;

[0223] The optimal transformation is constructed based on the first residual, the second residual and the third residual to complete the point cloud matching.

[0224] The degraded environment map construction module 102 is used to integrate the IMU pre-integration factor and the lidar factor into the factor graph to perform factor graph optimization, output pose information according to the optimized factor graph, and obtain a degraded environment map based on the pose information and the adaptive pose.

[0225] In this embodiment of the present invention, factor graph optimization refers to a nonlinear optimization method based on a graphical model. This method models multi-sensor data as factors and constructs a global optimization problem to solve the robot's pose and environmental map. In the context of multi-sensor fusion involving LiDAR and IMU sensors, the core concept is to model the observation constraints of each sensor as factor nodes, express the dependencies between variables through a graph structure, and achieve high-precision state estimation.

[0226] In the embodiment of the present invention, Global Navigation Satellite System (GNSS) data may also be introduced into the factor graph as a factor.

[0227] Furthermore, the global navigation satellite system (GNSS) data conversion factor in this patent application adopts the following implementation steps:

[0228] The positioning of the embodiment of the present invention adopts the local northeast sky ENU coordinate system, while the GNSS data is based on the LLA coordinate system, so it needs to be converted from the LLA coordinate system to the ENU coordinate system before fusion.

[0229] Assume that the GNSS data in the LLA coordinate system are longitude λ, latitude φ and altitude h. First, they need to be converted into the three-axis coordinates (x, y, z) in the LLA coordinate system. L , the calculation formula is as follows:

[0230]

[0231] Where RN is the radius of curvature and e is the eccentricity of the ellipse.

[0232] If the LLA coordinate corresponding to the starting point of the local ENU coordinate system is (x0, y0, z0) L . Get new LLA coordinates (x,y,z) L Then, through the rotation matrix Convert to ENU coordinates (x, y, z) E :

[0233]

[0234] Through the above steps, the conversion of GNSS data from LLA to local ENU is completed.

[0235] In the embodiment of the present invention, the Global Navigation Satellite System (GNSS) refers to a space radio navigation system that provides position, velocity, and time information to global or regional users through a satellite constellation.

[0236] Reference Figure 10 , which is a structural diagram of an electronic device for implementing the method for constructing a degraded environment map based on adaptive pose estimation provided by an embodiment of the present invention.

[0237] The electronic device may include a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a method program for constructing a degraded environment map based on adaptive pose estimation.

[0238] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor 10 is the control core (Control Unit) of the electronic device, connecting the various components of the entire electronic device using various interfaces and lines, and executing the programs or modules stored in the memory 11 (for example, executing a method program for constructing a degraded environment map based on adaptive pose estimation, etc.), as well as calling the data stored in the memory 11, to execute various functions of the electronic device and process data.

[0239] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 may also be an external storage device of the electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card, etc. equipped on the electronic device. Furthermore, the memory 11 may also include both an internal storage unit of the electronic device and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device, such as the code of a program for a method for constructing a degraded environment map based on adaptive pose estimation, but can also be used to temporarily store data that has been output or is to be output.

[0240] The communication bus 12 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0241] The communication interface 13 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0242] Figure 10 Only the electronic device with components is shown, and it can be understood by those skilled in the art that Figure 10 The structure shown does not limit the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0243] For example, although not shown, the electronic device may further include a power source (such as a battery) for powering various components. Preferably, the power source may be logically connected to the at least one processor 10 via a power management device, thereby implementing functions such as charge management, discharge management, and power consumption management through the power management device. The power source may further include any components such as one or more DC or AC power sources, a recharging device, a power failure detection circuit, a power converter or inverter, a power status indicator, etc. The electronic device may further include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0244] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0245] The program of a method for constructing a degraded environment map based on adaptive pose estimation stored in the memory 11 of the electronic device is a combination of multiple instructions. When running in the processor 10, it can achieve:

[0246] Obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values;

[0247] Obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data;

[0248] Extract feature points from standard lidar data, perform degraded environment recognition based on the feature points, and obtain degraded environment recognition results;

[0249] Calculate the degradation factor based on the degradation environment identification result, and adjust the feature weight corresponding to the feature point according to the degradation factor;

[0250] Adaptive pose update is performed based on the degraded environment recognition results, and point cloud matching is performed with standard radar data in the adaptive pose to obtain the lidar factor;

[0251] The IMU pre-integration factor and lidar factor are integrated into the factor graph for factor graph optimization, and the pose information is output according to the optimized factor graph. The degraded environment map is obtained based on the pose information and the adaptive pose.

[0252] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.

[0253] Furthermore, if the modules / units integrated into the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0254] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor of an electronic device, the computer program can implement:

[0255] Obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values;

[0256] Obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data;

[0257] Extract feature points from standard lidar data, perform degraded environment recognition based on the feature points, and obtain degraded environment recognition results;

[0258] Calculate the degradation factor based on the degradation environment identification result, and adjust the feature weight corresponding to the feature point according to the degradation factor;

[0259] Adaptive pose update is performed based on the degraded environment recognition results, and point cloud matching is performed with standard radar data in the adaptive pose to obtain the lidar factor;

[0260] The IMU pre-integration factor and lidar factor are integrated into the factor graph for factor graph optimization, and the pose information is output according to the optimized factor graph. The degraded environment map is obtained based on the pose information and the adaptive pose.

[0261] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and other division methods may be used in actual implementation.

[0262] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0263] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0264] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0265] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0266] Blockchain, as used in this article, refers to a novel application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each block contains information about a batch of online transactions, used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0267] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0268] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. Terms such as "first" and "second" are used to indicate names and do not imply any particular order.

[0269] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for constructing a degraded environment map based on adaptive pose estimation, characterized in that: The method comprises: Obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values; Obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data; Extract feature points from standard lidar data, perform degraded environment recognition based on the feature points, and obtain degraded environment recognition results; Calculate the degradation factor based on the degradation environment identification result, and adjust the feature weight corresponding to the feature point according to the degradation factor; Adaptive pose update is performed based on the degraded environment recognition results, and point cloud matching is performed with standard radar data in the adaptive pose to obtain the lidar factor; The IMU pre-integration factor and lidar factor are integrated into the factor graph for factor graph optimization, and the pose information is output according to the optimized factor graph. The degraded environment map is obtained based on the pose information and the adaptive pose.

2. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 1, wherein: Calculating the IMU pre-integration factor according to the IMU measurement value includes: Where Δp ij is the translation of the IMU state from time i to time j, p i is the translation of the IMU state at time i, p j is the translation of the IMU state at time j, Δv ij is the linear velocity change of the IMU state from time i to time j, v i is the linear velocity of the IMU state at time i, g is the gravity vector, ΔR ij is the rotation change of the IMU state from time i to time j, is the transpose of the rotation matrix of the IMU state at time i, R j is the rotation matrix of the IMU state at time j, Δt ij is the time difference from time i to time j, δp ij is the first noise, δv ij is the second noise, δφ ij The third noise.

3. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 1, wherein: The point cloud dedistortion processing is performed on the point cloud data in the lidar data according to the IMU pre-integration factor to obtain standard lidar data, including: Calculate the rotation pose of the point cloud frame in the lidar data in the most recent time frame from the current moment according to the IMU pre-integration factor; The relative rotation amount is calculated based on the rotation pose, and the relative rotation amount is used to correct the pose of the point cloud in the lidar data at the current moment to complete the point cloud distortion processing.

4. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 1, wherein: The extracting of feature points from the standard laser radar data includes: Calculate the curvature of the point cloud set in the standard lidar data, and identify the edge points and plane points in the standard lidar data according to the curvature value; Calculate the angle between the point cloud scan line and the point cloud scan point in the standard lidar data, and identify the ground points in the standard lidar data according to the preset angle threshold.

5. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 4, wherein: After extracting the feature points from the standard lidar data, the method further includes scoring the importance of the feature points and evaluating the effect of feature point extraction based on the importance scores.

6. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 1, wherein: The step of adjusting the feature weight corresponding to the feature point according to the degradation factor includes: Weighting function w of ground points g , weighted function w of edge points e They are defined as: in, is the initial weight of the ground point, is the initial weight of the edge point, α is the first weight magnification coefficient, β is the second weight magnification coefficient, s is the ground score, and D is the final degradation factor.

7. The method for constructing a degraded environment map based on adaptive pose estimation according to claim 1, wherein: The step of performing degraded environment identification based on the feature points to obtain a degraded environment identification result includes: Extract the non-ground point cloud set from the feature points, and perform eigenvalue decomposition on the non-ground point cloud set to obtain the eigenvalue size; Identify whether the current scene is a single feature degraded environment based on the eigenvalue size and the degree of discreteness of the point cloud in the direction of the corresponding eigenvector; Identify whether the current scene is a feature-deficient degraded environment based on the number of feature points and plane points.

8. A device for constructing a degraded environment map based on adaptive pose estimation, characterized in that: The device implements the method for constructing a degraded environment map based on adaptive pose estimation according to any one of claims 1 to 7, and the device includes: A factor calculation module is used to obtain IMU measurement values and calculate IMU pre-integration factors based on the IMU measurement values; obtain lidar data and perform point cloud dedistortion processing on the point cloud data in the lidar data based on the IMU pre-integration factors to obtain standard lidar data; extract feature points in the standard lidar data and perform degraded environment identification based on the feature points to obtain degraded environment identification results; calculate degradation factors based on the degraded environment identification results and adjust the feature weights corresponding to the feature points based on the degradation factors; perform adaptive pose update based on the degraded environment identification results and perform point cloud matching with the standard lidar data in the adaptive pose to obtain the lidar factor; The degraded environment map construction module is used to integrate the IMU pre-integration factor and the lidar factor into the factor graph for factor graph optimization, output the pose information according to the optimized factor graph, and obtain the degraded environment map based on the pose information and the adaptive pose.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for constructing a degraded environment map based on adaptive pose estimation according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for constructing a degraded environment map based on adaptive pose estimation according to any one of claims 1 to 7 is implemented.

Images