Camera Point Cloud Research Articles

Bush spherical center detection algorithm based on depth camera 3D point cloud

Automated pruning is an inevitable trend in the improvement of modern gardens. In order to provide necessary information for automatic garden robots and satisfy the requirement of target detection and positioning during pruning, this paper proposed a bush spherical center detection algorithm based on a 3D depth camera point cloud. Firstly, the depth camera collected the bush image, and the results were aligned to the depth image to obtain the 3D point cloud of bush. Then the ROI was extracted by preprocessing, and the 3D point clouds of bush was obtained after filtering and coordinate transformation. Finally, the spherical center coordinates of the bush were extracted by the minimum bounding box method. Four groups of tests on the bush spherical coordinates detection were carried out outdoors. The maximum location error and the minimum location error of the spherical bush center were 10.23mm and 8.65 mm, respectively, and the average location error was 9.51mm. The bush spherical center detection algorithm based on depth camera 3D point clouds proposed in this paper provides a technical reference for the information acquisition of automatic pruning robot.

Automatic Registration Algorithm for the Point Clouds Based on the Optimized RANSAC and IWOA Algorithms for Robotic Manufacturing

In order to solve the problems of low accuracy and low efficiency of point cloud registration for stereo camera systems, we propose a binocular stereo camera point cloud registration method based on IWOA and Improved ICP. We propose the following approaches in this paper—the registration process is divided into two steps to complete the initial coarse registration and the exact registration. In the initial registration stage, an improved Whale Optimization Algorithm (IWOA) based on nonlinear convergence factor and adaptive weight coefficients was proposed to realize the initial registration in combination with the RANSAC algorithm, and the obtained transformation matrix was used as the initial estimate of the subsequent exact registration algorithm. In the second step of the exact registration stage, an IICP algorithm with the introduction of normal vector weighting constraints at key points was proposed for achieving point cloud exact registration. This algorithm was verified by using Stanford point clouds (bunnies and monkeys) and our own point clouds algorithm, and the proposed algorithm in this paper has high registration accuracy, improved registration speed, and convergence speed.

Event Camera Point Cloud Feature Analysis and Shadow Removal for Road Traffic Sensing

Event cameras have a wide range of applications in the field of traffic flow detection owing to their ability to accurately identify moving targets while being insensitive to stationary targets. In this study, an event camera is installed in a roadside scenario to collect the point cloud data of moving targets. For traffic identification, the geometrical, quantitative and Gaussian projection characteristics of the point clouds for motor/non-motor vehicles and pedestrians are extracted, and their feature distributions are analyzed. Furthermore, to address the problem of shadow noise caused by sunlight, a shadow removal method based on feature similarity is proposed considering the point cloud distribution characteristics. Experimental results show that there are significant differences in the length-width ratio, pixel points, horizontal and vertical projection features among different traffic targets specified in vehicles, non-motor vehicles and pedestrians. In addition, the proposed shadow removal scheme demonstrates high accuracy of 96.5%, and the standard deviation is 1.3%.

Application of fusion 2D lidar and binocular vision in robot locating obstacles

If robot uses 2D lidar or binocular camera to locate obstacles, there will be some problems such as missing obstacle information or inaccurate obstacle locating, which will affect the normal work of the robot. In order to obtain accurate 3D obstacle information, this paper proposes an algorithm for fusing 2D lidar and binocular vision to complete the obstacle location. In this paper, the depth value of the 2D lidar point cloud is used as a benchmark. By fitting the error equation of the binocular camera point cloud depth value, the depth value of the camera point cloud is modified to obtain an accurate 3D camera point cloud, thereby obtaining an accurate 3D obstacle information. Many experiments have proved that the fusion algorithm of 2D lidar and binocular vision can obtain accurate 3D obstacle information. The method of fusion 2D lidar and binocular vision can approximately achieve the measurement effect of 3D lidar, and the point cloud of obstacles is relatively dense, so the accurate 3D obstacle information can be obtained. This method can reduce the influence of single sensor on the robot locating obstacles, thus completing the accurate locating of obstacles, which is of certain significance to robot navigation.

Research on SLAM Algorithm of Mobile Robot Based on the Fusion of 2D LiDAR and Depth Camera

This paper proposes a new Simultaneous Localization and Mapping (SLAM) method on the basis of graph-based optimization through the combination of the Light Detection and Ranging (LiDAR), RGB-D camera, encoder and Inertial Measurement Unit (IMU). It can conduct joint positioning of four sensors by taking advantaging of the unscented Kalman filter (UKF) to design the related strategy of the 2D LiDAR point cloud and RGB-D camera point cloud. 3D LiDAR point cloud information generated by the RGB-D camera under the 2D LiDAR has been added into the new SLAM method in the sequential registration stage, and it can match the 2D LiDAR point cloud and the 3D RGB-D point cloud by using the method of the Correlation Scan Matching (CSM); In the loop closure detection stage, this method can further verify the accuracy of the loop closure after the 2D LiDAR matching by describing 3D point cloud. Additionally, this new SLAM method has been verified feasibility and availability through the processes of theoretical derivation, simulation experiment and physical verification. As a result, the experiment shows that the multi-sensor SLAM framework designed has a good mapping effect, high precision and accuracy.

An Indoor Wayfinding System Based on Geometric Features Aided Graph SLAM for the Visually Impaired.

This paper presents a 6-degree of freedom (DOF) pose estimation (PE) method and an indoor wayfinding system based on the method for the visually impaired. The PE method involves two-graph simultaneous localization and mapping (SLAM) processes to reduce the accumulative pose error of the device. In the first step, the floor plane is extracted from the 3-D camera's point cloud and added as a landmark node into the graph for 6-DOF SLAM to reduce roll, pitch, and Z errors. In the second step, the wall lines are extracted and incorporated into the graph for 3-DOF SLAM to reduce X , Y , and yaw errors. The method reduces the 6-DOF pose error and results in more accurate pose with less computational time than the state-of-the-art planar SLAM methods. Based on the PE method, a wayfinding system is developed for navigating a visually impaired person in an indoor environment. The system uses the estimated pose and floor plan to locate the device user in a building and guides the user by announcing the points of interest and navigational commands through a speech interface. Experimental results validate the effectiveness of the PE method and demonstrate that the system may substantially ease an indoor navigation task.

ONE STEP MOBILE MAPPING LASER AND CAMERA DATA ORIENTATION AND CALIBRATION

Abstract. Modern mobile mapping systems include one or several laser scanners and cameras. The main outputs of these systems are oriented camera images and 3D point clouds. These point clouds can be derived from pairs of overlapping images or using the laser raw data together with platform trajectory. A mobile mapping campaign may include several overlapping areas, generally, the derived point clouds of the same area are not properly registered, due to partial or total GNSS occlusions, multipath and inertial drift and noise. Nowadays, the standard procedure for co-registration between laser and laser and between camera and laser, includes several steps. The first ones are the system calibration where the lever arm and boresight between laser and IMU, and between camera and IMU must be determined. After the calibration steps, a camera and LiDAR point cloud can be derived. Then, a co-registration between LIDAR points clouds and between camera point cloud and LiDAR point cloud are computed. In contrast to the standard approach, in this paper we propose to solve the orientation and calibration of laser and camera data in a single, combined adjustment. Solving the orientation and calibration allows us to implicitly deal with the co-registration problem. The proposed method is based on the identification of common tie features between images and point clouds and their use in a combined adjustment. This common tie features are straight line segments. The preliminary results indicate the feasibility and the potential of the approach.