Yungeun Choe

Urban structure classification using the 3D normal distribution transform for practical robot applications

Previous urban structure classification methods are intractable for practical robots in two viewpoints: storing point clouds and complex computation for using conditional random fields. This paper presents a classification method based on normal distribution transform (NDT) grids for practical robots. NDT grids store point clouds in the form of the mean and covariance, instead of directly dealing with huge point clouds. By taking the advantage of NDT grids, we design geometric-featured voxel (GFV) based on NDT grids to represent urban structures as a voxel model. The proposed method consists of three steps: GFV generation, segmentation, and classification. In the segmentation, GFVs are clustered according to units of urban structures by spectral clustering. For the classification, the clustered GFVs are classified as one kind of urban structures by supervised learning. Geometric characteristics of urban structures are expressed by a histogram of geometric words. Experimental results prove that the proposed method based on NDT grids is suitable for practical robots in terms of memory requirement, computation time, and even classification accuracy.

Fast point cloud segmentation for an intelligent vehicle using sweeping 2D laser scanners

The previously developed radially bounded nearest neighbor (RBNN) algorithm have been shown a good performance for 3D point cloud segmentation in indoor scenarios. In outdoor scenarios however it is hard to adapt the original RBNN to an intelligent vehicle directly due to several drawbacks. In this paper, drawbacks of RBNN are addressed and we propose an enhanced RBNN for an intelligent vehicle operating in urban environments by proposing the ground elimination and the distance-varying radius. After the ground removal, objects can be remained to segment without merging the ground and objects, whereas the original RBNN with the fixed radius induced over-segmentation or under-segmentation. We design the distance-varying radius which is varied properly from the distance between a laser scanner and scanning objects. The proposed distance-varying radius is successfully induced to segment objects without over or under segmentation. In the experimental results, we have shown that the enhance RBNN is preferable to segment urban structures in terms of time consumption, and even segmentation rates.

Online urban object recognition in point clouds using consecutive point information for urban robotic missions

Urban object recognition is the ability to categorize ambient objects into several classes and it plays an important role in various urban robotic missions, such as surveillance, rescue, and SLAM. However, there were several difficulties when previous studies on urban object recognition in point clouds were adopted for robotic missions: offline-batch processing, deterministic results in classification, and necessity of many training examples. The aim of this paper is to propose an urban object recognition algorithm for urban robotic missions with useful properties: online processing, classification results with probabilistic outputs, and training with a few examples based on a generative model. To achieve this, the proposed algorithm utilizes the consecutive point information (CPI) of a 2D LIDAR sensor. This additional information was useful for designing an online algorithm consisting of segmentation and classification. Experimental results show that the proposed algorithm using CPI enhances the applicability of urban object recognition for various urban robotic missions. The online processing is achieved in order to cope with streaming point cloud data.The classification results contain probabilistic outputs in terms of confidence levels.It is available to train the classifier with a few examples based on generative model.

Geometrical Featured Voxel Based Urban Structure Recognition and 3-D Mapping for Unmanned Ground Vehicle

Recognition of structures in urban environments is a fundamental ability for unmanned ground vehicles. In this paper we propose the geometrical featured voxel which has not only 3-D coordinates but also the type of geometrical properties of point cloud. Instead of dealing with a huge amount of point cloud collected by range sensors in urban, the proposed voxel can efficiently represent and save 3-D urban structures without loss of geometrical properties. We also provide an urban structure classification algorithm by using the proposed voxel and machine learning techniques. The proposed method enables to recognize urban environments around unmanned ground vehicles quickly. In order to evaluate an ability of the proposed map representation and the urban structure classification algorithm, our vehicle equipped with the sensor system collected range data and pose data in campus and experimental results have been shown in this paper.

Mobile robot localization by matching 2D image features to 3D point cloud

In this paper we describe a method for solving a mobile robot localization problem using prior data. By matching 2D image features to a 3D point cloud, the robot position is estimated in the prior point cloud. We prove our method by testing at specific locations over the whole point clod data.

The Development of Sensor System and 3D World Modeling for Autonomous Vehicle

This paper describes a novel sensor system for 3D world modeling of an autonomous vehicle in large-scale outdoor environments. When an autonomous vehicle performs path planning and path following, well-constructed 3D world model of target environment is very important for analyze the environment and track the determined path. To generate well-construct 3D world model, we develop a novel sensor system. The proposed novel sensor system consists of two 2D laser scanners, two single cameras, a DGPS (Differential Global Positioning System) and an IMU (Inertial Measurement System). We verify the effectiveness of the proposed sensor system through experiment in large-scale outdoor environment.

Feature-Based 6-DoF Camera Localization Using Prior Point Cloud and Images

In this paper, we present a new localization algorithm to estimate the localization of a robot based on prior data. Over the past decade, the emergence of numerous ways to utilize various prior data has opened up possibilities for their applications in robotics technologies. However, challenges still remain in estimating a robot’s 6-DoF position by simply analyzing the limited information provided by images from a robot. This paper describes a method of overcoming this technical hurdle by calculating the robot’s 6-DoF location. It only utilizes a current 2D image and prior data, which consists of its corresponding 3D point cloud and images, to calculate the 6-DoF position. Furthermore, we employed the SURF algorithm to find the robot’s position by using the image’s features and the 3D projection method. Experiments were conducted by the loop of 510m trajectory, which is included the prior data. It is expected that our method can be applied to broad areas by using enormous data such as point clouds and street views in the near future.

3D mapping in urban environment using geometric featured voxel

In recent year, much progress has been made in outdoor 3D mapping. However, 3D mapping in real time is still a daunting challenge in urban environment. This paper addresses the problem of 3D mapping from 3D laser scans in urban environments in real time. To do this, we proposed geometric featured voxel which can efficiently represent 3D urban structure without loss of geometric properties. For evaluation of the proposed voxel, we use Oakland dataset.

Position estimation of landmark using 3D point cloud and trilateration method

We demonstrate the appropriacy of using laser distance meter as the accurate, cost-effective, simple solution for position estimation within precise 3D point cloud map in urban environment scenario. Our approach treats trilateration method with minimum error selection algorithm for better position accuracy of landmark. We validate the performance of our approach through 22 landmarks measured using RTK GPS for error analysis of proposed position estimation method based on trilateration.

Building wall extraction method using land registration map image

Digital map can be a good materials for some robotic problems such as localization, segmentation, classification, SLAM algorithm. In this paper, we propose building wall extraction algorithm from land registration map image using corner, edge detection methods and simple matching approach. Experimental results show that this algorithm is valid for general land registration map.