Hernsoo Hahn

Real-Time Lane Departure Detection Based on Extended Edge-Linking Algorithm

Lane detection can provide important information for safety driving. In this paper, a real time vision-based lane detection method is presented to find the position and type of lanes in each video frame. In the proposed lane detection method, lane hypothesis is generated and verified based on an effective combination of lane-mark edge-link features. First, lane-mark candidates are searched inside region of interest (ROI). During this searching process, an extended edge-linking algorithm with directional edge-gap closing is used to produce more complete edge-links, and features like lane-mark edge orientation and lane-mark width are used to select candidate lane-mark edge-link pairs. For the verification of lane-mark candidates, color is checked inside the region enclosed by candidate edge-link pairs in YUV color space. Additionally, the continuity of the lane is estimated employing a Bayesian probability model based on lane-mark color and edge-link length ratio. Finally, a simple lane departure model is built to detect lane departures based on lane locations in the image Experiment results show that the proposed lane detection method can work robustly in real-time, and can achieve an average speed of 30~50ms per frame for 180x120 image size, with a correct detection rate over 92%.

A new algorithm for ellipse detection by curve segments

This paper proposes a new ellipse detection scheme using curve segments. It detects curve segments in an edge image and selects every pair of them to test whether they pertain to the same ellipse or not. If they pass the test, they are merged as a new segment. Since the proposed algorithm uses curve segments, it reduces the computation time of the conventional algorithm significantly, and detects all ellipses included in an image without missing. The experimental results have shown that its performance is more successful than other approaches in detection of ellipses when they are overlapped and partially occluded.

Visual tracking of a moving target using active contour based SSD algorithm

Abstract This paper presents a new image based visual tracking scheme for a mobile robot to trace a moving target using a single camera mounted on the mobile robot. To accurately estimate the position of the target in the next image, it decomposes the effect of the camera motion on the velocity vector of the target in the image frame. Based on the estimated velocity of the target and the image Jacobian, the control inputs of the mobile robot are determined in such a way that the target may appear inside the central area of the image frame. Since the shape of the target in the image frame varies due to rotation and translation of the target, a new shape adaptive Sum-of-Squared Difference (SSD) algorithm is proposed which uses the extended snake algorithm to extract the contour of the target and updates the template in every step of the matching process. The proposed scheme has been implemented using a Nomad Scout Robot II. The experimental results have shown that the proposed scheme follows the target within a negligible error range even when the target is temporarily lost due to various reasons.

Remote Detection and Monitoring of a Water Level Using Narrow Band Channel

To measure water levels from remote sites using a narrowband channel, this paper propose a difference image based JPEG communication scheme and water level measurement scheme using sparsely sampled images in time domain. In the slave system located in the field, the image is converted to difference image, and are compressed using JPEG, then larger changes are sampled and transmitted. To measure the water level from the images received in the master system which may contain noises caused by various sources, the averaging filter and Gaussian filter are used to reduce the noise and the Y-axis profile of an edge image is used to read the water level. Considering the wild condition of the field, a simplified camera calibration scheme is also introduced. The implemented slave system was installed in a river and its performance has been tested with the data collected for a year.

License Plate Detection Using Topology of Characters and Outer Contour

This paper presents a license plate detection method for detecting license plates on vehicles, using the topology of painted numbers and letters and the outer contour of the license plate. Even in the situation of poor illumination, the contrast is often high between the background and the characters on the license plate; this makes the detection of characters on the license plate possible, when the license plate is free from occlusion within the image. The characters are detected using the topological information from the character candidates, and then the license plate is detected by extracting the vertical and horizontal edges in proximity of the detected characterpsilas region. In the case of a failure to detect the license plate region by the edge detection method, an extended region encompassing the detected character is converted to a binary image, and then all the characters present on the license plate are detected. We verify the performance of this algorithm by experiments on a large dataset of images with varying complexity.

Fuzzy Controller based Biped Robot Balance Control using 3D Image

In this paper, we propose a balance control algorithm based on fuzzy controller using 3D (Dimensional) geometric information obtained from sequential images. The pose of a biped robot is acquired from image information instead of different sensors, and this pose information is used for fuzzy controller to keep biped robot in a stable balanced condition. Initially, the camera mounted on the head of robot obtains the pose information of the target object while walking. We acquired corresponding points between characteristic features on the target object and CAD (Computer Aided Design) information in DB (Data Base) using SURF (Speeded Up Robust Features) algorithm. Three arbitrary features from SURF algorithm are able to generate a virtual plane, which represent the geometric information of the target objects pose data. The camera pose information for the target object is calculated using pose information derived from the virtual planes. As the camera pose information is related to pose of a biped robot, we can easily get the pose information of a biped robot. With this the difference between current and desired pose of a biped robot is obtained. This difference is used as input data for the fuzzy controller. The fuzzy controller keeps the biped robot in stable pose using the recent pose and velocity of a biped robot. The efficiency of the proposed algorithm has been proven by the experiments performed on even floor using a mini humanoid robot.

Natural Gait Generation of Biped Robot based on Analysis of Human's Gait

This paper proposes a method of adaptively generation a natural gait pattern of biped robot. The gait synthesis is based on humans gait pattern analysis. The proposed method can easily be applied to generate the natural and stable gait pattern of any biped robot. To analyze the humans gait pattern, sequential images of the humans gait on the sagittal and frontal plane are acquired from which the gait control values are extracted. The gait pattern of biped robot on the sagittal plane is adaptively generated by a genetic algorithm using the humans joint torque energy. However, gait trajectories of the biped robot on the sagittal plane are not enough to construct the complete gait pattern because the biped robot moves on 3-dimension space. Therefore, the gait pattern on the frontal plane, generated using Zero Moment Point(ZMP) and humans frontal plane gait pattern, is added to the gait one acquired on the sagittal plane. Consequently, the natural and stable walking pattern for the biped robot is obtained, as proved by the experiments.

Recognition of 3d object using attributed relation graph of silhouette's extended convex hull

This paper presents a new approach of recognizing a 3D object using a single camera, based on the extended convex hull of its silhouette. It aims at minimizing the DB size and simplifying the processes for matching and feature extraction. For this purpose, two concepts are introduced: extended convex hull and measurable region. Extended convex hull is consisted of convex curved edges as well as convex polygons. Measurable region is the cluster of the viewing vectors of a camera represented as the points on the orientation sphere from which a specific set of surfaces can be measured. A measurable region is represented by the extended convex hull of the silhouette which can be obtained by viewing the object from the center of the measurable region. Each silhouette is represented by a relation graph where a node describes an edge using its type, length, reality, and components. Experimental results are included to show that the proposed algorithm works efficiently even when the objects are overlapped and partially occluded. The time complexity for searching the object model in the database is O(N) where N is the number of silhouette models.

Top-View-Based Guidance for Blind People Using Directional Ellipse Model

The guidance system proposed in this paper aims to complement the white cane by monitoring road conditions in a medium range for blind pedestrians in real time. The system prototype employs only one webcam fixed at the waist of user. One of the main difficulties of using a single camera in outdoor obstacle detection is the discrimination of obstacles from a complex background. To solve this problem, this paper re-formulates top-view mapping as an inhomogeneous re-sampling process, so that background edges are sub-sampled while obstacle edges are oversampled in the top-view domain. Morphology filters are then used to enhance obstacle edges as edge-blobs, which are further represented using a directional ellipse as a new model for obstacle classification. Based on the identified obstacles, safe walking area is estimated by tracking a polar edge-blob histogram. To transfer the information obtained from image domain to language domain, this paper proposes a verbal message generation scheme based on fuzzy logic. The efficiency of the system is confirmed by testing the system with visually impaired people on outdoor pedestrian paths.

An Efficient Extraction of On-Road Object and Lane Information Using Representation Method

Robust and reliable lane detection is an important issue for driver assistance systems and self-guided systems. Input images acquired from a camera on moving vehicle need to be processed for lane and vehicular traffic detection. We propose an algorithm that uses a chain code detector for detecting lanes and detects vehicles using the horizontal lines formed by the vehicles at the edge of the images. In the first step, a vanishing point is obtained from both sides of the lane, which are obtained from the chain code analysis using the canny edge operator. In the second step, the driveways are decided by combining the vanishing point and the lane information on both sides of the moving vehicle. We define the lane properties and the limiting conditions for reliable lane detection. The robustness of proposed algorithm has been verified from experimental results of complex background images.