Hyun-Uk Chae

Vehicle License Plate Detection Method Based on Sliding Concentric Windows and Histogram

Detecting the region of a license plate is the key component of the vehicle license plate recognition (VLPR) system. A new method is adopted in this paper to analyze road images which often contain vehicles and extract LP from natural properties by finding vertical and horizontal edges from vehicle region. The proposed vehicle license plate detection (VLPD) method consists of three main stages: (1) a novel adaptive image segmentation technique named as sliding concentric windows (SCWs) used for detecting candidate region; (2) color verification for candidate region by using HSI color model on the basis of using hue and intensity in HSI color model verifying green and yellow LP and white LP, respectively; and (3) finally, decomposing candidate region which contains predetermined LP alphanumeric character by using position histogram to verify and detect vehicle license plate (VLP) region. In the proposed method, input vehicle images are commuted into grey images. Then the candidate regions are found by sliding concentric windows. We detect VLP region which contains predetermined LP color by using HSI color model and LP alphanumeric character by using position histogram. Experimental results show that the proposed method is very effective in coping with different conditions such as poor illumination, varied distances from the vehicle and varied weather.

Parallelogram and Histogram based Vehicle License Plate Detection

This paper describes a new approach to analyze road images which often contain vehicles and extract license plate (LP) from natural properties by finding vertical and horizontal edges from vehicle region. The proposed technique consists of three main modules: (a) segmentation technique named as sliding concentric windows (SCW) on the basis of a novel adaptive image for detecting candidate region, (b) refining by using HSI color model on the basis of using hue and intensity in HSI color model verifying green and yellow LP and white LP, respectively and (c) finally, verify and detect VLP region which contains predetermined LP alphanumeric character by using position histogram. In the proposed method, input vehicle images are converted into gray images. After then the candidate regions are found by sliding concentric windows. We detect vehicle license plates (VLP) region which contains predetermined LP color by using HSI color model and LP alphanumeric character by using position histogram. Experimental results show that the proposed method is very effective in coping with different conditions such as poor illumination and varied weather. Experimental results also show that the distance from the vehicle varied according to the camera setup.

Multiple Camera-Based Correspondence of Ground Foot for Human Motion Tracking

In this paper, we describe correspondence among multiple images taken by multiple cameras. The correspondence among multiple views is an interesting problem which often appears in the application like visual surveillance or gesture recognition system. We use the principal axis and the ground plane homography to estimate foot of human. The principal axis belongs to the subtracted silhouette-based region of human using subtraction of the predetermined multiple background models with current image which includes moving person. For the calculation of the ground plane homography, we use landmarks on the ground plane in 3D space. Thus the ground plane homography means the relation of two common points in different views. In the normal human being, the foot of human has an exactly same position in the 3D space and we represent it to the intersection in this paper. The intersection occurs when the principal axis in an image crosses to the transformed ground plane from other image. However the positions of the intersection are different depend on camera views. Therefore we construct the correspondence that means the relationship between the intersection in current image and the transformed intersection from other image by homography. Those correspondences should confirm within a short distance measuring in the top viewed plane. Thus, we track a person by these corresponding points on the ground plane. Experimental result shows the accuracy of the proposed algorithm has almost 90% of detecting person for tracking based on correspondence of intersections.

An Informationally Structured Room for Robotic Assistance

The application of assistive technologies for elderly people is one of the most promising and interesting scenarios for intelligent technologies in the present and near future. Moreover, the improvement of the quality of life for the elderly is one of the first priorities in modern countries and societies. In this work, we present an informationally structured room that is aimed at supporting the daily life activities of elderly people. This room integrates different sensor modalities in a natural and non-invasive way inside the environment. The information gathered by the sensors is processed and sent to a centralized management system, which makes it available to a service robot assisting the people. One important restriction of our intelligent room is reducing as much as possible any interference with daily activities. Finally, this paper presents several experiments and situations using our intelligent environment in cooperation with our service robot.

The Intelligent Room for Elderly Care

Daily life assistance for elderly is one of the most promising and interesting scenarios for advanced technologies in the near future. Improving the quality of life of elderly is also some of the first priorities in modern countries and societies where the percentage of elder people is rapidly increasing due mainly to great improvements in medicine during the last decades. In this paper, we present an overview of our informationally structured room that supports daily life activities of elderly with the aim of improving their quality of life. Our environment contains different distributed sensors including a floor sensing system and several intelligent cabinets. Sensor information is sent to a centralized management system which processes the data and makes it available to a service robot which assists the people in the room. One important restriction in our intelligent environment is to maintain a small number of sensors to avoid interfering with the daily activities of people and to reduce as much as possible the invasion of their privacy. In addition we discuss some experiments using our real environment and robot.

Identification of a human using accorded blobs on the varied region from image sequence by multiple cameras

This paper proposes a method to identify the same person in different-view images using plane homography. We address the problem of correspondence under multiple uncalibrated cameras. For solving this problem, we propose a three steps process. The first, we detect intensity-varied regions in the each image sequence by background subtraction. The background model has the temporal-median value for each pixel and it updates by itself in the process. The temporal difference between consecutive images contributes to detect accurately in the result of background subtraction. The second step is generating the set of blobs from detected varied regions in the each image sequence. Those blobs are generated using criterion with specific color range. Each blob is described probability distribution with color and location information in the image plane. The final step is matching the current color blobs with other blobs in the different images using maximum likelihood estimation. The matched color blobs are reconstructed to the human silhouette by context information that includes restriction conditions on the human body. Finally we identify matched human in the different images without calibration.

Appearance feature based human correspondence under non-overlapping views

In this paper, a method is proposed, to solve correspondence problem under structured space which is installed multiple cameras. The correspondence between different cameras is an important task to use the multiple camera system. For solving this problem, the proposed method is consists of three steps which are detection of moving object, feature extraction and correspondence among different cameras. First step is to detect moving people by background subtraction from multiple background model. The temporal difference is used jointly to remove noise occurred from temporary change. The detected regions are divided using labeling as individual person. The second step is to segment the each person by a criterion with appearance and context information. The segmented regions in a person are estimated as Gaussian mixture model (GMM) for correspondence. The final step is process of correspondence between different cameras. A GMM from a camera is matched with another GMM from other cameras. A ratio of those GMMs is used as a criteria to identify same person. The experiment was performed with the specific scenarios in quantitative results.

Multiple Camera-based Person Correspondence using Color Distribution and Context Information of Human Body

In this paper, we proposed a method which corresponds people under the structured spaces with multiple cameras. The correspondence takes an important role for using multiple camera system. For solving this correspondence, the proposed method consists of three main steps. Firstly, moving objects are detected by background subtraction using a multiple background model. The temporal difference is simultaneously used to reduce a noise in the temporal change. When more than two people are detected, those detected regions are divided into each label to represent an individual person. Secondly, the detected region is segmented as features for correspondence by a criterion with the color distribution and context information of human body. The segmented region is represented as a set of blobs. Each blob is described as Gaussian probability distribution, i.e., a person model is generated from the blobs as a Gaussian Mixture Model (GMM). Finally, a GMM of each person from a camera is matched with the model of other people from different cameras by maximum likelihood. From those results, we identify a same person in different view. The experiment was performed according to three scenarios and verified the performance in qualitative and quantitative results.

Comparison of Vehicle Control Systems Based on Multiple Features

In striving to achieve autonomous navigation, the guidance system plays an essential role. In this article, the authors present two recently developed vision-based methods to estimate the heading angle of the vehicle. The authors propose a real-time guidance application based on edge and color information using an omnidirectional image. First, line segments were extracted. Second, the Random Sample Consensus (RANSAC) curve fitting method was implemented. Third, the set of intersection points for each pair of curves was extracted. Finally, by implementing the Density Based Spatial Clustering of Applications with Noise (DBSCAN) algorithms, the heading angle was computed. The main contribution of this work is in the evaluation of the methods applied, which included the “edge-based extraction in road scenes” method, which uses the spatial density information around the vehicle (curbs, barrier, gutters, side strip areas), and the “edge-based lane marking segmentation.” Both of these methods were evaluated in terms of the performance assessment of the length of the line segments in relation to the heading angle measurement. In that sense, the experiments were conducted for the purpose of testing the processing time and the number of extracted line segments. The preliminary results were gathered and tested on a group of consecutive frames to demonstrate the effectiveness of the proposed methods.

Auto-surveillance for object to bring in/out using multiple camera

This paper describes an auto-surveillance system which tracks a person who comes in/out an office using multiple camera system. Furthermore it automatically recognize whether the person bring an object in/out. For this purpose, we set three steps. The first step is detecting a person using MBM(Multiple Background Model) and TMB(Temporal Median Background). The second step is calculation of correspondence between persons detected by different view-point cameras in the multiple camera system. We simply calculate the correspondence based on the principal axis and homography. The last step is generating global color model, which includes every local color model organized by GMM (Gaussian Mixture Model) from each camera, of the person. The global color model represented by GMM checks the temporally varied error and detects the object to bring in or out objects. In the experiment, we show the detected human silhouette by background subtraction and the tracking result by correspondence of multiple views. We also show the color segmentation using GMM and the recognition result for detecting objects brought in/out by the tracked person.