• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.44 no.6
• /
• pp.11-20
• /
• 2007

This paper proposes a new efficient method to detect ellipses using a segment merging based Randomized Hough Transform. The key idea of the proposed method is to separate single line segments from an edge image, to estimate ellipses from any pair of the single line segments using Randomized Hough Transform (RHT), and to merge the ellipses. This algorithm is able to accuracy estimate the number of ellipses and largely improves the computational time by reducing iterations.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.27 no.10
• /
• pp.1508-1515
• /
• 2003

In this study, the image processing program for deducing parameters of the elliptic shape of the partially overlapped liquid droplets was developed using the randomized Hough transform and the parameter decomposition. The procedure for the shape detection consists of three steps. For the first step, the candidate centers of ellipses are determined by the geometric property of the ellipse. Next, the rest parameters are estimated by the randomized Hough transform. In the final step for the post-processing, optimally approximated parameters of ellipses are determined. The developed program was applied to the simulated overlapped ellipses, real overlapped droplets, and real spray droplets. The shape detection was very excellent unless there existed inherent problems in original images. Moreover, this method can be used as an effective separating method for the overlapped small particles.

• Proceedings of the KSME Conference
• /
• 2003.04a
• /
• pp.1783-1788
• /
• 2003

In this study, the image processing program for deducing parameters of the elliptic shape of the partially overlapped liquid droplets was developed using the randomized Hough transform and the parameter decomposition. The procedure for the shape detection consists of three steps. For the first step, the candidate centers of ellipses are determined by the geometric property of the ellipse. Next, the rest parameters are estimated by the randomized Hough transform. In the final step for the post-processing, optimally approximated parameters of ellipses are determined. The developed program was applied to the simulated overlapped ellipses, real overlapped droplets, and real spray droplets. The shape detection was very excellent unless there existed inherent problems in original images. Moreover, this method can be used as an effective separating method for the overlapped small particles.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.04a
• /
• pp.157-160
• /
• 2007

본 논문에서는 입력 영상의 에지를 단일 세그먼트로 구성하고 같은 타원에 속하는 에지 세그먼트를 병합하여 타원검출의 속도와 정확도를 향상시키는 방법을 제안한다. 먼저 분기점은 이용한 라벨링 기법과 코너 패턴 정합 기법으로 연속된 화소들의 집합인 에지 세그먼트를 만든다. 구성된 에지 세그먼트와 Randomized Hough Transform에 의해 타원을 추정하여 병합하고 타원을 결정한다. 위 과정으로부터 얻어진 병합된 에지 세그먼트 집합 하나가 타원 하나를 구성하므로 입력 영상 내의 전체 타원의 개수를 정확하게 추정할 수 있다. 또한 전체 에지 화소들로 타원을 검출하는 기존 방법과 달리 분리된 에지 세그먼트 단위로 타원 변수를 결정하기 때문에 전체 수행시간을 크게 줄일 수 있다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.1
• /
• pp.22-30
• /
• 2011

Plane detection is very important information for mission-critical of robot in 3D environment. A representative method of plane detection is Hough-transformation. Hough-transformation is robust to noise and makes the accurate plane detection possible. But it demands excessive memory and takes too much processing time. Iterative randomized Hough-transformation has been proposed to overcome these shortcomings. This method doesn't vote all data. It votes only one value of the randomly selected data into the Hough parameter space. This value calculated the value of the parameter of the shape that we want to extract. In Hough parameters space, it is possible to detect accurate plane through detection of repetitive maximum value. A common problem in these methods is that it requires too much computational cost and large number of memory space to find the distribution of mixed multiple planes in parameter space. In this paper, we detect multiple planes only via data sampling using Self Organizing Map method. It does not use conventional methods that include transforming to Hough parameter space, voting and repetitive plane extraction. And it improves the reliability of plane detection through division area searching and planarity evaluation. The proposed method is more accurate and faster than the conventional methods which is demonstrated the experiments in various conditions.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.11
• /
• pp.2086-2094
• /
• 2008

Finding a planar surface on 3D space is very important for efficient and safe operation of a mobile robot. In this paper, we propose a method using a plane detection cell (PDC) and iterative randomized Hough transform (IRHT) for finding the planar region from a 3D range image. First, the local planar region is detected by a PDC from the target area of the range image. Each plane is then segmented by analyzing the accumulated peaks from voting the local direction and position information of the local PDC in Hough space to reduce effect of noises and outliers and improve the efficiency of the HT. When segmenting each plane region, the IRHT repeatedly decreases the size of the planar region used for voting in the Hough parameter space in order to reduce the effect of noise and solve the local maxima problem in the parameter space. In general, range images have many planes of different normal directions. Hence, we first detected the largest plane region and then the remained region is again processed. Through this procedure, we can segment all planar regions of interest in the range image.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.5
• /
• pp.419-428
• /
• 2011

This paper suggests a judgement method for an inclined plane before entrance of it and the detection of obstacle position. Main idea is started from the assumption that obstacle is always on the bottom plane, and corner appears at this position. The process to detect the obstacle consists of three steps. First the 3D data using stereo matching is acquired to detect an obstacle. Second a bottom plane is extracted by using limit condition. Last the obstacle position is found by using Harris corner detection. Obstacle position detection on an inclined plane was verified by outdoor and indoor experiment. In error analysis, it is confirmed that an average error of obstacle detection in outdoor was larger than the error in indoor but the error are within about 0.030 m. This method will be applied to unmanned vehicles to navigate under various environment.

• Journal of Biosystems Engineering
• /
• v.36 no.5
• /
• pp.377-383
• /
• 2011

This study proposed the new method of discerning the assembled parts and presuming the position of central point in a Cowl Cross Bar (CCB) using a Charge-Couple Device (CCD) camera attached to a robot in the auto assembly line. Three control points of an ellipse were decided by three reference points, which were equally distanced. The radii of these reference points were determined by the size of the object, and the repeated presumption secured the precise determination. The comparison of the central point of ellipse presumed by Randomized Hough Transform (RHT) with the part information stored in a database was used for determining the faulty part in an assembly. The method proposed in this study was applied for the real-time inspection of elliptical parts, such as bolt, nut hole and so on, connected to a CCB using a CCD camera. The findings of this study showed that the precise decision on whether the parts are inferior or not can be made irrespective of the lighting condition of industrial site and the noises of the surface of the part. In addition, the defect decision on the individual elliptic parts assembled in a CCB showed more than 98% accuracy within a 500-millisecond period at most.

• Journal of Korea Multimedia Society
• /
• v.17 no.7
• /
• pp.818-828
• /
• 2014

After emerging of Microsoft Kinect, the interest in three-dimensional (3D) depth image was significantly increased. Depth image data of an object can be converted to 3D coordinates by simple arithmetic calculation and then can be reconstructed as a 3D model on computer. However, because the surface coordinates can be acquired only from the front area facing Kinect, total solid which has a closed surface cannot be reconstructed. In this paper, 3D registration method for multiple Kinects was suggested, in which surface information from each Kinect was simultaneously collected and registered in real time to build 3D total solid. To unify relative coordinate system used by each Kinect, 3D perspective transform was adopted. Also, to detect control points which are necessary to generate transformation matrix, 3D randomized Hough transform was used. Once transform matrices were generated, real time 3D reconstruction of various objects was possible. To verify the usefulness of suggested method, human arms were 3D reconstructed and the volumes of them were measured by using four Kinects. This volume measuring system was developed to monitor the level of lymphedema of patients after cancer treatment and the measurement difference with medical CT was lower than 5%, expected CT reconstruction error.

• Korean Journal of Remote Sensing
• /
• v.31 no.2
• /
• pp.111-125
• /
• 2015

In this study, the target detection using both high-resolution satellite SAR and Elecro-Optical (EO) images such as TerraSAR-X and WorldView-2 is performed, considering the characteristics of targets. The targets of our interest are featured by being stationary and appearing as cluster targets. After the target detection of SAR image by using Constant False Alarm Rate (CFAR) algorithm, a series of processes is performed in order to reduce false alarms, including pixel clustering, network clustering and coherence analysis. We extend further our algorithm by adopting the fast and effective ellipse detection in EO image using randomized hough transform, which is significantly reducing the number of false alarms. The performance of proposed algorithm has been tested and analyzed on TerraSAR-X SAR and WordView-2 EO images. As a result, the average false alarm for group of targets is 1.8 groups/$64km^2$ and the false alarms of single target range from 0.03 to 0.3 targets/$km^2$. The results show that groups of targets are successfully identified with very low false alarms.