• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.796-798
• /
• 2004

기존의 번호판 인식 시스템에서는 대부분 카메라가 고정 상태에서 차량의 전면부를 찍어 영상을 획득하고, 이로부터 번호판을 추출하고 인식한다 그러나 본 연구에서는 기존 연구들과 달리 이동 중인 자동차에 카메라를 설치하여 움직이는 자동차의 영상을 획득하여 번호판을 추출하고 인식한다. 인식하고자 하는 영상이 잡음이나 왜곡 없이 깨끗하다면 인식 과정은 간단하게 수행될 것이다. 그러나, 실제로 얻어진 영상은 간단한 방법으로 인식하기에는 어려올 정도로 왜곡이나 변형이 심한 경우가 많다. 따라서 본 논문에서는 SVM 전단에 모듈라 신경망을 결합하여 인식하는 방법을 사용함으로써 잡음과 같은 변형에 덜 민감하도록 하고자 하였다. 실험결과, 제안하는 분류기를 이용한 방법이 번호판 인식에 우수한 성능을 보임을 확인하였다.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.45 no.2
• /
• pp.8-17
• /
• 2008

This paper proposes a robust license plate extraction method from images taken under unconstrained environments. Utilization of the color and the edge information in complementary fashion makes the proposed method deal with not only various lighting conditions, hilt blocking artifacts frequently observed in compressed images. Computational complexity is significantly reduced by applying Hough transform to estimate the skew angle, and subsequent do-skewing procedure only to the candidate regions. The true plate region is determined from the candidates under examination using clues including the aspect ratio, the number of zero crossings from vertical scan lines, and the number of connected components. The performance of the proposed method is evaluated using compressed images collected under various realistic circumstances. The experimental results show 94.9% of correct license plate extraction rate.

• The KIPS Transactions:PartB
• /
• v.13B no.4 s.107
• /
• pp.353-360
• /
• 2006

This paper presents a robust algorithm for segmenting a vehicle license plate area from a road image. We consider the features of license plates in three aspects : 1) edges due to the characters in the plate, 2) colors in the plate, and 3) geometric properties of the plate. In the preprocessing step, we compute the thresholds based on each feature to decide whether a pixel is inside a plate or not. A statistical approach is applied to the sample images to compute the thresholds. For a given road image, our algorithm binarizes it by using the thresholds. Then, we select three candidate regions to be a plate by searching the binary image with a moving window. The plate area is selected among the candidates with simple heuristics. This algorithm robustly detects the plate against the transformation or the difference of color intensity of the plate in the input image. Moreover, the preprocessing step requires only a small number of sample images for the statistical processing. The experimental results show that the algorithm has 97.8% of successful segmentation of the plate from 228 input images. Our prototype implementation shows average processing time of 0.676 seconds per image for a set of $1280{\times}960$ images, executed on a 3GHz Pentium4 PC with 512M byte memory.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.2
• /
• pp.40-47
• /
• 2009

This paper proposes a method for vehicle license plate detection using mathematical morphology and symmetry. In general, the shape, color, size, and position of license plate are regulated by authorities for a better recognition by human. Among them, the relatively big intensity difference between the letter and the background region of the license plate and the symmetry about the plate are major discriminating factors for the detection. For the first, the opened image is subtracted from the closed image to intensify the region of plate using the rectangular structuring element which has the width of the distance between two characters. Second the subtraction image is average filtered with the mask size of the plate. Third, the column maximum graph of the average filtered image is acquired and the symmetry of the graph is measured at every position. Fourth, the peaks of the average filtered image are searched. Finally, the plate is assumed to be positioned around the one of local maxima nearest to the point of the highest symmetry. About 1,000 images taken by speed regulation camera are used for the experiment. The experimental result shows that the plate detection rate is about 93%.

• Journal of Biomedical Engineering Research
• /
• v.25 no.6
• /
• pp.551-556
• /
• 2004

The purpose of this study was to suggest a systematic and scientific method for measurement of children's growth development, in which the accuracy of existing diagnosis method has not been concretely examined yet. The most popular method for diagnosis of children's growth is to analyze the opening degree of growth plate in each joint by X-ray image. However, X-ray method has some disadvantages; it is impossible to measure the diagnosis of growth periodically and repeatedly due to the radiation problem. Hence, this study introduced a profile analysis and the algorithm of analyzing the image of growth plate with the BUA(Broadband Ultrasound Attenuation) of calcaneus, to verify the possibility of alternative ultrasonic method harmless to human body. We obtained the images of growth plate in proximal tibiae, phalanges, and calcanei of 269 children (7∼16 years old) with X-ray. And the image of growth plate in calcanei was also obtained front those children with ultrasound. The results showed that the time of the opening degree of growth plate in each joint was almost consistent between X-ray and ultrasonic images. Also, the images of growth plate measured by X-ray and ultrasound showed the high correlation. Therefore, it is expected that the algorithm of ultrasonic profile analysis introduced in this study can replace the existing X-ray method to measure the growth plate correctly.

• Journal of the Korean Arthroscopy Society
• /
• v.16 no.2
• /
• pp.147-152
• /
• 2012

Purpose: The purpose of this study was to compare the preoperative magnetic resonance image (MRI) findings with postoperative arthroscopic findings on meniscus injury with anterior cruciate ligament (ACL) rupture. Materials and Methods: We reviewed MRI images and arthroscopic findings of 225 patients treated by ACL reconstruction due to rupture, from February 2001 to November 2010. There were 154 cases of meniscus tear in arthroscopic findings. We examined the sensitivity for detecting meniscal tears varied with the presence of a rupture of the ACL, with the location of the tear within the meniscus, and among configurations of meniscal tears. Results: In the presence of a rupture of the ACL, the sensitivity of MRI was 0.88 for medial meniscal tears and 0.69 for lateral meniscal tears. And sensitivity of MRI was lowest in posterior horn and peripheral portion tears in lateral meniscus injury accompanying ACL rupture, sensitivity was low in anterior horn and flap shape tears in medial meniscus injury accompanying ACL rupture. Conclusion: In Meniscus injury with ACL rupture, a special attention shoulder given to the posterior horn and peripheral portion injury in lateral meniscus during arthroscopic surgery due to difficulty in detecting on MRI.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.8
• /
• pp.1947-1954
• /
• 2009

Camera-captured car plate images contain much variation and noise and the character images in a plate are typically very small. We attempted to improve the plate identification efficiency suitable for this undesirable condition. We experimented various image preprocessing and feature extracting methods and the very effective features that can compensate one feature's limitation is determined through extensive experiments. Finally two very effective features that can complement the limitations of each other feature(classifier) are determined and the efficiency is proved by recognition experiments. This approach is very necessary when handling plate character images which are typically small, various, and noisy. Individual classification result, confidence factor, region name relation and feedback verification are comprehensively considered to enhance the overall recognition efficiency. The efficiency of our method is verified by a recognition experiment using real car plate images taken from traffic roads.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• v.9 no.1
• /
• pp.457-461
• /
• 2005

본 논문에서는 RGB 컬러 정보와 오류 역전파 알고리즘을 이용한 신 차량 번호판 인식 방법을 제안한다. 차량 영상에서 평균 Blue값을 이용하여 차량 영상을 보정한다. 보정된 차량 영상에서 순수 Red픽셀과 현재 픽셀의 차이와 순수 Green 픽셀과 현재의 픽셀의 차이를 각각 구하여 Red 후보 영역과 Green 후보 영역으로 구분한다. 구분된 2개의 후보 영역의 픽셀 값을 오류 역전파 알고리즘에 적용하여 최종 Green 영역을 찾는다. 그리고 오류 역전파 알고리즘에 의해서 Green 영역으로 판명된 영역을 제외한 영역들은 잡음으로 처리한다. 잡음이 제거된 영역에 대해 수평 및 수직 히스토그램의 빈도수를 이용하여 번호판 영역을 추출한다. 추출된 번호판 영역에서 윤곽선 추적 알고리즘을 적용하여 개별 코드들을 추출하고, 오류 역전파 알고리즘을 적용하여 개별 코드들을 인식한다. 제안된 차량 번호판 추출 및 인식 방법의 성능을 평가하기 위하여 실제 비영업용 신 차량 번호판에 적용한 결과, 제안된 번호판 추출 방법이 기존의 HSI 정보를 이용한 번호판 추출 방법보다 추출률이 개선되었고 제안된 차량 번호판 인식 방법이 효율적인 것을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2013.05a
• /
• pp.277-280
• /
• 2013

원자로 건물 내부에는 원자로의 안전운전을 위한 다수의 계통, 구조물 및 기기들이 위치하고 있다. 원자로의 안전과 직결되는 안전계통 변수 (수위, 압력) 들은 경우에 따라서는 현장확인이 요구된다. 원자로 가동 중에 운전원에 의한 안전계통변수의 현장확인이 용이하도록 원자로 건물 내부의 출입구 (2 중문) 부근에 수위계, 압력계 등의 현장계기가 위치하고 있다. 본 논문에서는 일본의 (주) 동경전력이 공개한 후쿠시마 제 1 원자력발전소 1 호기 원자로건물의 IC (비상용 복수기, isolation condenser) 조사영상에 나타나는 현장계기의 영상인식에 대해 기술한다. 조사 영상의 분석에 의하면 현장계기들은 기계식의 아날로그 타입이다. 아날로그 타입의 기하구조를 이용하여 계기판 눈금을 인식하고, 계기판 바늘의 기울기 계산을 통해 계기판을 영상 판독할 수 있었다. 이러한 계기판의 영상판독은 후쿠시마 원전 사고와 같이, 고방사선 피폭 우려로 인해 사람대신에 로봇이 원자로 건물내부에 진입하여 주요 계통, 구조물 및 기기의 현장계기를 판독한다고 가정하면, 유용한 기능이다.

• Journal of Korea Multimedia Society
• /
• v.14 no.9
• /
• pp.1117-1124
• /
• 2011

This paper proposes a plate number extraction method which uses the information of both the colors and the arrangement of numbers in a vehicle image with complex background. In a number plate, color of the numbers is usually black or white, and numbers are arranged in a row. At first, a raw image is partitioned into the plate number candidate regions and non-interest region. The number candidate regions are thresholded in mean binarization. After eliminating the illegal candidate regions using the aspect ratio of the plate number, the plate number region is finally extracted by using the arrangement information among the numbers. To evaluate the proposed mothed, 292 images are taken in various places and at different times. The experimental results show that the rate of the proposed number regions extraction is about 89.8%, 95.5% for the plate of green and white backgrounds, respectively.