• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.4
• /
• pp.10-21
• /
• 2008

The present paper propose a real-time moving object recognition and tracking method using the wavelet-based neural network and invariant moments. Candidate moving region detection phase which is the first step of the proposed method detects the candidate regions where a pixel value changes occur due to object movement based on the difference image analysis between continued two image frames. The object recognition phase which is second step of proposed method recognizes the vehicle regions from the detected candidate regions using wavelet neurual-network. From object tracking Phase which is third step the recognized vehicle regions tracks using matching methods of wavelet invariant moments bases to recognized object. To detect a moving object from image sequence the candidate regions detection phase uses an adaptive thresholding method between previous image and current image as result it was robust surroundings environmental change and moving object detections were possible. And by using wavelet features to recognize and tracking of vehicle, the proposed method decrease calculation time and not only it will be able to minimize the effect in compliance with noise of road image, vehicle recognition accuracy became improved. The result which it experiments from the image which it acquires from the general road image sequence and vehicle detection rate is 92.8%, the computing time per frame is 0.24 seconds. The proposed method can be efficiently apply to a real-time intelligence road traffic surveillance system.

• Journal of the Korea Convergence Society
• /
• v.12 no.1
• /
• pp.29-35
• /
• 2021

License plate recognition plays a key role in intelligent transportation systems. Therefore, it is a very important process to efficiently detect the number and character areas. In this paper, we propose a method to effectively detect license plate number area by applying deep learning and semantic image segmentation algorithm. The proposed method is an algorithm that detects number and text areas directly from the license plate without preprocessing such as pixel projection. The license plate image was acquired from a fixed camera installed on the road, and was used in various real situations taking into account both weather and lighting changes. The input images was normalized to reduce the color change, and the deep learning neural networks used in the experiment were Vgg16, Vgg19, ResNet18, and ResNet50. To examine the performance of the proposed method, we experimented with 500 license plate images. 300 sheets were used for learning and 200 sheets were used for testing. As a result of computer simulation, it was the best when using ResNet50, and 95.77% accuracy was obtained.

• KIISE Transactions on Computing Practices
• /
• v.22 no.1
• /
• pp.8-19
• /
• 2016

Potholes are caused by the presence of water in the underlying soil structure, which weakens the road pavement by expansion and contraction of water at freezing and thawing temperatures. Recently, automatic pothole detection systems have been studied, such as vibration-based methods and laser scanning methods. However, the vibration-based methods have low detection accuracy and limited detection area. Moreover, the costs for laser scanning-based methods are significantly high. Thus, in this paper, we propose a new pothole detection system using a commercial black-box camera. Normally, the computing power of a commercial black-box camera is limited. Thus, the pothole detection algorithm should be designed to work with the embedded computing environment of a black-box camera. The designed pothole detection algorithm has been tested by implementing in a black-box camera. The experimental results are analyzed with specific evaluation metrics, such as sensitivity and precision. Our studies confirm that the proposed pothole detection system can be utilized to gather pothole information in real-time.

• Journal of Korean Society of Transportation
• /
• v.17 no.1
• /
• pp.105-112
• /
• 1999

Image Processing Technique has been used as an efficient method to collect traffic information on the road such as vehicle counts, speed, queues, congestion and incidents. Most of the current methods which have been used to detect vehicles by the image processing are based on point processing, dealing with the local gray level of each pixel in the small window. However, these methods have some drawbacks. Firstly, detection is restricted by image quality. Secondly, they can not deal with occlusion and perspective projection problems, In this research, a new method which possibly deals with occlusion and perspective problems will be proposed. It extracts spatial information such as the position, the relationship of vehicles in 3-dimensional space, as well as vehicle detection in the image. The main algorithm used in this research is based on an extension of the Hough Transform. The Hough Transform which is proposed to estimates parameters of vertices and directed edges analytically on the Hough Space, is a valuable method for the 3-dimensional analysis of static scenes, motion detection and the estimation of viewing parameters.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.10a
• /
• pp.645-648
• /
• 2003

This paper presents an active vision system for on-line traffic sign recognition. The system is composed of two cameras, one is equipped with a wide-angle lens and the other with a telephoto lends, and a PC with an image processing board. The system first detects candidates for traffic signs in the wide-angle image using color, intensity, and shape information. For each candidate, the telephoto-camera is directed to its predicted position to capture the candidate in a large size in the image. The recognition algorithm is designed by intensively using built in functions of an off-the-shelf mage processing board to realize both easy implementation and fast recognition. The results of on-road experiments show the feasibility of the system.

• Annual Conference on Human and Language Technology
• /
• 2004.10d
• /
• pp.20-24
• /
• 2004

본 논문에서는 환경 잡음과 원거리 음성 입력 그리고 노인 발화 등의 열악한 음성 인식 환경에서의 음성 언어이해(spoken language understanding)를 위한 정보 추출 접근 방식에 대해 논하고 있다. 정보 추출의 목적은 미리 정의된 slot에 적절한 값을 찾는 것이다. 음성 언어 이해를 위한 정보 추출은 필수적인 요소만을 추출하는 것을 목적으로 하는 개념 집어내기(concept spotting) 접근 방식을 사용한다. 이러한 방식은 미리 정의된 개념 구조 slot에만 관심을 가지기 때문에. 음성 언어 이해에서 사용되는 정보 추출은 언어를 완전히 이해한다기보다는 부분적으로 이해하는 방식을 취하고 있다. 음성 입력 언어는 주로 열등한 인식 환경에서 이루어지기 때문에 많은 인식 오류를 가지고 이로 인해 텍스트 입력에 비해 이해하기 어렵다. 이러한 점을 고려하여, 특정 정보에 집중함으로써 음성 언어를 이해하고자 시도하였다. 도로 정보 안내 영역을 대상으로 한 실험에서 텍스트 입력(WER 0%)과 음성 입력(WER 39.0%)이 주어졌을 때, 개념 집어내기 방식의 F-measure 값은 각각 0.945, 0.823을 나타내었다.

• Proceedings of the Korean Association of Geographic Inforamtion Studies Conference
• /
• 2010.06a
• /
• pp.183-183
• /
• 2010

위치기반서비스(Location Based Service)는 갈수록 고도화 되어 가고 있다. 특히 최근의 대형 포털을 중심으로 지오웹 서비스가 활성화 되어 있고 이를 스마트폰과 같은 개인용 이용기기를 통해 연속적으로 제공하려는 경향이 뚜렷하다. 이와 같은 시점에서 정부와 민간에서 구축 중이거나 보유 중인 전국적 규모의 데이터 간 상호 연동과 융합을 도모하려는 시도 또한 불가결하다. 이는 고도화된 LBS를 위하여 반드시 필요한 과정이기 때문이다. 이에 따라 몇 가지 주요한 전국 데이터를 대상으로 상호 연동과 융합을 위한 기술개발을 시도하였다. 우선 도로명주소기본도와 수치지형도 간 POI의 연계를 위한 연구를 수행하고 있다. 이 연구에서는 두도면 내의 POI를 대상으로 다양한 매칭과 이에 기반 한 의사결정 방법론을 이용하여 자동으로 상호 인식 및 연계가 될 수 있도록 하고 있다. 다음으로 지적도와 수치지형도 간의 객체 매칭에 관한 연구이다. 수치지형도와 지적도의 불부합으로 인하여 그 동안 지적도를 수치지형도에 맞춘 형태의 편집지적도를 지속적으로 생산하여 왔고 앞으로도 그럴 것이다. 문제는 여기에 필요한 많은 예산이다. 만일 수치지형도와 지적도를 자동으로 매칭하여 편집지적도를 자동으로 생산할 수 있게 된다면 많은 예산 절감과 함께 편집지적도의 현시성을 확보할 수 있게 될 것이다. 다음으로 항공사진과 도로망도의 매칭이다. 현재 주요 포털에서 제공하고 있는 항공사진 기반의 도로망도는 기복변위와 같은 문제로 인하여 시각적으로 많은 위치오차를 보이고 있다. 만일 항공사진의 도로영역을 자동으로 추출하여 벡터 도로망도와 매칭을 할 수 있다면 보다 시각적으로 안정된 항공사진 상의 도로망도를 제공할 수 있게 되고 나아가 이는 차량이나 보행자 네비게이션에 매우 요긴하게 이용될 수 있을 것이다. 다음으로 서로 LOD가 다른 도로망도의 매칭 문제이다. 많은 기관에서 독자적으로 생산한 도로망도는 LOD의 상이에 기인한 문제가 많아 서로 연계 활용되지 않는다. 이를 자동으로 매칭하여 서로 연계할 수 있다면 두 도로망도가 보유하고 있는 속성정보를 공동으로 이용할 수 있는 이익을 얻게 된다. 다음으로 지도 일반화 기술이다. 지도일반화는 지적도내 수치지형도와 같은 대규모 데이터를 스마트폰과 같은 저용량 사양의 기기에 서비스 할 때 불가결한 기술이다. 지도상 객체들의 기하학적 정보 손실을 최소화하면서 메모리 측면에서 경량의 지도를 자동으로 만들어 낸다면 이는 매우 요긴하게 이용될 것이다. 마지막으로 보행자 네트워크의 생성기술이다. 보행자 네트워크는 그 상세함과 정보용량에 있어서 차량용 네트워크에 견줄 수 없다. 이를 현행의 차량용 네트워크와 같이 수동으로 생성하는 데에는 경제적으로나 시간적으로 막대한 투자가 필요하다. 따라서 이를 기존의 공간정보들을 활용하여 자동으로 생성해 낼 수 있다면 그 파급효과는 매우 크리라 판단된다. 본 발표에서는 위와 같은 주제에 관하여 그간의 연구 성과를 개략적으로 소개해본다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.39 no.2
• /
• pp.235-240
• /
• 2015

Active Safety system is requested on the market of the medium and heavy duty commercial vehicle over 4.5ton beside the market of passenger car with advancement of the digital equipment proportionally. Unlike the passenger car, the mounting position of camera in case of the medium and heavy duty commercial vehicle is relatively high, it is disadvantaged conditions for lane recognition in contradiction to passenger car. In this work, we show the method of lane recognition through the Sobel edge, based on the spatial domain processing, Hough transform and color conversion correction. Also we suggest the low error method of front vehicles recognition in order to reduce the detection error through Haar-like, Adaboost, SVM and Template matching, etc., which are the object recognition methods by frontal camera vision. It is verified that the reliability over 98% on lane recognition is obtained through the vehicle test.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.9
• /
• pp.5446-5451
• /
• 2014

Because there are some mistakes by hand in processing electronic maps using a navigation terminal, this paper proposes an automatic offline recognition for traffic signs, which are considered ingredient navigation information. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), which have been used widely in the field of 2D face recognition as computer vision and pattern recognition applications, was used to recognize traffic signs. First, using PCA, a high-dimensional 2D image data was projected to a low-dimensional feature vector. The LDA maximized the between scatter matrix and minimized the within scatter matrix using the low-dimensional feature vector obtained from PCA. The extracted traffic signs under a real-world road environment were recognized successfully with a 92.3% recognition rate using the 40 feature vectors created by the proposed algorithm.

• Journal of Gifted/Talented Education
• /
• v.18 no.3
• /
• pp.371-400
• /
• 2008

The purposes of this study were to develop and validate Korean MHBT for identification of giftedness. MHBT in this study consists of KFT-HB and MHBT-5. MHBT-S was composed of 1) space presentation and thinking ability, space perception,. physics/technic tasks 2) affective domain; creativity, achievement motivation, desire of knowledge, social competence questionnaire 3) performance attitude questionnaire 4) interest questionnaire. The subject were 489 middle school students (1 or 2grade) in the education centers for gifted youth and general classes. Except a few subscales, internal consistent reliability was considered good. Korean MHBT discriminated well gifted students from general students in KFT-HB and some subtests of MHBT-5. As results, Korean MHBT in this study was expected to be a reliable and valid instrument for identification of korean gifted students.