http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Vision-based Railway Inspection System using Multiple Object Detection and Image Registration
Jinbeum Jang,Heegwang Kim,Minwoo Shin,Jonggook Park,Joungyeon Kim,Joonki Paik 대한전자공학회 2018 IEIE Transactions on Smart Processing & Computing Vol.7 No.6
Image processing and computer vision techniques have been utilized for safety and maintenance in the railway field. Although a lot of research has been proposed to automatically inspect a facility, most diagnosis for facility maintenance is still dependent on a manager’s subjective judgment. This paper presents a novel railway-inspection system using object detection and image subtraction based on registration. For accurate deformation and defect inspection, the proposed system compares a pair of two high-resolution images acquired by a laser scan camera equipped on a railway vehicle. The proposed system consists of three parts: i) object detection using classifiers learned by random forest, ii) facility position alignment using phase correlation matching, and iii) deformation and defect detection using image registration and subtraction. The proposed inspection system performs automatic inspections by detecting facilities and any deformed regions. Therefore, the proposed system can provide improvement of a maintenance system at a cost reduction.
Person Re-identification using Sparse Representation with a Saliency-weighted Dictionary
Miri Kim,Jinbeum Jang,Joonki Paik 대한전자공학회 2017 IEIE Transactions on Smart Processing & Computing Vol.6 No.4
Intelligent video surveillance systems have been developed to monitor global areas and find specific target objects using a large-scale database. However, person re-identification presents some challenges, such as pose change and occlusions. To solve the problems, this paper presents an improved person re-identification method using sparse representation and saliency-based dictionary construction. The proposed method consists of three parts: i) feature description based on salient colors and textures for dictionary elements, ii) orthogonal atom selection using cosine similarity to deal with pose and viewpoint change, and iii) measurement of reconstruction error to rank the gallery corresponding a probe object. The proposed method provides good performance, since robust descriptors used as a dictionary atom are generated by weighting some salient features, and dictionary atoms are selected by reducing excessive redundancy causing low accuracy. Therefore, the proposed method can be applied in a large scale–database surveillance system to search for a specific object.