BROWSE
Detail
Conference Paper
생산 공정 자동화를 위한 물체 인식 기술
Cited - time in 
Share 
Share
- Issue Date
- 2015-06
- Citation
- 대한전자공학회 종합 학술 대회 (하계) 2015, pp.1703-1706
- Publisher
- 대한전자공학회
- Language
- Korean
- Type
- Conference Paper
- Abstract
- In this paper, object recognition using multiple local features and classifier has applied for automating manufacturing system. Object recognition has challenged in real world application field because of illumination effect and complicate background. We have enhanced accuracy of object recognition using multiple local features, which are used to detect object location and to estimate pose. Local feature classifier using NN has used to obtain more reliable object recognition rate. The proposed object recognition method has evaluated in pilot system, in which we have tested object detection rate and pose estimation rate of randomly piled objects. And also, we have tested object recognition accuracy on database captured in real world environment, which has various lighting condition and time passing.
- KSP Keywords
- Enhanced Accuracy, Estimation rate, Illumination effect, Lighting condition, Local features, Manufacturing system, Object recognition, Pose estimation, Real-world applications, Recognition Rate, Recognition method