1-404 Study on a Landmark Detection and Recognition Technique
スポンサーリンク
概要
- 論文の詳細を見る
For a mobile robot vision system, the colour based detection and recognition technique is more efficient than others. It will be more practical if its stability can be improved under different lighting conditions. In our case, the colour models of landmarks were built, and the colour models was used to separate landmark colour regions from non-landmark colour regions in images taken by a CCD video camera. In order to recognize landmarks, geometric features and colours of landmarks were used. After the landmark's region was detected, the colour of landmark is recognised by the models of landmark colours. And then the boundary of landmark was detected, and recognizes the shape of landmark by its curvature characteristic of boundary. The reliability of this technique was tested with practical different lighting conditions.
- 一般社団法人日本機械学会の論文
- 2002-11-14
著者
-
Sugisaka Masanori
Department Of Electrical And Electronics Engineering Oita University
-
Sugisaka Masanori
Department Of Electrical & Electronic Engineering Oita University
-
Chen Renping
Department of Electrical and Electronics Engineering Oita University
関連論文
- A Selective Vision and Landmark based Approach to Improve the Efficiency of Position Probability Grid Localization
- Fast and Robust Face Detection Using Genetic Search
- Development of adaptive self-driving control strategies for an intelligent mobile vehicle
- 1-404 Study on a Landmark Detection and Recognition Technique
- Improvement on the Digit Recognition Technique of Backpropagation Neural Networks
- Letter recognition and obstacle avoidance by ART neural network
- An Effective Search Method for Neural Network Based Face Detection Using Particle Swarm Optimization(Artificial Intelligence and Cognitive Science)
- Adaptive Fuzzy Control Based on Neural Network for a Mobile Vehicle
- 1-203 A Control of Two DC Stepping Motor Using Control Vector and Control Trajectory for Electric Car