Dynamic Background Modeling System Based on Spatial Extended Local Binary Pattern
碩士 === 國立高雄應用科技大學 === 電子工程系 === 100 === In this thesis, a background modeling system using the method of local binary pattern (LBP) by the spatial extension to process feature extraction in dynamic environment is proposed. This modeling for each pixel of the video builds the feature pattern, which u...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
101
|
Online Access: | http://ndltd.ncl.edu.tw/handle/73448897520653883440 |
Summary: | 碩士 === 國立高雄應用科技大學 === 電子工程系 === 100 === In this thesis, a background modeling system using the method of local binary pattern (LBP) by the spatial extension to process feature extraction in dynamic environment is proposed. This modeling for each pixel of the video builds the feature pattern, which uses a hexagon and rhombus arrangement structure to improve the detection rate for a motion object. Experimental results show that the proposed method can enhance about the detection rate of 10% as compared with the SCS-LBP algorithm. Therefore, the proposed system is very suitable for applications in the adaptive background environment and motion object detection.
|
---|