An Object Segmentation System Design using Group-based Color Algorithm

• Main Authors: Hsieh, Ping-Chuan, 謝秉荃
• Other Authors: 林正基
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/67356739774419598014

碩士 === 東海大學 === 資訊工程學系 === 101 === Background segmentation is a very important technology to get information about moving objects. In order to get the right background information, you must create a background model to segment the foreground and background. The traditional segmentation algorithm have a good effect on dealing with complex background information, but it also brings more processing time and larger memory space. This paper presents a new algorithm to detect moving objects in the foreground, the use of block-based processing of pixels in the image, and then one by one grouping whose pixels are subsequently acquires the background color information. The block-based and pixel grouping process can significantly reduce the time requirement to create a background model and accelerate get background information. Through grouping color after background information you can extract information about moving objects, and then provided to the user-related alerts, to maintain social security purposes.