低位机处理图像识别技术的设计与实现  

Design and Implementation of a Low Level Single-chip Processing Image Identifying Technology

作　　者：江爱荣[1] 王超荣[1] 曾晰[1] 王渊平[1] 

机构地区：[1]浙江工业大学特种装备制造与先进加工技术教育部重点实验室,浙江杭州310014

出　　处：《轻工机械》2011年第2期95-98,共4页Light Industry Machinery

摘　　要：由于低位单片机在处理图像识别技术中存在数据存储量不够、实时性差等缺点,许多工业领域不得不采用价格更高的高位单片机(如arm9、DSP等)来实现图像处理。为了使图像处理技术在一些高成本的工业领域中既能满足设计要求又具有较高的经济性,文章提出了一种由8bit单片机atmega32作为主芯片采集OV7620图像信号的控制方法,以应用于电梯门控系统来说明,提出了4种方案解决了8bit单片机在图像识别中的弊端,并且根据像素点的特性采用了半邻域平均算法结合背景差分法来识别物体。实验结果表明该系统的实时性、适应性都基本满足要求,具有较大的实用价值,能够带来可观的经济效益。In the image identifying technology,the low level single-chip exists the disadvantages of limited data-storage and bad real-time.Lots of industrial fields have to choose higher-priced single-chip（such as arm9,DSP etc.） to achieve the image processing technology.In some high-cost industrial technology,in order to make image processing technology to meet the design requirements with economic benefit,the paper presented a new control approach with 8 bit single-chip wit-atmega32-main core chip collect processing OV7620 image signal,and took elevator door control system to illustrate with four plans to overcome the defects in the application of 8 bit single-chip identifying technology.According to the feature of image element,the paper used half-neighbourhood averaging algorism combined with background difference method for animated objects identifying.Experimental results show that the system real-time performance,adaptability etc.can meet all basic requirements.It will bring considerable economic benefit and a greater practical value.[Ch,6 fig.10 ref.]

关 键 词：图像识别技术 半邻域平均算法 背景差分法 atmega32单片机 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]