【作者】 曹小秋； 王晓舟； 刘娟； 高小勇；

【Author】 Cao Xiaoqiu Wang Xiaozhou Liu Juan Gao Xiaoyong (College of Electronic and Control Engineering of Beijing University of Technology Beijing 100124)

【机构】 北京工业大学电子信息与控制工程学院；

【摘要】 匹配算法在图像处理中多应用于实时性要求较高的系统,灰度图像由于数据量大,严重的影响了匹配的实时性。本文实现了一种利用小波的多尺度分解,对图像进行金字塔式的分层匹配的方法,在保证精度的前提下大大减小了运算量。并充分发挥ADSP TS201S的优良性能,在DSP平台上进行优化实现。经过结构和算法的优化后,算法的匹配速度得到了显著的提高。更多还原

【Abstract】 Image matching is widely used in real-time image processing system;The gray value matching of images;because of the large calculating demand;badly weaken the processing performance;This article introduces a way for image matching based on wavelet multi-resolution pyramidal structure which can reduce the large computations and have a good precision;After having optimized the structure;algorithm and having taken full advantage of the ADSP TS201S’s features;the experimental results show that this method greatly improved processing speed;performance,and accuracy.更多还原