针对网域内简单背景下目标感知识别处理速度慢、准确率低的问题,提出了基于最优算法组合的目标检测方法；该方法将多种图像处理方法融入目标识别算法中,在采用不同算法对采集到的图像进行预处理、分割和特征提取效果比较的基础上,确定出最优算法组合并设计实验验证其可靠性和鲁棒性;试验结果表明该方法能够快速有效的判别行人目标,识别率达到96.67％,解决了网域内目标检测问题；与其他单一算法相比具有处理速度更快、判别更加有效,实时性高等特点,明显优于一般算法.%In view of the problems in the simple background of target perceptual recognition is slower and accuracy is lower, an target detection of optimal algorithm combination is proposed. This method integrates a variety of image processing methods with the target recognition algorithm. To determine the optimal algorithm combinations, different algorithms as preprocessing, segmentation and feature extraction on the system with collected images are compared. Reliability and robustness are being designed to verify the optimal algorithm combinations. Experimental results show that the method can distinguish pedestrians quickly with the recognition rate at 96. 67% and effectively solve the target detection within the network. Also show that compared with other single algorithm, the optimal algorithm combination is faster in speed, more effective in identification and higher in real-time. Significantly better than the general algorithm
展开▼
