摘要
在分析肤色的信息特征和手势的运动特性基础上,构建一种基于肤色集、运动集、模糊手势集的手势分割及跟踪算法并采用硬件实现.该算法对颜色空间建立肤色集和对视频流空间建立运动集,通过模糊运算得到模糊手势集,其中加入跟踪算法和背景实时更新,克服了复杂背景和光照因素的干扰.整个算法在现场可编程门阵列(FPGA)上实现,利用流水线和并行处理技术实现了手势的实时分割和跟踪.实验结果表明,该方法有较好的准确性、实时性和鲁棒性,在系统时钟100 MHz时,对640×480分辨率的图像可实现100帧/s的处理速度.
Based on the analysis of the characteristics of skin color information and hand motion proper-ties ,a gesture segmentation and tracking algorithm on account of the color set, the motion set and the vague gesture set is constructed and realized on hardware. With this algorithm, the skin color set is estab-lished in color space while the motion set in video flow space. And through blurring operation joining tracking algorithm and the background real-time updating to overcome the interference of complex back-ground and light factors, the vague gesture set is obtained. The whole algorithm is implemented on the field programmable gate array （FPGA） , realizing gesture real-time segmentation and tracking using stre-amlining and parallel processing technology. The experimental result shows that this method has good accuracy, timeliness and robustness. On a 100 MHz system clock, the speed of processing an image of 640 × 480 resolution can be as fast as 100 frames / second.
出处
《广东工业大学学报》
CAS
2016年第5期54-58,共5页
Journal of Guangdong University of Technology
基金
广东省信息产业发展专项现代信息服务业项目(2150510)
广东省高等学校学科建设专项资金(2014KTSCX058)
关键词
手势分割
FPGA
实时性
模糊处理
gesture segmentation
FPGA （ f ield programmable gate array）
real-time
fuzzy processing
