中文摘要：

目前国内普遍采用的整纬装置为光电整纬,依靠采集透射过织物的光强度来检测纬斜,但这种方式对织物的厚度及纹理程度具有一定制约性;针对这个问题设计了一种基于DM642的摄像式的纬斜检测系统,充分利用DM642的高速运算能力,对织物的纹理进行图像处理,来实现纬斜角度实时的检测;该系统以TMS320DM642处理器为核心,包括视频采集、视频显示、视频存储、视频处理、外部通信5个模块,并在该硬件平台基础上实现了基于投影映射的纬斜检测算法;映射算法的阈值采用20,根据统计的大于阈值的数目来得到纬线偏移的角度;应用结果表明本系统具有良好的实时性并且精度可达到95%,可用于工业现场的实时纬斜检测。

英文摘要：

Currently the traditional straightening machines which widely used, relies on the transmission light intensity of the fabric to detect the skew, with certain reslrictionson the thickness of the fabric and texture degree. We designed a weft- straightening system based on camera detection, the design full use of the DM642 high-speed computing power, and image processing on the texture of the fabric to a cbieve real time detection of skew perspective. The system is consist of video capture, video display, video storage, video processing, ex ternal communication five modules and on the hardware platform we realized the skew detection algorithm based on projection mapping. The tbreshold value was chosen 20, According to the value which is greater than the threshold we got the angular deflection of weft. The experi ments showed that this system has good real time and higher accuracy, it can be used in the industrial field real time skew detection.