基于梯度极值边缘跟踪与嵌入式系统的实时绞距检测算法

Real-time twisted ranging algorithm based on gradient extreme value edge tracking and embedded system

针对双绞线生产流水线上双绞线图像采集时产生的模糊与外部光照条件不均等问题使得图像测距误差较大的问题，基于双绞线模糊图像的灰度变化特征提出了一种快速边缘识别与绞距测量算法。快速绞距测量算法首先寻找绞线图像边缘跟踪起点像素，然后跟踪梯度极值最大的点至图像最右侧得到双绞线边缘近似图像，同时根据得到的近似边缘点计算双绞线的绞距。实验证明，该方法能够在双绞线图像边缘识别的同时计算出绞距，受外部光照等条件的影响较小。并且在嵌入式IMX6Q平台上实现的绞距检测系统能实现流水线上绞距的实时检测。

Aim to the image acquisition in twisted pair production line is too blurry and the external light is not uni- form which make the error of image ranging. Therefore, this paper proposes a quick edge coghitive and twisted ranging al- gorihtm based on gray scale variation. The algorihtm firstly find the starting pixel. Then track the pixel with gradient maxi- mum point to the right of the image to get the edge approximate image. At the same time, calculate the twisted - pair range based on the edge pixel. Experimental results prove that the algorithm can calculate the twisted - pair range and i- dentify the edge points at the same time. The algorithm is less affected by conditions such as external light. And the algo- rithm tuning on the IMX6Q platform shows that the system can achieve the real -time twisted -pair measuring.