中文摘要：

滴灌带是常用的农业节水灌溉产品之一，它利用镶嵌在管道内壁上的滴头将水送到农作物根部。在生产线上采用计算机视觉系统可以提高滴灌带在线检测质量和效率，其中为降低滴头漏打／打偏率在图像处理过程中对滴头的准确定位是比较关键的一步。利用边缘检测和模板匹配相结合的算法定位滴头。算法首先通过自适应Canny边缘检测算子进行粗定位，然后采用模板匹配的方法完成精定位。大量现场实验结果表明该算法能够快速有效地实现滴灌带图像中滴头的实时准确定位。

英文摘要：

The dripping irrigation pipe is a common agricultural irrigation water-saving product. It transports water to the plant root by tile dripping heads that are imbedded in the inner walls of the pipes. A production line that uses computer vision system can increase online inspection quality and efficiency of dripping irrigation pipes. More specifically, in order to decrease the escape or miss rate of drippers, during the image process, it is an essential step to precisely locate the drippers. The paper locates drippers by an algorithm that integrates edge detection and template matching. The algorithm first of all utilizes adaptive Canny edge detection operator for coarse location. Next it utilizes template matching method to implement precise location. A lot of vivid experiment results illustrate that the algorithm can quickly and effectively realize the real-time accurate location of drippers in dripping irrigation images.