摘要
为实现在工业测量中对工件边缘的快速准确定位,以满足在线测量系统的实时性的要求,利用脉冲耦合神经网络(PCNN,Pulse Coupled Neural Network),并结合前三阶灰度矩实现了对工件边缘的精确定位。该方法首先利用脉冲耦合神经网络对待测量工件进行初步定位,然后利用初步定位的结果设置感兴趣的区域,再利用前三阶灰度矩在原始灰度图像上的感兴趣区域内进行边缘的亚像素细分。实验结果表明该方法抗噪声能力强,边缘定位准确,同时能够提高边缘的检测速度。
To achieve a quick locating for the edge of workpiece in the industrial online measurement system to meet the real-time requirement,an accurate workpiece edge locating method is implemented combines pulse coupled neural network with first three sample gray moments.At first,an initial position is located with pulse coupled neural network,and according to the results of first step the region of interest is set,then the first three sample gray moments is employed for the sub-pixel edge segmentation from the original gray-scale image in the region of interest.The simulation results show that the proposed approach has abilities of strong anti-noise,quick and accurate edge locating and improves the speed of edge detection.
出处
《计算机技术与发展》
2010年第6期221-224,共4页
Computer Technology and Development
基金
江苏省高校重点实验室开放基金(KXJ07128)
关键词
工件
脉冲耦合神经网络
灰度矩
边缘检测
亚像素
workpiece
pulse coupled neural network
gray moment
edge detection
subpixel