基于图像识别技术的机械手臂过角度判断.pdf

第33卷第10期
计算机仿真
2016年10月
文章编号:1006-9348(2016)10-0391-04
基于图像识别技术的机械手臂过角度判断
李晓峰
(南京航空航天大学计算机科学与技术学院,江苏南京210068)
摘要:针对传统方法容易受到噪声及外界条件的干扰,判断时误差较大、无法有效实现机械手臂过角度的判断的问题。提出
采用机械手臂过角度判断方法,通过最大类间方差法计算机械手臂图像的自适应阈值,完成图像分割。利用中值滤波法对
图像进行滤波,完成对图像的膨胀处理。将机械手臂图像边缘点梯度向量作为匹配信息对相似度量进行计算。依据曲面拟
合原理对边缘梯度方向进行计算,对机械手臂图像进行求导,对图像的梯度进行计算,判断图像的值与极大值抑制获取像
素级的边缘位置,将像素级边缘点某范围内的梯度值作为拟合信息,完成梯度曲面拟合,通过最近邻法实现机械手臂过角度
的判断。仿真结果表明,改进方法进行机械手臂过角度判断时的平均实测角度更接近实际值;平均误差和所需时间都有了
较大的改善,具有较好的实用性。
关键词:图像识别;机械手臂;角度;判断
中图分类号:TP273文献标识码:A
Based on Image Recognition Technology of the
Mechanical Arm Angle Judgment
L Xiao-feng
( School of Computer Science and Technology Nanjing University of Aeronautics& Astronautics
Jiangsu Province, Nanjing Jiangsu 210068, China)
ABSTRACT: The traditional method is easily affected by the noise and interference of external conditions, so that the
judgment error is big, and it is unable to effectively realize the over-angle judgment of mechanical arm. An over
gle judgment method of mechanical arm is proposed. By using maximum between-cluster variance method, the a
daptive threshold of the mechanical arm image is calculated, and the image segmentation is accomplished. The image
is filtered by median filtering to complete the expansion processing of image. The edge point gradient vector of me
chanical arm image is as matching information to calculate similarity measurement. Based on the principle of surface
fit ing the edge gradient icon is calculaed. The mechanical arm image is made derivation, and the image gradi
ent is calculated, the image threshold and maximum suppression are determined, and the edge position of the pixel
level is obtained. The gradient values of pixel level edge points in a range are as fitting information, and the gradient
surface fitting is completed. Through the nearest neighbor method, the over-angle jud ment of mechanical arm is
chieved. The simulation results show at the average measured angle using the improved method used for the over
angle judgment of mechanical arm is more closed to the actual value, the average error and the required time have
been greatly improved, and it has good practicability
KEYWORDS: Image recognition; Mechanical arm; Over-angle; Judgement
角度参数进行髙精度快速自动判断才能保证机械运行的准
1引言
确性与快速性-】。因此,研究机械手臂过角度判断方法具
近年来,随着工业的逐渐发展,各个领域对机械手臂角有重要意义,成为相关学者研究的重点课题,受到越来越广
度精度的要求越来越高,而只有利用视觉系统对机械手臂过泛的关法(、-6°过角度进行判断的过程中,容易受到噪声
在对机械手
收稿日期:2015-07-21修回日期:2016-01-28
及外界条件的干扰,导致传统机械手臂过角度判断方法,由
万方数据