摘要
通过机器视觉获取脐橙截面积来估测脐橙体积,首先进行图像分割、腐蚀,然后从腐蚀图像中脐橙的像素点得出脐橙的截面积,接着计算出脐橙半径,通过最小二乘法估测脐橙体积。为提高精度,采取2个不同拍摄角度拍摄2幅脐橙图像,然后以2幅脐橙图像计算出脐橙截面积的平均值作为脐橙截面积。结果表明,估测误差将近9.36%,说明基于机器视觉将脐橙作为球体进行估测是低精度的。
The navel orange volume is estimated from cross selection area acquired through machine vision.The navel orange image is first segmented and eroded,and the cross selection area is obtained via the navel orange pixels＇ number in erosion image.Then the radius of navel orange is computed out,and the volume is estimated with least square method.To improve the estimation effect,two cross selection areas are obtained from two images from different shooting angles and the cross selection area is the average of them.The experiment result shows that the average error is about 9.36%.It shows that the navel orange volume estimation as a ball based on machine vision lacks precision.
出处
《安徽农业科学》
CAS
北大核心
2011年第32期20237-20238,共2页
Journal of Anhui Agricultural Sciences
关键词
脐橙
体积
球体
Navel orange
Volume
Ball
