摘要
在线检测是实现自动生产的重要环节,计算机视觉图像信息处理技术已经广泛应用于农产品品质检测中。表面缺陷检测一直是困扰苹果自动化分级的难题,本文采用特征提取,即提取出缺陷部位的轮廓,再进行填充求其面积的方法确定缺陷面积大小,然后用缺陷面积结合最大横径、颜色色度来判定等级,实现机器化苹果分拣。经实验证明本方法准确率达94.2%,且具有较好的稳定性,为农业自动化提供了一种可靠的苹果在线检测的方法。
On-line detection is a critical part of automatic production, and visual image information processing technique has been gradually applied in the field of fruit quality detection. Surface defect detection is a difficult problem for fruit automatic grading. The feature extraction method was used which first gets the extraction of defective parts and fills the area to determine the defect size, then combines maximum transverse diameter and color tone with defect size to realize apple sorting mechanization. This method has 94.2% precision rate and better stability, could provide a reliable method to accelerate fruit on-line detection for agriculture automation.
出处
《中国农机化学报》
北大核心
2014年第5期169-172,共4页
Journal of Chinese Agricultural Mechanization
关键词
品质检测
机器视觉
图像处理
苹果分级
quality detection
computer vision
image process
apple classification
