针织手套轮廓分割图像质量评价研究
Image Quality Evaluation of Knitted Glove Contour Segmentation
潘利鑫 1茅木泉 2袁嫣红1
作者信息
- 1. 浙江理工大学机械与 自动控制学院,浙江杭州 310018
- 2. 杭州高腾机电科技有限公司,浙江杭州 310018
- 折叠
摘要
为实现更高效、更准确的纯色手套轮廓分割,文中构造一种以轮廓分割为目标,基于图像灰度直方图特征参数的针织手套图像质量评价指标.通过拍摄不同质量等级手套图像,分析灰度直方图形态与轮廓可分割难易程度关系,总结由灰度直方图中波峰值、波谷值、峰距和类内方差构成的图像质量评价模型,此模型可定量计算图像指标,该指标是自动拍摄调参的基础.结果表明,根据模型求解的评价指标值与人眼主观评价等级有较好一致性,且评价指标值与手套本身颜色无关,通用性较好.因此,完成评价指标可改变依靠人眼观察的手动调参拍摄方法,为在线检测需要的相机智能调参拍摄提供依据和基础.
Abstract
In order to achieve more efficient and accurate contour segmentation of solid colored gloves,this paper constructs a knitted glove image quality evaluation index based on image grayscale histogram feature parameters with contour segmentation as the objective.By shooting glove images of different quality levels,the relationship between the shape of grayscale histograms and the difficulty of segmenting contours is analyzed.An image quality evaluation model composed of peak value,valley value,peak distance,and intra class variance in grayscale histograms is summarized.This model can quantitatively calculate image indicators,which are the basis for automatic shooting and parameter adjustment.The experimental results show that the evaluation index values solved by the model have good consistency with the subjective evaluation level of the human eye,and the evaluation index values are independent of the color of the gloves themselves,indicating good universality.Therefore,the completed evaluation indicators can change the manual parameter adjustment shooting method that relies on human eye observation,providing an adjustment basis and foundation for intelligent camera parameter adjustment shooting required for online detection.
关键词
直方图/图像质量评价/针织手套/轮廓分割/双峰性Key words
Histogram/Image Evaluation Index/Knitted Gloves/Image Segmentation/Bimodality引用本文复制引用
出版年
2024
NETL
NSTL
万方数据