作物数字图像分割误差的影响因素研究

背景分割是机器视觉图像处理的重要环节，在室外自然光照条件下，影响作物图像分割质量的因素有很多，为了解多种影响因素对作物图像分割质量产生的影响及其影响规律，给机器视觉实时识别作物提供参考，该文通过正交试验的方法，以作物图像的分割误差为指标，并借助计算机统计分析软件，研究了光照强度、太阳高度角、土壤干湿程度和拍摄角度四种因素对青椒图像分割质量的影响，获得了影响青椒图像分割质量的主次因素及影响规律。
关键词：图像分割；误差；正交试验；作物
Studies on Factors Influencing Digital Image Segmentation of Crop Mao Hanping，Qian Dan
(Institute of Agriculture Engineering, Jiangsu University, Jiangsu Province Key Laboratory of Modern Agriculture Equipment and Technology. Zhenjiang 212013, China) Abstract: Image segmentation is important in digital image processing, there are multi-factors influencing image segmentation of crops in outdoor entironments，in order to find out main factors and their rules, orthogonal experiments are performed and four factors are studied. Through computer statistic analysis software, the main and subordinate factors and their influence rules on image segmentation of green pepper are obtained.
Key words: image segmentation; error; orthogonal experiment; crop