摘要
目前国内外普遍采用的鲜切花品质检测方法是感官评审法,此方法易受个人主观因素和外界因素的影响。该研究利用计算机视觉技术来代替人的感官对玫瑰切花品质等级进行分类研究,试验中对玫瑰切花图像进行图像分割、图像去噪等图像处理后选取其7个形状特征参数作为玫瑰切花的品质评价指标,并借助BP神经网络建立玫瑰切花品质分级模型,分级正确率达到94%以上。试验表明,基于计算机视觉的玫瑰切花品质分级是可行的,并且具有较高的分级正确率。
At present, the prevailing method to determinate flower quality is through organoleptie way at home and abroad, this method is affected by external and personal subjective factors easily. This article uses the computer vision technology to replace human's sense organ to study on classification of cut roses quality grade, in the experiment, after preprocess the cut roses image with image segmentation, image denoising, then abstract 7 shape parameters as cut roses quality evaluation index, and through the establishment of rose cut flower quality grading model based on BP neural network, the classification accuracy rate was above 94%. Experiments show that classifying cut roses based on com- puter vision is feasible and has highly accuracy.
出处
《安徽农业科学》
CAS
2015年第5期322-326,共5页
Journal of Anhui Agricultural Sciences