图像智能分割下的果实成熟度评估计算与研究  

CALCULATION AND RESEARCH ON HARVEST RIPENESS EVALUATION WITH INTELLIGENT IMAGE SEGMENTATION

作　　者：胡玲艳[1] 李俐 许巍 李国强 宋旦旨 刘艳[1] 裴悦琨 汪祖民[1] Hu Lingyan;Li Li;Xu Wei;Li Guoqiang;Song Danzhi;Liu Yan;Pei Yuekun;Wang Zumin(School of Information Engineering,Dalian University,Dalian 116000,Liaoning,China;Dalian Modern Agricultural Production Development Service Center,Dalian 116000,Liaoning,China;Dalian University Archive,Dalian 116000,Liaoning,China)

机构地区：[1]大连大学信息工程学院,辽宁大连116000 [2]大连市现代农业生产发展服务中心,辽宁大连116000 [3]大连大学档案馆,辽宁大连116000

出　　处：《计算机应用与软件》2025年第3期149-155,195,共8页Computer Applications and Software

基　　金：国家自然科学基金青年科学基金项目(61601076);大连市科技创新基金项目(2020JJ26SN058,2021JJ13SN78)。

摘　　要：基于果实大小和颜色指标评估果实成熟度,为樱桃储运中的最佳采收时间提供数值化判断依据。利用改进型智能剪刀算法,从光照背景中分割不同时期果实特征区域;计算特征区域的图像投影面积,图像在HSV颜色空间模型下H分量的方差和均值;根据获取的果实面积和颜色指标数据,采用加权评估法计算不同时期果实成熟度。果实发育过程各指标动态变化,当果实面积达到190000像素,果皮颜色为紫红色时,成熟度达最大值。实时获取樱桃果实成熟度双重评估的指标,有效实现了对果实成熟度的数字化评估。Assessment of fruit ripeness based on fruit size and color indicators provides a numerical basis for optimal harvesting time in cherry storage and transportation.The improved intelligent scissors algorithm was used to segment the fruit feature regions from the light background at different periods.The image projection area of the feature area,the variance and mean of the H component of the image under the HSV color space model were calculated.Based on the obtained fruit area and color index data,the weighted assessment method was used to calculate the fruit maturity at different periods.Each index changed dynamically during fruit development.The ripeness reached the maximum when the fruit area reached 190000 pixels and the skin color was purple-red.Real-time access to indicators for dual assessment of cherry fruit ripeness effectively enables digital assessment of fruit ripeness.

关 键 词：计算机视觉 智慧农业 樱桃 改进型智能剪刀算法 果实图像投影面积 颜色提取 采收成熟度 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]