果蔬智能分拣系统中的目标提取算法的实现  

Implementation of target extraction algorithm in fruit and vegetable intelligent sorting system

作　　者：蔡鑫垚 何开纪 张晨阳 李高平[1] CAI Xin-yao;HE Kai-ji;ZHANG Chen-yang;LI Gao-ping(School of Computer Science & Technology,Southwest Minzu University,Chengdu 610041 ,P.R.C.)

机构地区：[1]西南民族大学计算机科学与技术学院,四川成都610041

出　　处：《西南民族大学学报（自然科学版）》2019年第2期178-184,共7页Journal of Southwest Minzu University(Natural Science Edition)

基　　金：中央高校基本科研业务费专项资金项目(2018YXXS16)

摘　　要：随着农产品加工自动化程度越来越高,采用工业机器人代替人工分拣水果、蔬菜等农产品是适应现代农业发展所必需的.在蔬菜水果分拣过程中的前景目标提取质量决定了分级的准确性,基于混合高斯背景模型和仿射变换模型建立相应算法,对蔬菜水果分拣过程中所拍摄到的前景目标进行提取.以水果猕猴桃为例的实验结果表明,该方法所提取到的前景目标质量能够满足蔬果分拣需求,与其它方法相比也具有一定优势.With the increasing degree of automation of agricultural product processing,the use of industrial robots instead of manually sorting agricultural products such as fruits and vegetables is necessary for the development of modern agriculture.The quality of foreground target extraction in the process of sorting vegetables and fruits determines the accuracy of classification.Based on the Gaussian mixture background model and the affine transformation model,the corresponding algorithm is established to extract the foreground targets captured during the vegetable and fruit sorting process.The experimental results of kiwifruit show that the quality of the foreground target extracted by this method can meet the demand of vegetable and fruit sorting,and it has certain advantages compared with other methods.

关 键 词：水果蔬菜 分类拣选 图像处理 仿射变换模型 混合高斯背景模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]