基于计算机视觉的番茄催熟与正常熟识别  

Recognition of artificial ripening tomato and nature mature tomato based on computer vision

作　　者：赵海波[1,2] 周向红[3] 

机构地区：[1]光电子应用安徽省工程技术研究中心,铜陵244000 [2]铜陵学院电气工程系,铜陵244000 [3]中国电子科技集团公司第43研究所,合肥230088

出　　处：《农业工程学报》2011年第2期355-359,共5页Transactions of the Chinese Society of Agricultural Engineering

基　　金：光电子应用安徽省工程技术研究中心资助项目

摘　　要：国内常有菜农采摘远离成熟期的番茄,采用乙烯利处理进行催熟,为了阻止催熟番茄进入瓜果市场危害食用者的身体健康,给出了催熟番茄识别系统的硬件组成,通过计算机视觉装置获取番茄透射光颜色参数(R、G、B),并将RGB值转换成HIS值,采用遗传算法训练的多层前馈神经网络实现催熟番茄的自动识别。试验结果表明,系统正确识别率为91.7%,为进一步进行番茄催熟与正常熟识别的研究提供参考。Nowadays some vegetable farmers pick unripe tomatoes and treat them with ethylene to quicken ripeness in China.In order to keep artificial ripening tomato which harming consumer’s health from entering into melon and fruit market,the hardware structure of artificial ripening tomato recognition system was given.The colour parameters RGB （red,green,blue） of transmitted light of tomatoes were obtained through computer vision device,and the RGB values were converted into HIS （hue,intensity,saturation） values.The multilayer feedforward neural networks with genetic algorithm training realized the automated recognition of artificial ripening tomato.The results of test showed that accurate recognition rate of the system was 91.7%,and the method can provide references for further research on recognition of artificial ripening tomato and nature mature tomato.

关 键 词：计算机视觉 神经网络 遗传算法 催熟番茄 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP274[自动化与计算机技术—计算机科学与技术]