水果智能采摘质量识别技术  

Technology of Fruit Picking Quality Intelligent Identification

作　　者：王坤[1,2] 

机构地区：[1]泰州市农业物联网工程技术中心,江苏泰州225300 [2]江苏农牧科技职业学院,江苏泰州225300

出　　处：《科技通报》2014年第10期139-141,共3页Bulletin of Science and Technology

基　　金：江苏省农产品质量安全追溯管理体系研究与试点项目(2130109)

摘　　要：研究一种采用机器视/嗅觉和联网技术的水果智能采摘识别系统,传统的水果采摘和质量鉴别采用人工识别的方法,劳动量大且标准不统一导致水果采摘不适时而质量不好。提出一种基于气味传感识别和颜色纹理识别的双重识别技术对水果的质量进行追踪识别,使用传感器进行气味感知,获得较高精度的水果纹理像素级视差,避免了因为果叶的遮挡导致的识别不准问题,实现基于气味和视觉的高精度水果质量分析。基于AD9225高速DSP芯片进行硬件设计,搭建物联网平台,设计出水果智能采摘质量识别检测仪。系统测试实验结果表明,采用该水果智能采摘仪,能有效跟踪树上水果的成熟度和质量,实现准确分类识别,并根据水果质量进行选择性实时采摘。The intelligent fruit picking and recognition system based on the computer vision and olfaction and in Internet of Things technology, the traditional method is manual and costs amount of labor, the quality is not good. Double recognition indexes are used with the odor sensing and the color texture recognition. Sensors are used for odor perception, and multiple window spectrum estimation method is used for feature extraction, and the singular value decomposition solution and phase winding technology of fruit color and texture features are extracted, finally get the texture the pixel level machine vision characteristic through least square fitting. The higher accuracy of fruit texture is obtained. The hardware design are ob-tained based on AD9225 DSP chip, network platform is established. The system test results show that it can effectively track the tree fruit maturity and quality, the real-time picking of fruit are completed.

关 键 词：水果采摘 物联网 视觉 嗅觉 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]