猕猴桃外观尺寸在线检测分级系统设计与试验

目前市场上存在的猕猴桃分级机械主要是针对质量特征进行分级的大型机械,而消费者在购买猕猴桃时往往更注重它的外观品质,且大型设备难以在以农户小规模营销为主的市场环境下得到普及。为了实现猕猴桃外观品质的自动分选,适应广大猕猴桃种植农户的需求,设计了一种基于机器视觉的小型移动式猕猴桃外观尺寸在线检测与分级系统。该系统主要由输送机构、检测机构、分级执行机构和控制系统组成。输送机构采用倾斜式输送带方案,结构简单,便于实现猕猴桃的输送和分级;检测机构采用图像处理的方法得出猕猴桃的大小等级信息;分级执行机构借助猕猴桃的重力与旋转磁铁的开合实现猕猴桃的分离。对样机进行了试制和验证试验,结果表明:该系统的平均分级成功率为96.3%,单个猕猴桃分级时间约为2.5s。该猕猴桃检测分级系统的设计为今后完成多特征指标的融合分级提供了基础和依据。

Design and Experiment on Automatic Grading System of Appearance Size for Kiwi

Existing kiwi grading machine mainly classified fruits based on the weight, while consumers focus more on the appearance quality of fruit and the popularity of large equipment is quite difficult under this small-scale farmers oriented market environment. In order to achieve automatic sorting of kiwi and meet the needs of kiwi farmers, we designed a small online detection and classification systems of kiwi based on machine vision. This system is mainly composed of transport institutions, testing institutions, grading actuator and control system. The transport mechanism uses inclined conveyor which is with a simple structure and easy to deliver and grade kiwi. Detecting mechanism using image process-ing methods derives the kiwis’ size classify information. The classifying actuator achieves separation of kiwi through the kiwis’ gravity and the opening and closing of the rotating magnet. Finally, a prototype was made and the verification test results showed that, the system's average grading accuracy could reach to 96. 3%, and a single kiwi classification time is about 2. 5s. The design of this detection and grading system provide the basis for the completion of multi-feature fusion index rating in the future.