实现识别行间苗草的机器视觉系统设计  

Design of Machine Vision System for Real-time Inline Seedling Identification

作　　者：李谦[1] 蔡晓华 

机构地区：[1]黑龙江八一农垦大学信息技术学院,大庆163319 [2]黑龙江农业机械工程科学研究院

出　　处：《黑龙江八一农垦大学学报》2014年第1期67-70,共4页journal of heilongjiang bayi agricultural university

基　　金：国家科技支撑计划资助项目(2011BAD20B07)

摘　　要：在耕作试验室的试验环境之下,开发并测试了能够识别行间苗草作物的机器视觉系统,硬件系统主要由速度可控制的土槽试验车装备、固定在可升降平台上可实时采集图像的imagesoure工业摄像头和作为控制台的工控计算机组成;机器视觉系统根据植物和背景的颜色特征二值化图像与田间作物的位置特征识别苗草。结果表明,采集并处理一副大小为640×480像素的彩色图像的平均时间为291 ms,在行驶速度为1.8 km·h-1条件下系统正确识别率达到了95%。Under the farming of the lab testing environment,the machine vision system could identify line thinning grass crop was developed and tested,whose hardware system was mainly composed of the controlled speed testing lorry,imagesoure industrial camera collecting real-time images was fixed in liftable platform,industrial control computer as the console. The machine vision system identified seedlings by binarization image of the plants and background color features and the location characteristics of the seedlings. The results showed that the average time of processing 640x480 pixel color images was 291 ms,and the system recognition rate could reached 95%at the speed of 1.8 km·h-1.

关 键 词：机器视觉 图像处理 实时识别 行间苗草 

分 类 号：S482[农业科学—农药学]