当前位置: 首页 > 文章 > 基于激光视觉的智能识别苹果采摘机器人设计 农机化研究 2016,38 (7) 60-64
Position: Home > Articles > Design for Apple-picking Robot of Intelligent Recognition Based on Laser Vision Journal of Agricultural Mechanization Research 2016,38 (7) 60-64

基于激光视觉的智能识别苹果采摘机器人设计

作  者:
张宾;宿敬肖;张微微;邓明华;汪小志
单  位:
河北工程技术学院;武汉工商学院信息工程学院;武汉理工大学物流工程学院
关键词:
采摘机器人;视觉识别;抗干扰性;自适应性;激光扫描
摘  要:
为了提高苹果采摘视觉识别系统的精度,增强视觉系统的抗干扰能力和自适应能力,设计了一种新的苹果采摘机器人激光视觉识别系统,可以直接获得层次关系的深度图像,实现了果园非结构化环境中果实的识别与定位。为了测试激光识别系统苹果采摘机器人的采摘效果,在果园中对其采摘性能进行了测试:首先采用高清相机完成了对果实图像的采集,通过图像处理准确地实现了苹果的识别,在遮挡率低于50%时其识别率达到了90%以上;然后利用激光测距方法对苹果进行距离测量,成功定位了果实位置,其响应时间仅为3.58s,动作效率快,实现了苹果的高效率、高精度采摘功能。
译  名:
Design for Apple-picking Robot of Intelligent Recognition Based on Laser Vision
作  者:
Zhang Bin;Su Jingxiao;Zhang Weiwei;Deng Minghua;Wang Xiaozhi;Hebei Polytechnic Institute;School of Information Engineering,Wuhan Technology and Business University;School of Logistics Engineering,Wuhan University of Technology;
关键词:
picking robot;;visual recognition;;anti jamming;;adaptive;;laser scanning
摘  要:
In order to improve apple- picking accuracy of visual identification system,enhance anti- interference ability and adaptive ability of visual system,it designs a new apple- picking robot with laser vision recognition system,which can direct access to the depth image of the relationship. It can realize the identification and localization of the unstructured environment of orchard fruit. In order to test laser identification system of apple- picking robot in the orchard of the picking performance were tested. First of all,the high- definition camera completed the acquisition of fruit image,accurate implementation of the recognition of Apple by image processing,the recognition rate in shielding rate of less than 50% to more than 90%. Then the laser ranging method to measure the distance of apple and the successful positioning of the fruit position,its response time is only 3. 58 s and fast operation efficiency,which realize the function of picking apples with high efficiency and accuracy.

相似文章

计量
文章访问数: 17
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊