CAJ | 학술논문

文章针对传统码垛机器人作业时存在的不足，从视觉的角度研究了相关技术难点，为实现生产自动化提供了解决方案。文章完成了基于机器视觉的码垛机器人系统平台的搭建，首先通过摄像机对传送带上进入工作区的工件进行图像采集，然后对图像处理分析，接着用不变矩对工件进行粗略快速识别，之后用SIFT算法对工件进行局部特征匹配，最后通过中心矩计算出工件的位置，引导机器人对工件进行抓取，机器人对抓取后的工件进行分类码垛。同时，文中还提出了多目标分块处理算法、SIFT和HU特征融合的单目视觉识别算法。实验结果表明，该码垛系统可以有效解决工件的分类码垛的问题。
문장침대전통마타궤기인작업시존재적불족，종시각적각도연구료상관기술난점，위실현생산자동화제공료해결방안。문장완성료기우궤기시각적마타궤기인계통평태적탑건，수선통과섭상궤대전송대상진입공작구적공건진행도상채집，연후대도상처리분석，접착용불변구대공건진행조략쾌속식별，지후용SIFT산법대공건진행국부특정필배，최후통과중심구계산출공건적위치，인도궤기인대공건진행조취，궤기인대조취후적공건진행분류마타。동시，문중환제출료다목표분괴처리산법、SIFT화HU특정융합적단목시각식별산법。실험결과표명，해마타계통가이유효해결공건적분류마타적문제。
For the existing palletizing workpieces problems in the industrial production lines, this paper has discussed the technical difficulties in visual perspective and providing a solution to realize automation. First-ly, this paper built the grasping intelligently platform on the industrial production line. Firstly, the experi-mental system using the industrial camera which on the production line acquired pictures of workpieces that placed on the platform, next preprocessed the digital image and distinguished the models quickly by moment invariants, then recognized the workpieces by sift, later calculated the geometric center coordinates of the model and guided the manipulator to sort the models,finally the robot placed the workpieces regularly. Ex-perimental shows that the palletizing system can effectively solve the placements of workpieces.