CAJ | 학술논문

主要对创新实验室二维视觉跟踪转台的算法进行改进。传统的算法只能实现在白色背景下对黑色目标的识别与跟踪，为实现对不同目标复杂背景的跟踪，针对跟踪系统设备和系统指标要求，确定了图像预处理、图像分割、形心跟踪等算法。针对目标的特点，根据识别的目标用形心跟踪算法进行跟踪。文中所有控制程序都基于LabVIEW,在MFC框架下用C++编写图像处理程序。结果表明，在LabVIEW平台下的识别跟踪过程中，二维转台最终实现良好的控制,使整个系统快速、准确的跟踪运动的特征物体,完成了二维目标视觉检测与跟踪的目的。
주요대창신실험실이유시각근종전태적산법진행개진。전통적산법지능실현재백색배경하대흑색목표적식별여근종，위실현대불동목표복잡배경적근종，침대근종계통설비화계통지표요구，학정료도상예처리、도상분할、형심근종등산법。침대목표적특점，근거식별적목표용형심근종산법진행근종。문중소유공제정서도기우LabVIEW,재MFC광가하용C++편사도상처리정서。결과표명，재LabVIEW평태하적식별근종과정중，이유전태최종실현량호적공제,사정개계통쾌속、준학적근종운동적특정물체,완성료이유목표시각검측여근종적목적。
For realizing the tracking of different targets in complicated background, the algorithm such as im?age preprocessing, image segmentation, centroid tracking are confirmed according to the requirements of tracking system equipment and system characteristic. According to the characteristics of targets and the identified targets, centroid tracking algorithm is used for tracking. All the control programs are based on LabVIEW and C++is used to write image processing program under MFC frame. Results show that good control can be realized by two-dimension?al turret finally during identification and tracking process under LabVIEW platform. The whole system can track the moving feature object quickly and accurately. And two-dimensional visual target detection and tracking purpose can be completed.