CAJ | 학술논문

本文基于DM6437-355开发板结合OpenCV与VC技术，设计出双跟踪算法的目标跟踪系统。本系统将形心跟踪算法和重心跟踪算法这两种算法同时设置其中，形心跟踪算法是以阈值分割图像预处理为基础的改进方法，重心跟踪算法则提出为每个采样点设置不同的权值I(x,y)，再来求取重心坐标的改进方法。这两种算法对于在目标跟踪的过程中出现的由于跟踪目标发生翻转、形变或是目标被部分遮挡而导致跟踪目标丢失现象的避免有了明显的提高。根据不同的情况要求选择不同的跟踪算法，加强了在复杂多变的实际环境中目标跟踪的效果。
본문기우DM6437-355개발판결합OpenCV여VC기술，설계출쌍근종산법적목표근종계통。본계통장형심근종산법화중심근종산법저량충산법동시설치기중，형심근종산법시이역치분할도상예처리위기출적개진방법，중심근종산법칙제출위매개채양점설치불동적권치I(x,y)，재래구취중심좌표적개진방법。저량충산법대우재목표근종적과정중출현적유우근종목표발생번전、형변혹시목표피부분차당이도치근종목표주실현상적피면유료명현적제고。근거불동적정황요구선택불동적근종산법，가강료재복잡다변적실제배경중목표근종적효과。
In this paper, a double tracking algorithm target tracking system based on DM6437-355 development board combining with OpenCV and VC technology is designed. Both the centroid tracking algorithm and the center of gravity tracking algorithm are set into the system,the centroid tracking algorithm is an improvement based on threshold image segmentation pretreatment,the center of gravity tracking algorithm is proposed for setting different weights I(x,y) for each sample point, it is an improvement solving the coordinates of the center of gravity. The phenomenon has been significantly improved, that the tracked target is lost by the target rolling and deforming and partially being shaded. A different tracking algorithm depending on different situations is chosen,the target tracking results in the midst of a real environment is strengthened.