摘要
在图像跟踪中主要因素是目标的几何中心,只考虑目标的相对运动特性,就可以设法减少图像处理的工作量。由于目标的边缘不可能占满整幅图像,在目标边缘检测时不必对整幅图像进行处理。为此,从估计的中心向给定方向寻找目标的边界,得到目标边界点的一个子集合,再由该子集得到目标中心的近似值,并以此为基础构造目标跟踪方法,利用图像检测实现无人战斗机对目标的跟踪。仿真结果表明,该方法对于目标的跟踪是有效的。
The heavy calculation to target center's matrix has influenced the speediness. We should pay much more attention to the relative movement character of a target and manage to reduce the work of image processing because the main factor is location of the target's geometry center. We needn't process the whole image when we test target's margin for the target's margin being full of whole image is impossible. We can get a son aggregate of target marginal points if we search the target's margin from estimated center to decided direction. We also receive a approximate value of target's center from the son aggregate. Based on that, construct the method of target tracking, that is to say, realizing self-service battle plane's tracking to the target by using the image testing. This tracking method provides the estimate of target's center location and real-time revise. The simulation result shows that it is effective to tracking.
出处
《飞行力学》
CSCD
北大核心
2005年第1期79-81,85,共4页
Flight Dynamics
基金
国家自然科学基金重大研究计划资助项目(90205019)
高等学校博士学科点专项科研基金(20020699001)
