摘要
设计了一种目标跟踪系统,该系统利用立体沙盘模拟复杂地面环境,通过红外和可见光传感器采集图像数据,利用红外图像传感器来检测目标,在检测到目标的基础上,利用可见光传感器来跟踪目标。目标跟踪采用粒子滤波算法实现,为了实现对目标的稳定跟踪,结合红外传感器的检测结果和可见光图像的跟踪结果对目标模板的更新。该图像处理算法在以VPX为设计架构、FPGA+DSP为核心的高速嵌入式图像处理平台实现,最终生成目标坐标偏差来控制伺服云台的转动,使其保持在视场的中心,并输出目标跟踪图像,实现对目标的跟踪。
In this paper,a target tracking system is designed. The system simulated a complex ground environment by utilizing a sand table and adopted infrared and visible sensor to collect image data,and then the target is detected by infrared image sensor and the target is tracked by visible sensor. Particle filter algorithm is used to realize target tracking. To realize the stable track,target template is updated according to tracking results of two kings of sensors. The image processing algorithm is realized through image processing platform based on VPX and FGPA ＋ DSP. The rotation of servo is controlled by target coordinate deviation,and the target is kept in the center of viewing field,which realized the target tracking.
出处
《激光与红外》
CAS
CSCD
北大核心
2015年第7期839-843,共5页
Laser & Infrared
