摘要
为了保证处理效率,需要通过图像分割出可疑目标来减少干扰、减少运算,提出了一种基于知识的目标分割方法.将图像区域作为研究对象,结合目标先验知识通过区域划分、特征提取及判定来选取合适的目标分割阈值.实验结果表明:该方法能够克服全局场景的影响,有效地将可疑目标分割出来,提高了系统的实时性,为目标的识别提供依据,同时降低了后面目标识别跟踪的难度.
Strapdown seekers abandon the traditional stable platform, through a large field of view toensure effective detection of the target, which leads to more complexes imaging target scene. Whetherpeople in the loop or automatic target recognition, the real-time and reliability are affected worse.Therefore image segmentation is needed to improve treatment efficiency, segmentation of suspicioustargets to reduce interference while reducing operational. A target segmentation method based onscene knowledge was proposed,the image area as research object,combined with a priori knowledge ofthe target through zoning, feature extraction and judgment to select the appropriate targetsegmentation threshold. The experimental results showed that.this method can overcome the effectsof the global scene, effectively split up the suspicious targets, improved real-time of the system,provides the basis for identifying targets ,while reducing the difficulty of target tracking.
出处
《东北师大学报（自然科学版）》
CAS
CSCD
北大核心
2014年第3期75-80,共6页
Journal of Northeast Normal University(Natural Science Edition)
基金
中国科学院"三期创新"平台资助项目
关键词
目标分割
捷联导引头
基于知识
target segmentation
strapdown seeker
knowledge-based
