摘要
根据平动物体的运动学特征采用层的图像处理方法,实现自然环境中平动军事目标的识别和运动估计。在彩色图像的组成层上,进行图像恢复、背景噪音消除和轮廓提取等灰度图像处理。依据军事目标的特点,在各个图像组成层上采用独立的参数控制以实现最佳彩色图像处理效果。在已知目标模型的情况下,根据连续帧的处理识别确定模型的运动状态,结合平动的特点,实现目标模型的运动估计。在层上的图像处理效果优于灰度处理,利用目标模型模板,能较准确地判断目标模型运动状态。
With the application of the kinematical features to the layer image processing of translation objects,translation military target recognition and movement estimation could be obtained in natural environment.Grey-scale image processing such as image restoration,background noise elimination and contour extraction,etc.is drawn from the color image layer.In light of the characteristics of military target,separate parameter is adopted on each image layer to achieve the best effect of the color image processing.Given the object model,its motion state is identified and defined according to the processing of the continuous frames;relating to the translation features,the movement estimation of the object model is obtained.The results obtained from the layer image processing excels those of the grey-scale processing;when the object model template is confirmed,the motion state of the object model could be determined more accurately.
出处
《光学技术》
CAS
CSCD
北大核心
2006年第z1期162-165,168,共5页
Optical Technique
基金
高等学校博士学科点专项科研基金资助项目(20010610028)
华南理工大学博士后基金资助项目
关键词
图像识别
图像层
图像轮廓
平动
运动估计
image recognition
image layer
image contour
translation
movement estimation
