航空图像的分割与定位算法研究  

Segmental and determining approaches based on aerial films

作　　者：李迎春[1] 陈贺新[1] 张建新[2] 

机构地区：[1]吉林大学通信工程学院,吉林长春130012 [2]空军第二航空学院,吉林长春130022

出　　处：《吉林大学学报（信息科学版）》2003年第3期243-246,共4页Journal of Jilin University（Information Science Edition）

基　　金：空军装备部重点科研基金资助项目(kj02426)

摘　　要：通过航空图片的图像分割和特征提取,对几种分割算法进行了比较。通过比较发现,当图像的目标与背景的灰度值相差比较大时,用多阈值的灰度分割方法来获取目标;而当图像的目标与背景的灰度值相差不大时,使用模糊差影滤波算法可获得理想的分割结果。同时,对采用多阈值分割方法后的图像进行目标定位。结果表明,对于复杂航空图像,该算法分割效果好,噪声小,定位算法简单,定位准确,有较好的应用前景。Several segmental approaches for image segmentation and feature extraction of aerial films are presented and compared. Experiments show that multithreshold segmental approaches can be used to deal with bigger grayvalue differences between object and background. However, Applying the fuzzy difference filtering arithmetic results in a perfect segmental conclusion in less grayvalue differences between object and background. The objects in the image after being processed by multithreshold segmental approaches can be detected. The approaches presented are simpler and have better effects, less noise and accurate situation in the complicated aerial images.

关 键 词：图像分割 灰度 多阈值 模糊差影滤波 检测定位 

分 类 号：TN941.1[电子电信—信号与信息处理]