| 改进的二维最大熵分割夜视图像融合目标检测 |
| 投稿时间：2013-01-18 修订日期：2013-01-27 点此下载全文 |
| 引用本文： |
| 摘要点击次数: 832 |
| 全文下载次数: 1 |
|
| 基金项目:国家自然科学基金 |
|
| 中文摘要:夜视图像的目标检测主要利用红外和微光图像来实现。针对夜视成像的特性和目标检测算法的不足，提出了一种改进的二维最大熵分割的红外和微光图像融合目标检测的方法。首先通过对二维直方图的改进，即灰度级-加权的区域灰度最大值的灰度级建立二维直方图，选取权值利用该直方图计算最大熵对红外和微光图像分割，较传统最大熵分割算法在目标检测方面效果明显，具有抑制背景和提取目标的作用。然后验证多维特征的相与运算的有效性，对分割后的红外和微光图像特征级融合检测出目标。检测算法对于复杂背景下的目标检测及多目标检测方面都具有较好的效果和适用性。 |
| 中文关键词:微光和红外图像 目标检测 二维最大熵 信息融合 |
| |
| Improved 2D Maximum Entropy Segmentation Fusion ForTarget Detection |
|
|
| Abstract:Infrared image and LLL image are used for target detection of night vision image. In allusion to the characteristics of night vision imaging and disadvantages of target detection algorithm, we propose a method of infrared image and LLL image fusion for target detection with improved 2D maximum entropy segmentation. Firstly, two-dimensional histogram is created by gray level and maximum gray level in weighted area, and then weights are selected to calculate the maximum entropy with infrared image and LLL image segmentation by using the histogram, which, compared with the traditional maximum entropy segmentation, has significant effect in target detection and the functions of background suppression and target extraction. And then, the validity of multi-dimensional characteristics with AND operation in the characteristic infrared image and LLL image fusion for target detection is verified. Detection algorithm has a relatively good effect and application in target detection and multiple targets detection in complex background. |
| keywords:Infrared and LLL image target detection 2D maximum entropy data fusion |
| HTML 查看/发表评论 下载PDF阅读器 |