chapter-5

基于统计算法的红外图像非均匀性矫正算法，场景类算法，实现实时校正(IRFPA NUC)

说明：  基础图像处理语言，RGB各图层提取，量化图像，lena典型案例代码。(Basic image processing language, RGB layer extraction, quantization image, Lena typical case code.)

改进的数学形态学小目标检测，用于红外图像中的点目标检测，比传统的Top-Hat变换检测精度高(Small target detection improved mathematical morphology for point target detection in infrared images, higher detection accuracy than traditional Top-Hat transform)

自己用MATLAB编的一维信号数学形态学的腐蚀与膨胀运算以及开闭运算，可用于滤去信号中的尖锐脉冲信号，适合于用于振动信号的预处理(Own series of one-dimensional signal using MATLAB mathematical morphology, corrosion and expansion of operations, as well as the opening and closing operation can be used to filter the signal sharp pulse signal suitable for vibration signal preprocessing)

this code is for the paper of image segmentation based on level set methods. in this paper improves Tony Chan s paper and improves the speed.

：介绍了一种基于小波变换的图像融合算法，它利用像素邻域平均梯度作为选择依据，对小 波变换后的高频系数进行选择，再逆变换完成图像融合．通过对红外热图与可见光图像的融合实验表明： 基于小波系数区域平均梯度的融合能使红外热图更加清晰，融合的图像光谱变异性较小，图像效果较好(: This paper presents a wavelet-based image fusion algorithm, which uses an average gradient of pixel neighborhood as selected on the basis of high frequency coefficients of wavelet transform to choose, and then inverse transform of image fusion. Through the infrared image and visible image fusion experiments show that: Based on the average gradient of wavelet coefficients of the regional integration of thermal infrared images can be more clarity, variability in integrated spectrum of images smaller, better image)

混合高斯背景建模，可以把连续帧图片前景与背景区分开来(Mixed gaussian background modeling, can continuous frame picture foreground and background to distinguish )

本程序可以实现MPEG-2I帧基本层编码和解码。图像大小为352*288.(This program implements the MPEG-2 I base layer frame encoding and decoding. Image size is 352* 288.)

说明：  分类，纹理特征，提取特征向量，matlab 7.0 原程序，classifer,texture(Classification, texture features, feature vector extraction, matlab 7.0 original procedure, classifer, texture)

在GDAL库下，采用主成分变换实现多光谱影像与全色高分辨影像的融合。(Under the GDAL library, the principal component transformation is used to fuse multispectral images with panchromatic high resolution images.)