SIFT-EMGUCV

bresenham画直线，WIN32，计算机图形学基础类源代码(bresenham画直线)

指纹图像预处理程序，包括分割，二值化，去噪和细化(Fingerprint image pre-processing procedures, including segmentation, binarization, denoising and refined)

数字图像处理编程入门（吕风军）第一章第1章 Windows位图和调色板(Windows Bitmap and palette of the first chapter, Chapter 1 of the digital image processing program entry (Lv Fengjun))

使用C++和CUDA编写的双边滤波器。在平滑图像的同时可以保持边界(Using bilateral filter the C++ and CUDA written. Boundary can be kept in the smooth image while)

以SIFT特征为目标输入，LDAHash方法进行降维实现目标识别(SIFT features input LDAHash method to reduce the dimensionality of the target recognition)

matlab小波分解信号的几个源程序，包括小波分解，时频转换一级功率谱分析(Matlab wavelet decomposition of the signal source program, including several wavelet decomposition, time-frequency transition level 1 power spectrum analysis)

Matlab图像处理中的Harris角点提取算法， 代码中有详细的解说，方便阅读。 (Matlab Image Processing of Harris corner extraction algorithm, the code has a detailed explanation, easy reading.)

高斯滤波是一种线性平滑滤波，适用于消除高斯噪声，广泛应用于图像处理的减噪过程。通俗的讲，高斯滤波就是对整幅图像进行加权平均的过程，每一个像素点的值，都由其本身和邻域内的其他像素值经过加权平均后得到。高斯滤波的具体操作是：用一个模板（或称卷积、掩模）扫描图像中的每一个像素，用模板确定的邻域内像素的加权平均灰度值去替代模板中心像素点的值。(Gaussian smoothing filter is a linear filter for the elimination of Gaussian noise, the noise reduction process is widely used in image processing. Popular speaking, a Gaussian filter to the whole image is a weighted average of the process, the value of each pixel, both by itself and other neighborhood pixel values obtained after weighted average. The specific operation Gaussian filter is: a template (or a convolution mask) to scan each pixel in the image, a weighted average gray value of the neighborhood is determined using a template to replace the pixel value of the center pixel of the template.)

说明：  本设计为基于MATLAB的指纹识别系统。本设计系统主要对指纹图像进行三方面处理：图像预处理、特征提取和特征匹配。图像预处理包括四个步骤：图像灰度化、滤波增强、二值化、细化，对指纹图像进行预处理后，去除了原图像的冗余部分，方便后续的识别处理；特征提取主要是提取指纹图像细化后的端点和分叉点；特征匹配是利用两个指纹的图像进行特征点比较，来确定两幅图像是否来自于同一手指。(This design is a fingerprint identification system based on MATLAB. This system mainly deals with fingerprint image in three aspects: image preprocessing, feature extraction and feature matching. Image preprocessing includes four steps: image grayscale, filter enhancement, binarization and thinning. After preprocessing the fingerprint image, the redundant part of the original image is removed for the convenience of subsequent recognition processing. Feature extraction is mainly to extract the endpoint and bifurcation points after thinning the fingerprint image. Feature matching is to determine the two images by comparing the feature points of the two fingerprint images Whether the image comes from the same finger.)

白噪化处理，这是高光谱图像数据降维的前预处理，也称为球化，在数据降维中必须用到(White noise process, which is pretreatment hyperspectral image data dimensionality reduction, also called the ball in the data dimensionality reduction must be used)