SaliencyToolbox2.2

将RGB彩色空间转换到HSI彩色空间，HSI彩色空间更接近人眼对彩色的认知。(The RGB color space conversion to the HSI color space, HSI color space is closer to the human eye color perception.)

这MFC图像处理源代码总集包括了均值滤波，直方图均衡，高斯平滑，，锐化，拉普拉斯锐化（边缘检测），阈值变(With the MFC implementation of an image processing example code)

说明：  图像分块处理和计算图像灰度平均值的算法，可用于路面识别预处理(Average value of image block calculation)

读入光流估计的结果，将运动矢量转化成颜色，生成颜色的光流估计图像。(Read optical flow estimation result, the color motion vector conversion, optical flow estimation to generate a color image.)

使用GDAL库，用于遥感影像的特征点提取(Using the GDAL library, used for remote sensing image feature point extraction)

本压缩包有canny算子的源程序代码和测试图片。(The archive has canny operator of the source code and test images.)

说明：  （1）图像基本操作：不同格式（大于3种）图像的读入与存盘、文字叠加、不同彩色空间的转换、图像的DCT及FFT变换等； （2）图像增强：包括直方图拉升（线性和非线性）、直方图均衡、平滑与锐化（采用不同的滤镜），美颜（加分项）； （3）图像恢复：几何操作（如旋转、缩放、投影校正等）、模糊恢复（如运动模糊消除等，加分项）； （4）图像合成（加分项）：实现换头、换背景、图像拼接等功能。((1) Basic operation of image: reading and saving of images in different formats (more than 3 kinds), text superposition, conversion of different color spaces, DCT and FFT transformation of images, etc; (2) Image enhancement: including histogram lifting (linear and non-linear), histogram equalization, smoothing and sharpening (using different filters), beauty (bonus item); (3) Image restoration: geometric operations (such as rotation, scaling, projection correction, etc.), blur restoration (such as motion blur elimination, etc., plus points); (4) Image synthesis (Bonus): realize the functions of head changing, background changing, image splicing, etc.)

3、 请采用学过的图像复原算法（鼓励自己研究新算法），对 blur3.jpg、 sadhna_53_45.bmp和telescope.bmp 图像进行复原处理。(3, please use the learned image restoration algorithm (to encourage their study new algorithm), to blur3.jpg, sadhna_53_45.bmp and telescope.bmp image restoration processing.)

使用OpenCV实现特征点的查找以及两幅图像中的特征点匹配(Use the function findHomography to find the transform between matched keypoints. Use the function perspectiveTransform to map the points)

说明：  将图像进行各个角度的旋转，以及对两幅或多幅图像进行叠加(Each rotation angle of the image, and the two or more images are superimposed)