2

利用数学形态学对云层背景下的目标进行检测识别，有两种方法。 (Using mathematical morphology of clouds in the context of target identification, there are two ways.)

对图像目标识别和提取，主要通过改进非线性阈值分割实现(For image target recognition and extraction, the main method is to improve the non-linear threshold segmentation.)

OPENCV做图像变形的经典代码啊，适合比较初级的同学看看(OPENCV do image warping the classic code ah, suitable for elementary students to see more)

说明：  使用matlab编写的源程序 高斯噪声产生的源代码(prepared by the use of Matlab source Gaussian noise source)

利用摄像头工具箱标定得到的内部参数还原透视图像为“鸟瞰图”。本例中使用的图像时棋盘格。(Using the internal parameters obtained from the camera calibration toolbox. reducting perspective images as the "bird s eye view". In this case, use the image of the checkerboard.)

自己编写图像平滑算法：均值滤波和中值滤波，设置不同的均值滤波模板，以得到最好的图像输出结果。(I have written the image smoothing algorithm: mean filter and median filter, mean filter set different templates to get the best image output.)

说明：  图像分块的处理和matlab水印的提取，(Image sub-block processing and matlab watermark extraction,)

人头检测 通过颜色分割产生人头头发部分的二值图像，计算此部分的面积和长宽比来判断是否有人头(human detection)

说明：  基于MATLAB，对逐日的sst遥感数据做月平均并制作gif动图(Based on MATLAB, the daily SST remote sensing data are averaged monthly and the GIF map is made)

说明：  用POCS方法对图像进行超分辨率重构，matlab源码，有解释(POCS method using super-resolution reconstruction of images, matlab source code, to explain)