source

说明：  基于Kmeans算法的图像分割，一般Kmeans是数据挖掘中用来聚类的，本试验利用图像中的灰度值作为Kmeans算法的原始点，进行图像分割(Kmeans algorithm based on image segmentation, data mining in general Kmeans is used to clustering, the trial use of the image gray value as the original algorithm Kmeans spots for Image Segmentation)

图像球面化，对于输入的图片进行球面变换达到图形变形的效果(Spherical image of, for input images to spherical transformation to achieve the effect of graph distortion)

canny边缘检测（自动设置Sigma的高斯平滑+Sobel(原始)+自适应直方图阈值）(canny edge detection (automatically set Sigma Gaussian smoothing+ Sobel (original)+ adaptive histogram threshold))

图像去雾 暗原色 导向滤波 最好的去雾算法 何恺明的matlab代码的opencv的实现(haze removal guided image filtering the best algorithm to fog )

VC++Digital image processing

用kinect获取3D图像，并从3D图像中用OpenCV方法从深度图中获取梯度边缘，从而得到更多信息，做后续处理(Get with kinect 3D images, and get the edge gradient from the depth map from a 3D image using OpenCV method to obtain more information, do follow-up treatment)

说明：  ：针对显微图像分析、识别需要的全自动控制显微镜，研究开发其相应的图像处理算法。提 出了能量谱自动聚焦评价函数算法、自适应选取基准图的图像拼接算法、改进 Laplacian算子的多层 聚焦图像叠合算法等。该算法模块已成功地应用于所开发的CMIAS 显微医学图像分析系统中，取得 了满意的应用效果。(【ABSTRACT】For the automatic microscope required by micro-image analysis and recognition, the relevant image processing a lgorithms are investigated. Three algorithms are described: energy-spectrum image auto focusing evaluation function, image merging based on the adaptive selection of standard image and sum-modified-Laplacian (SML) operator for multi-focus image fusion. The algorithms have been applied successfully to CMIAS medical micro-optical image analysis system and testified to be feasible and effective. )

desing and develop a cube rotation

对图像进行纹理处理，为图像识别、目标跟踪创造条件。主要分析：惯性矩、熵以及局部平衡性。(The image texture processing for image recognition, target tracking and create conditions. Main analysis: moment of inertia, entropy, and local balance.)

根据RBM深度网络（Hinton，2006）进行MINST手写文字的分类器的训练，利用训练得到的权值制作了这个小程序，通过这个程序可以看出训练结果对数据集内的测试样本和训练样本都能进行很好的识别，但是对其他的手写字体识别就没有那么好了。(According to RBM depth network (Hinton, 2006) conducted MINST handwritten text classifier training, using the training to get the right value to produce this small program, this program can be seen through the results of the training data set and the training sample test samples are can be well identified, but on the other handwriting recognition is not so good.)