use-matlab-for-GPU

说明：  序贯重要性采样——标准粒子滤波算法。可直接使用。(Sequential importance sampling- standard particle filter. Can be used directly.)

CVPR2011文章Efficient_Rotation_Invariant_Object_Detection_using_Boosted_Random_Ferns(paper of CVPR2011 Efficient_Rotation_Invariant_Object_Detection_using_Boosted_Random_Ferns)

将经过二值化的车道线图利用hough变换提取车道线(After binarization the lane line graph using hough transform to extract lane)

通过sift算法进行图像配准，配准效果好，精度高(Image registration algorithms sift through the good registration results, high precision)

实现基于混合高斯模型的运动目标检测算法的仿真(Achieved based on Gaussian mixture model simulation of moving target detection algorithm)

基于点云的谷脊线特征提取算法研究：提出一种基于多步逼近策略的点云特征提取算法(Exaction of ringle and ravine.)

视频编解码技术，在有限带宽的情况下，能够大幅度的压缩编码比特数，图像的质量没有丝毫的下降。(Video codec technology, in the case of limited bandwidth can be substantial coding bits, the image quality is not the slightest decline.)

立体视觉代码。主要过程包括：harris角点检测，zncc最优策略匹配，扩散得到更多特征对应，计算disparity和对应颜色值。另外提供了两幅图片以及对应的特征匹配的结果和disparity图像。(Three-dimensional visual codes. The main process includes: harris corner detection, zncc optimal strategy for matching, the proliferation of more features of the corresponding calculation of disparity and the corresponding color value. Also provides a picture and two corresponding feature matching results and disparity images.)

图像匹配，将两幅不同的角度拍的图像的对应点都找到(Corner detection)

合成孔径雷达（SAR）回波模拟半实物仿真平台的构建(Synthetic aperture radar (SAR) analog echo-loop Simulation Platform)