资源列表
Imageedgedetect
- 图像处理示例程序，能相识不同处理方法的效果-Image processing sample application can be met the effect of different treatment methods
detecthand
- 识别手形，基于颜色信息。运行时需要配好opencv的环境-Recognition hand, based on color information. Need to run with a good environment for opencv
researchonDetectionandTrackofMovingHuman
- 本文构建了一套自动的运动人体检测跟踪原型系统，对从一台固定的数码摄 像机采集的视频图像进行处理与分析，能够检测和跟踪运动人体。该原型系统由 运动人体检测、运动人体跟踪、人群融合分离跟踪等部分构成 -This paper built an automatic detection of tracking the movement of human prototype system from a stationary digital video cameras capture the vi
main
- 图片显示的动态效果，如果想看到具体效果，请将图片换成大图片-the dynamic effects of showing picture
Analysisoftexturefeatureextractedbygraylevelco2occ
- 为使灰度共生矩阵(GLCM)提取的特征值较好地表达纹理信息,对Brodatz纹理库图片进行了大量实验。 首先测试了各构造参数对关键特征统计量的影响,给出了特征值随参数变化的规律,确立了构造参数的合理取值 然 后测试了图像旋转和大小变化对所提取特征值的影响-In order to grayscale co-occurrence matrix (GLCM) features extracted texture to express the value of good information
auto_detecthand_facecolor
- 基于颜色的自动找手程序，先由人脸检测取得脸的肤色，再以此去找手。运行需要配置好opencv环境。手放在脸部下方两侧-Color-based automated process to find hands, first face detection to obtain face color, and then in order to find hands. Need to configure your environment to run a good opencv. Hand on the bo
AReal-timeElectronicImageStabilizingAlgorithm
- 为了实时地稳定摄像系统的输出视频,提出了一种基于特征匹配的实时电子稳像算法。将改进 的Kanade2Lucas2Tomasi (KLT) 算法用于特征提取,并提出一种新的基于灰度投影均值的特征匹配算法。为了保 证稳像结果的鲁棒性,还给出了特征的有效性检验方法。-Real-time camera system in order to stabilize the output of video, presents a feature-based matching algorithm for
Real-TimeRobustAlgorithmofDigitalImageStabilizatio
- 电子稳像的实时鲁棒性算法，首先对采集到的每帧图像建立高斯金字塔。然后在高斯金字塔的顶层进行 初估计，采用图像块法得到整像素的运动估计矢量。精估计采用光流梯度法。为提高鲁棒性，运用简化的M 收敛法 进行梯度约束。最后阶段采用双线性内插对当前帧与参考帧进行匹准-Electronic image stabilization, real-time robust algorithm, first collected in the establishment of Gaussian pyramid
Textureclassificationusingspectralhistograms
- Based on a local spatial/frequency representation,we employ a spectral histogram as a feature statistic for texture classification. The spectral histogram consists of marginal distributions of responses of a bank of filters and encodes implicit
blinkeyes
- 通过眨眼来识别人的眼睛，比人脸识别速度要快些。能够辅助定位-Through the blink of an eye to identify the person' s eyes, faster than the speed of face recognition. We could help locate
noise
- 给图片加入噪声，有代码和示例图片 很简单的东西-To the image with noise, there is the code and sample images is very simple thing
imgstack
- java环境下开发的imagestack软件，绝对可以用，所谓iamgestack，就是说几幅连续的图像的合成，然后放在同一个图像中，以此来弥补焦距不同造成的部分失真，大家可以下载并使用-imagestack means several consecutive images combined into a sharp noisefree image,for example,from focus length 6.3 to 63mm ,we try to make up for the infor