彩色瓷砖的自动分类系统

将彩色图像处理技术应用于机器视觉。为了检测瓷砖表面的细微色差，使用低端采像设备实现精确色彩测量。研究了彩色成像系统在不同光照及及噪声条件下的色彩校正原理，提出了一种矩形物体图像分割的简易得法和对训练样本进行修正的方法。探讨了特征参数与分类器的关系，提出分层分类并采用不同特下参数的方法来加速计算。

Automatic Grading System for Color Ceramic Tiles

An application of color image processing to the machine vision is described. To detect color nuance between ceramic tiles, one needs to measure color value accurately by using low cost imaging devices. Noise is mainly introduced by the variance of illumination and also by the auto aperture of video camera. A color calibrating method is designed to reduce the noise. An easy method is also designed to separate square objects from the image background. The intrinsic and subtle relationships between selection of features and classifiers are discussed. The different features and classifiers in different hierarchies of classification process are used to minimize the computational burden.