【作者】 张庆利；

【导师】 张仁忠；

【作者基本信息】 哈尔滨工程大学 ， 模式识别与智能系统， 2007， 硕士

【摘要】 本文是针对某洗煤厂生产线提出的煤炭灰分在线检测系统。传统检测方法是取一定量待测煤样本，将煤燃烧后检测煤炭灰分(根据燃烧前、后的重量比得出杂质百分含量)，进而判断所生产煤炭灰分是否合格，从而控制洗煤生产设备。此法效率低且无法达到实时控制洗煤系统。本文针对传统方法的缺点，提出了一种新的基于ARM图像处理方法----利用图像处理方法实现煤炭灰分在线检测。该方法首先通过摄像头摄取煤炭原始灰度图像，然后将原始图像通过线性加权均值滤波器降低噪声，再通过线性灰度变换去除光照影响，最后从概率统计角度以概率及方差为特征，对预处理后的灰度值矩阵提取特征，通过分组特征匹配算法得到所洗煤炭是否合格的判定值，根据判定结果输出一组控制信号达到对洗煤机器的实时控制。更多还原

【Abstract】 This paper puts forward an Online Assessing System of coal ash targeting the Coal-Washing factory. The tradition approach for this purpose is entirely manual: firstly, we fetch some coal for assessing, and then judge whether the product is consistence with the expectant result by burning the coal to ashes; finally, we adjust the washing system according to the judge. This method not only has a lower efficiency, but also can hardly achieve the real-time control.To solve the problem, we develop a new image processing approach based on ARM system to implement the online assessing system. This method first obtains the primitive grayscale image of coal from the camera through the linux driver layer. And then reduce the image-noise through a liner-weighting-average filter and eliminate the illumination influence through liner transformation. Finally abstract the main features of the grayscale image matrix based on probability and deviation. By using the group feature matching arithmetic, we can get the final criterion according which to output a set of signals to achieve the real time control on the washing machine.更多还原