沙粒图像中沙粒粒径的检测与识别

Detection and identification of grains of sand average size based on computer images

基于小波变换和图像分析理论检测和识别沙粒的平均粒径,使研究风沙运动中沙粒的粒径分布规律成为可能．首先通过小波变换对静态的沙粒照片进行消噪和增强对比度,使之成为可以南Otsu方法和灰度直方图峰值法进行阈值化处理的灰度图像．再将经过以上两种阈值方法处理得到的二值化图像利用八邻域边界跟踪的连通域标号算法对其进行逐行逐列扫描和搜索来确定并提取目标的颗粒数,最终实现对沙粒平均粒径的检测和识别．结果表明,该方法能够满足沙粒粒径识别的精度．

The distribution laws of the average grains of sand size on sand blown by wind movement are investigated based on wavelet analysis and the theory of image analysis. We firstly convert the static picture of grains of sand from which noise has been eliminated and enhanced contrast with wavelet transformation into gray image, to be treated by the Otsu method and gray histogram method. The number of objects for the binary images obtained from the two previous methods of threshold are scanned line by line, searched and confirmed. The detection and identification of the average grains of sand size are finally carried out. The results indicate that the method presented in this paper could satisfy the precision of the grains of sand size.