颗粒跳跃的计算机图像预处理方法

在非均匀沙床面上进行颗粒跳跃跟踪电视摄像试验得到一系列序列图像，可据以确定不同水流条件下颗粒跳跃的轨迹及速度，后者在输沙力学中具有重要意义．为获得有关颗粒跳跃的高精度数据，作者对摄像试验图像作了计算机图像处理和识别．首先将真彩色图像转化为灰度级图像，并对有用部分进行预截取．然后对这些图像采用半邻域法滤波和十字形中值滤波等平滑滤波方法进行处理，取得了较好的结果．这些图像的预处理工作为今后对图像作诸如阈值分割、轮廓线抽取和直线检测等的进一步处理奠定了基础．

Preprocessing methods for computer images of particle saltation

Particle saltation sequential images taken during a video\|image experiment on the bed\|surface of non\|uniform sediment can be used for determining the saltating grain trajectory and velocity, which are very important for studying the mechanics of sediment transport. In order to obtain high precision data on grain saltation, the authors resorted to computerized image processing and recognition of the video images. The methods for preprocessing particle saltation single image are described in this paper. Real color video images are first transformed into gray\|level images, whose useful parts are pre\|intercepted. Then, half\|neighborhood\|area filter and the cross median filter are used for moving average processing of these images. A satisfactory result is obtained. The pre\|processing of the images laid the foundation of the further processing such as the threshold segmentation, configuration extraction and edge detection.