一种基于帧间差分与模板匹配的河水表面流速测量方法  

MEASURING SURFACE VELOCITY OF RIVER FLOW BASED ON IMAGES OF FRAME DIFFERENCE AND TEMPLATE MATCHING

作　　者：赵晓刚 冯全[1] 王书志[2] 

机构地区：[1]甘肃农业大学工学院,甘肃兰州730070 [2]西北民族大学电气工程学院,甘肃兰州730030

出　　处：《计算机应用与软件》2017年第9期68-71,107,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61461005)

摘　　要：为了解决常见几种流速测量方法受含沙量与漂浮物影响大、设备与人力成本高的局限性,提出一种基于机器视觉的河水表面流速测量方法。我们分析了图像空间与现实空间的位置映射关系,得出了物理速度与图像速度之间的关系。通过垂直于水面的摄像头获取河流视频,采用帧差处理获得帧差图像,检测出其中运动显著性比较明显的一些区域作为模板。最后,基于模板匹配的方法检测出模板的像素位移,联立图像坐标系和与现实空间坐标系的位置映射关系以及帧间时间计算出河流表面实际流速。实验结果表明,该测速方法对自然条件下流速相对平缓的河流视频具有较好的效果,误差小于3%,可持续监测水流流速情况。There are some limitations in traditional velocity measure methods of river flow. For example,sediment concentration,floating objects and the cost of device and human resource are too high. In order to overcome the limitations,this paper proposes a method to measure the surface velocity of the river flow based on computer vision. We analyzed the relationship between image space and real world space,and derived the physical velocity from the image velocity. We captured the video of river by a camera whose optic axis was perpendicular to the surface of the river. The frame difference method was used to produce difference image in which the salient moving regions were detected as templates. In the succeeding difference image,the template matching algorithm was employed to find out the regions matched those template best,indicating the templates＇ movement. The shift of a template between two images was evaluated by the pair of the template and the corresponding region and used to calculate the velocity of the template. The experimental results show that the proposed method can evaluate the surface velocity effectively and accurately for smooth flow. The error is lower than 3%,and the flow velocity can be monitored continuously.

关 键 词：流速测量 模板匹配 视频处理 帧差法 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]