Research on Automatic Measurement Technology of Water Level Based on Image Processing

A real-time hydrological monitoring system was designed using the combination method of the ARM processor and video server,which can realize the extraction of the characteristic values in the lower computer and the following communication with the upper computer through the GRPS.The system used the Wince programming to obtain the binarization of the collected image.The morphological algorithm was employed to remove the noncharacteristic values and the edge of the image was also refined.The Hough transform was used to extract the linear equation of the water level and bankline.According to the conversion relationship between the pixel space and the coordinate coefficient,the distance between two points in the image can be obtained,which can be used to estimate the physical distance based on the above conversion relationship.The results showed that the system has the characteristics of noise resistance,stability and reliability,and strong real-time,which has good recognition effects.