Abstract: Because of its low cost, carrying capacity, etc., waterway transport plays an important role in our comprehensive transmission system. However, due to the different size of ship construction, nonstandard navigation operations and the lack of effective control measures, accidents such as ships colliding, hitting the bridge or going aground often occur. When referring to technology from road traffic video surveillance, difficulties occur, such as the camera is difficult to calibrate without markings on water, it’s hard to distinguish ships from their stern ripple which move together, sparkling water doesn’t have high contrast like the black asphalt road and so on. So this paper proposes a series of ship parameters detection algorithm based on video: proposed grid interactive camera calibration algorithm to obtain the mapping image between coordinate system and the world coordinate system; optimization Vibe algorithm to remove the stern ripple interference, thereby extracting the ship contour; establish detection model of ship length and width, speed, traffic, in order to get the ship size and operating parameters, respectively. The system has been identified in the Beijing-Hangzhou Grand Canal. It is tested with width, speed, and flow detection accuracy, respectively, better than 90%, 95%, 98%.