| With the implementation of a series of national strategic plans for inland waterway shipping development,the volume of inland waterway shipping shows a rising trend.Inland river shipping has become an important transportation mode because of its advantages such as low energy consumption,large capacity,less land occupation and less pollution.In order to reduce cost and seek maximum benefit,some ship operators disregard safety and carry out inland waterway ship overloading and over-draft transportation,which brings serious threat to water transportation safety and the ecological environment.China’s maritime departments for a long time on the actual draught of ship inspection,are using manual observation to achieve,this way time-consuming,low efficiency.Therefore,the management of the phenomenon of overload and overdraft in inland navigation has become a hot issue for research.In order to effectively guarantee the navigation safety of inland waterway vessels,a method is proposed to detect the ship’s freeboard and draught line and measure the ship’s freeboard height without the close-up ship’s water gauge mark.The research of this paper is carried out in two aspects as follows:(1)This thesis presents a K-means clustering algorithm based on segmentation and a freeboard vision detection algorithm based on Hough transform.This algorithm does not require close-up image of the water gauge position of the ship,and does not need contact to detect the ship’s side and draft line.In the image distribution feature,the ship’s position is used as the center dividing out,above the sky and below the water surface.Then,based on the regularity of color pixels,Lab color space adaptive K-means clustering is used to locate to the ship area in the image to eliminating the influence of water waves on edge detection;Finally,the binary image after Canny edge detection is processed by morphological expansion operation,and straight line detection by Hough transform is performed,which effectively reduce the appearance of “false” straight lines and detects the ship’s side line and draft line more accurately.(2)Freeboard height is measured by monocular and binocular cameras.The monocular camera calculates the freeboard height value by using the reference object proportion idea,and then carries out Zhang’s calibration and fixation on the camera to make the camera image a certain size each time,and obtains the actual size of each pixel according to the reference object;The binocular camera uses 3D reconstruction technology to establish the relationship between the world coordinate system,and calculate the freeboard height.Firstly,we calibrate the chessboard grid and get the internal and external parameter matrix of the camera;Then,the depth information of the object is obtained by using the epipolar geometry and triangulation principle of the monocular camera;Secondly,the actual size of each pixel in the image is obtained according to the reference object ratio;Finally,according to the detected freeboard height pixel value,calculate the freeboard height.Its practical significance is to compare the freeboard height with the standard freeboard height.If the freeboard height is less than the standard height,it will give an early warning and inform the maritime department for supervision.The automatic detection of ship’s freeboard height not only helps the maritime department reduce management costs and replace manual inspection;it also effectively curbs the occurrence of ships violations and guarantees navigation safety;It is of great significance in reducing loss of life and property and protecting the environment. |