Research On Measurement Method Of Ship Clearance Height Based On Infrared Imaging

With the increasing development of China’s transportation industry and the increasingly busy water transportation,collisions between ships and bridges occur frequently.Among them,the collision of ultra-high ships with bridges has become one of the main causes of water traffic accidents,which not only threatens the lives of people on board,but also caused huge property loss and brought bad social impact.In order to realize ultra-high ship early warning and reduce the probability of ultra-high ship collision,this paper studies a method of measuring the ship’s headroom height based on infrared imaging.This method combines the advantages of infrared imaging with strong ability to penetrate smog and can be monitored day and night.Infrared imaging system is established in the water area of the bridge area where the ship’s navigation direction is vertical to realize the acquisition of the infrared image sequence of the ship in the water area of the bridge area.Infrared image sequence is used for ship target detection,high-precision ship edge contour extraction and perspective transformation correction,etc,and the ship’s headroom height is measured according to the principle of the camera pinhole model.Ship target detection is a prerequisite for the measurement of ship headroom.Currently,commonly used target detection algorithms include wavelet transform method,visual attention model algorithm and local peak algorithm.Among them,the wavelet transform algorithm has a strong suppression effect on the interference with strong texture directionality(such as river waves,clouds),which is suitable for the detection of small targets.The local peak algorithm also has a good detection accuracy for weak targets.The visual attention model algorithm can realize the simultaneous detection of large and small targets in the image.The measurement targets in this paper are mainly large ships,so the visual attention model detection algorithm is selected to realize the detection of ship targets.High-precision ship contour extraction is a key technique for accurately measuring the ship’s headroom.If high-precision ship contour cannot be extracted,the calculation error will increase.In the context of this project,ships often form reflections on the river surface,causing the gray values of the ship and the reflection to be similar.Common contour extraction methods,such as the Canny edge detection algorithm,cannot achieve good results.The traditional pulse-coupled neural network algorithm(PCNN)can effectively achieve the segmentation of similar gray value areas,but this algorithm usually selects the initial threshold based on experience,there is a certain irrationality,so this paper proposes a pulse-coupled neural network algorithm based on large law to achieve high precision Extraction of ship outlines.In order to reduce the calculation error caused by the oblique projection of the ship in the infrared image,the perspective transformation correction algorithm is used in this paper.The three points of the target point,the image point and the perspective center are collinear.Project the image.The experimental results show that the ship headroom height error calculated by the method in this paper is small,which is of great significance to prevent the occurrence of traffic accidents caused by ultra-high ships passing through the bridge.