Research On Quay Crane Location Guide Slot Key Technology Based On Computer Vision

To achieve automation,intelligent terminal is the future development trend of many ports and terminals.However,there is still a lack of efficient and stable approach for loading and unloading operations of quay crane.Therefore,this article focuses on the automatic port stevedoring about the guide slot problem of location and it puts forward the corresponding technical methods based on the mathematical model of the binocular vision system,detection the container ship and hold area and Location guide slot.In this paper,the main work and achievements are as follows:First,we build the binocular vision system and location model of the guide slot,which are the foundation for the further realization of the target location.Combined with the principle of binocular vision and the error factor in hardware installation of quayside,we obtain the error mathematic model for binocular vision.Then,based on the structure of the container ship and the installation process of the ship's guide rail,the problem of locating the hold and slot is transformed into a straight line detection problem with edge characteristics.Finally,according to the mathematical model calculate the location model of the hold and slot.Second,we propose a method of hold location based on visual scanning,which provides the displacement basis for next step of the quay-loading.Obtained the image of the hold area under the quayside by binocular vision system.Then,serveral lines can be detected by the mathematical morphology edge extraction and Hough techniques,based on clustering analysis,and prior knowledge,a filtering model is proposed to detect the hold from these lines.Lastly,the least squares method was used to fit the linear set and the variance of the pixel gray value was calculated for each straight line,and the minimum variance line as the boundary line to detection of hold.Third,we propose a method to locate the slot based on error compensation.Based on the position of hold's area,which is divided into independent rectangles according to the slot location model.Then,analyzing the binocular mathematical model of the ideal and actual type quayside,and the error correction in the X and Y directions are also divided into two parts,and the error of the spreader in the X direction is calculated by detecting the local boundary line of the hold.Finally,combined with the head feature of guide slot and the machine learning to achieve position.Next,extracted the edge feature and dectected the boundary line,calculated the error correction amount relative to the Y direction of the spreader,which finish the accurate loading of the container.