Research On Key Technology Of Rubber-Tyred Gantry Crane Automation Based On Machine Vision

Although,the container throughput of C hina's seaports was ranked first in the world for many years,but in Scheduling management level,automation degree of loading and unloading transport machinery are still far behind to foreign advanced ports.At present,the domestic port container loading is also at the level of manual operation,poor safety,low efficiency and competitiveness.So,how to improving the loading and unloading machinery effectly,especially the automatic level of the Rubber-Tyred Gantry Crane(RTG),has become a key problem of research.With the development of computer,image processing,artificial intelligence and intelligent control technology,the flexible automation technology based on machine vision can be realized and developed rapidly.This topic takes the improvement of RTG automation level as the research target,and has done the theoretical and experimental research to its two key technologies-container recognition,location and tracking based on vision;anti-sway control technology of ropes based on visual servo.The main work is as follows:(1)Put the machine vision technology into the container recognition and location,and the optimization of visual recognition algorithm was studied.Firstly,the image was segmented quickly by the combination of the color template pre-location and self-adaptive variable step size block matching segmentation method.Then,according to the segmentation of binary image using Blob particle characteristics analysis and shape invariant moment feature analysis method to extract the five shape invariant features of the container,and combined with the length-width ratio to compose an six dimensional feature vector as the basis for container recognition.(2)Hardware and software structure of the system based on vision measurement was designed,and applied it to the rope swing angle measurement.It realized the real-time and accurate measurement of rope swing angle.(3)The two degree of freedom dynamic model of RTG trolley and spreader was founded.A conclusion that the main influe nce factors of rope swing angle was obtained,that is the trolley's speed,or acceleration,hanging weight and the length of rope.In the analysis of advantages and disadvantages of traditional anti-sway method,A kind of fuzzy adaptive PID anti-sway control method of image-based visual servo was proposed.Its principle is to locate the container coordinate in horizontal plane in real-time,and guide the trolley to track it by a camera below the spreader.Besides,to be combined with the swing angle information of ropes obtained by other camera as two input variables of the controller,using fuzzy adaptive PID parameters adjustment on line to track the ideal velocity curve,so as to achieve the unification of container localization with the rope anti-sway control.(4)On the basis of theoretical research,the experimental platform of RTG container loading and unloading was built,and finished the verification of theoretical research on this platform.The main work includes the following three aspects: Firstly,the feasibility of remote control of the system was verified.Secondly,the theoretical verification for container visual recognition,location and tracking algorithms were accomplished.Thirdly,the validation of spreader fuzzy adaptive PID anti-sway control effect based on visual servo was completed.The experiment finished verification of planned contents,and achieved desired verification purposes basically.