Research On Intelligent Location System Of Container Spreader Based On Machine Vision

Because the global economic and trade development and full liberalization of China’s external trade process, the container traffic to be growing rapidly. This requires container cranes to complete handling operations efficiently. Domestic container crane operators complete the work of alignment of the container by experience. Such work is not only costly, labor large demand, high strength, and low efficiency. Therefore, the automatic positioning of container spreader technology seems increasingly critical.This paper researched on the intelligent location system of container spreader based on machine vision. The content and the results are as follows:(1) The binocular stereoscopic camera parameter calibration and correction: Combining with OpenCV libraries and Matlab software in the use of the Zhang checkerboard calibration method to achieve the binocular camera parameter calibration procedure design.(2) The image preprocessing program analysis, design and realization:By smoothing processing, morphological processing, and linearization processing to reanalysis the images. Ellipse fitting method combined with the center of gravity to calculate the target of the center coordinates.(3) The design and realization of the fog images analysis module program: Using extract the dark-channel images method to complete the images automatic defogging process. The evaluation image defogging system can evaluate the effect of the images to defog objectively. The platform tests showed that the program works well to defogging process.(4) The container color automatic identification program design and realization: The higher sensitivity of HSV color model can separate color efficiently. The platform tests showed that the color separation correct essentially.(5) Research on three-dimensional imaging based on binocular vision and target coordinates feedback technology:Deduce the three-dimensional coordinates of the calculation method. Achieve the functional modules embedded programs completely. The error of the three dimensional coordinates of container locks is small by computation on simulation computing platform.This paper researched on the binocular machine vision works for the container spreader positioning simulation system which researched in the working principle, system construction, and other aspects of the image analysis. In the live images processing, image preprocessing methods under inclement weather conditions, container color automatic identification and other aspects provided the new programs. This provided a software platform for the subsequent synchronous processing of binocular vision.