Research On Container Location System Based On Image Recognition

With the continuous development of container automation ports,container handling efficiency has become the bottleneck of container transportation.How to realize the automation of container loading and unloading in order to improve the efficiency of container loading and unloading in the port has become one of the hot spots in the field of logistics at home and abroad.At present,container intelligent operation mainly uses computer image processing technology to realize container identification and location.The identification of container accessories is very important for the automation of large mechanical equipment and more researchers are needed to develop smarter ways to automate them.In this paper,based on image processing and recognition technology,the container corner hole is mainly positioned to realize the positioning of the container.The basic flow of container image preprocessing and segmentation is studied based on image processing algorithm.Firstly,the weighted average method is used to convert the color image into gray scale image.Then an adaptive weighted median filter algorithm is used to filter the gray-scale image to remove noise.The algorithm adaptively adjusts the size of filter window and groups each pixel point and assigns corresponding weight.The contradiction between noise suppression and detail protection is solved,and the adaptability of image processing and the ability of detail protection are improved.Then the image is binarized and the threshold is automatically determined by discriminant analysis.This method can not only select the correct threshold value,but also meet the requirements of real-time and automaticity of the system.Finally,Canny operator edge detection algorithm is used for image segmentation.Acquiring edge contour image further simplifies image information and prepares for subsequent image recognition and location.Research on the Basic Flow of Lock Hole Recognition and Location for Container Corner Based on Image Recognition Algorithms.Firstly,the imaging model of monocular camera is analyzed.Then,two edge fitting algorithms,least square method and Hough method,are used to detect and recognize the contour of the preprocessed image.By arranging the area of the connected area to locate the corner lock hole,the minimum inclusion matrix of the connected area of the target locates the center point to realize the identification of the corner lock hole.Finally,combined with the practical application,the relationship between the pixel distance of the image under monocular camera and the actual distance is studied,and the positioning of corner lock hole is realized.At the same time,a comparison is made between the least square method and Hough-based location algorithm for identifying and locating corner keyholes.The differences,advantages and disadvantages between the two algorithms are analyzed in detail through processes and results in different images and environments,and the reasons for the results are explained.Finally,on the basis of the above research,the software implementation of container lock positioning system is studied.Complete an interface program that can demonstrate the algorithm,and show the process of image change in the process of the algorithm.