Location And Identification Of The Document Information Of Postal Package In High-Speed Sorting

By analyzing the logistics system in the actual field, it’s found out that the automatic scanning equipment in the postal package sorter will trigger signal to take pictures of postal package and discern the barcode on the pictures. But the postal packages corresponding to the indiscernible pictures will be sent to refusal region and disposed of manually. Thus, in order to improve the intellectualization of the logistics system, the further disposal of the postal packages corresponding to the indiscernible pictures--location and identification of barcode on pictures, must be taken into consideration.This paper mainly analyzes the location and identification of document of postal packages in overall treatment process of the system. This exordium gives a simple introduction of the application background and significance and research approach of the preprocessing and location of the documents of postal packages in high-speed sorting, and gives a summary of acquisition process of parcel pictures. Then subdivide entire system into detailed subsystem, according to the system requirements and the characteristics of the pictures of postal packages. And then, analyze and test algorithm associated with specific parcel image preprocessing module, positioning module and barcode image angle adjustment module. Finally formulate the appropriate algorithm corresponding to the characteristics of the postal packages image, and program to achieve.In the preprocessing module, segmentation histogram technology is adopted, instead of the traditional piecewise linear transformation and global histogram equalization. Meanwhile, median filter technology with the direction of information is used to replace the traditional median filtering. This is a highly targeted preprocessing algorithm. In the positioning module, bar code eigenvector are merged to make image matching more accurate, and screen and eliminate inappropriate points after the image matching, the effect of the positioning of the sift algorithm can be improved by this way. In the angle adjustment module, according to the characteristics of the parcel pictures, the combination of mathematical morphology and Hough transform are adopted to replace the traditional Hough transform, which make the module more targeted.The final stage of this paper tests respectively all of the sub modules of the system and gives the corresponding test results. In the end, the author summarizes outcomes and pitfalls of this research and then images the assumption of the subsequent research.