A Book Recognition Method Based on Improved Hu Invariant Moment Algorithm

Abstract

The invention proposes a book recognition method based on the Hu invariant moment improved algorithm. The method includes the following steps: (1) collect the color image of the book and filter it; (2) use the minimum enclosing rectangle algorithm to determine the edge of the book according to the filtered image; (3) process the book image inside the edge of the book into a fixed pixel size, and disassemble it. Divided into multiple ROIs; (4) Separate the channels of each ROI in turn, and separate them into three color channels of red, green, and blue; (5) Use the improved Hu invariant moment algorithm to calculate the The improved invariant moment of the color channel; (6) analyzing and comparing all the improved Hu invariant moment data of each book image for book recognition. Using this method greatly reduces the amount of human operations, and avoids the barcode design and pasting of books in the early stage, effectively saving manpower and material costs, accurate recognition, strong portability, and significantly improved standards. The moment algorithm has the disadvantage of poor robustness when the external light source is color cast or the brightness changes significantly.

Description

technical field [0001] The invention relates to the field of image recognition, in particular to the field of book image recognition, and specifically refers to a book recognition method based on an improved Hu invariant moment algorithm. Background technique [0002] At present, in terms of book identification, sorting and classification in bookstores and libraries, most of them use manual identification and sorting or identification methods using special barcodes and code scanning guns. As far as manual identification and sorting is concerned, the identification and sorting of some large-scale books requires a lot of manpower and time costs, and simple and repetitive work is likely to cause human fatigue and mistakes. method. For the identification method using special barcodes and code scanners, it is necessary to design different barcodes in advance according to different books, and when identifying and sorting, it is necessary to manually find out the position of the b...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of manual identification and sorting or identification methods using special barcodes and code scanning guns that rely too much on people and require a lot of energy in the early stage, the present invention provides a book identification method based on the Hu invariant moment improved algorithm, which greatly reduces It reduces the amount of human operations in the book sorting process, avoids the barcode design and pasting of books in the early stage, and avoids the robustness of the output results of the standard Hu invariant moment algorithm in image recognition when the color cast of the light source or the brightness change is obvious. sexual inadequacy

Images