Angle Detection Algorithm Based on Triangle Similarity Theorem

Abstract

The invention discloses an angle detection algorithm based on similar triangles, comprising the following steps: Step 1, image binarization; Step 2, performing expansion operation on the generated binarized image to highlight edge curves; Step 3, binarizing the binary image Project the image from top to bottom and from bottom to top to obtain the upper and lower edge curves of the image respectively, and obtain the effective point start address and end address of the first effective line of the upper and lower edge curves respectively; step 4, combine the upper and lower edge curves Form a complete rectangular edge curve for processing, respectively select a line segment with the same abscissa coordinates on the left side of the starting address of the upper and lower edge curves of the image; determine whether the angle between the two curves is a right angle, and whether the selected two curves are located at the edge of the image An effective curve on the curve; step five, select a line segment that meets the requirements, use the fast reciprocal square root algorithm and the Pythagorean theorem to calculate the reciprocal of the hypotenuse, and then obtain the sine and cosine values ​​of the inclination angle.

Description

technical field [0001] The invention belongs to the field of angle detection, in particular to an angle detection algorithm based on the triangle similarity theorem. Background technique [0002] In today's life, ID cards, business cards and other documents are often used for copying and faxing. Due to human operation, many scanned images will appear tilted, and these tilts will affect the layout analysis of business cards, etc. , character recognition and copying will cause adverse effects, so it is necessary to detect the tilt angle of the tilted image, and then perform image correction according to the tilt angle. [0003] In practical applications, most of the existing image tilt angle detection technologies first filter and binarize the input image to obtain a binarized image, clearly reflect the edge of the image, and then pass Hough Algorithms such as operator, Sobel operator, and least square method analyze the edge curve of the image, and finally obtain the image t...

AI Technical Summary

Problems solved by technology

[0006] Although the above two methods have high detection accuracy, they have a large amount of calculation and high memory usage, and have certain requirements for the configuration of the operating platform. For some operating platforms with low configuration, the above methods cannot be fully applied

Images