Wide-range license plate inclination angle rapid calculation method

Abstract

The invention discloses a wide-range license plate inclination angle rapid calculation method. The method comprises the following steps that an initial parameter of an angle parameter a is set, an image center is rotated through leftward and rightward deviation for x degrees respectively with the a acting as the center, horizontal projection after rotation is calculated so that projection data are obtained, variance of the projection data is respectively calculated and the variance data are analyzed; if an inflection point appears in the data, the angle parameter a in which the inflection point appears is a solved license plate image inclination angle; if the data are a decreasing curve, forward iterative calculation is performed on the angle parameter a until the inflection point appears in the data; if the data is an increasing curve, backward iterative calculation is performed on the angle parameter a until the inflection point appears in the data; and the license plate image inclination angle obtained each time is recorded, and the maximum value is taken to update the calculated center angle parameter a so as to output the image angle. The problem of solving the maximum value by performing traversal calculation on all the angles within a wide range in conventional solving of the license plate large angle can be overcome and calculation amount can be reduced.

Description

technical field [0001] The invention belongs to the field of intelligent transportation, and relates to a method for quickly calculating the inclination angle of a large-scale license plate. Background technique [0002] In the field of intelligent transportation, with the rapid development of semiconductor technology and the continuous emergence of new calculation methods, the current intelligent transportation algorithms are gradually moved from the back-end PC equipment to the front-end. The greater the amount of highly analyzable data, the more types of integrated algorithms will greatly improve the cost performance of the product. License plate recognition is the core of intelligent applications, and fast and accurate calculation of the tilt angle of the license plate plays a vital role in improving the overall performance of the license plate recognition algorithm. [0003] The traditional calculation method is as follows: [0004] 1. Rotate the center of the image a...

AI Technical Summary

Problems solved by technology

[0012] The disadvantage of this method is that it is necessary to traverse the parameter a to find the maximum value. Assuming that the range of the parameter a is between -30 degrees and +30 degrees, traversing with an accuracy of 1 degree requires 60 times of the above calculation

The disadvantage of this is that regardless of whether the image is tilted or not, the traversal calculation will be performed according to the parameters, which greatly reduces the system efficiency

Images