Research Of Text Image Binarization And Tilt Correction Methods Based On The LabVIEW

With the rapid development of computer multimedia technology,in recent years,as an important application of computer intelligence,optical character identification technology is developing very fast.Due to the influence of the shooting equipment and the shooting environment,the collected images must be pre-processed before recognition,including image enhancement,image segmentation and tilt correction,etc,which plays a vital role in the later identification.This paper first discusses the algorithm of image segmentation and tilt correction at home and abroad,and analyzes the shortcomings of these algorithms based on the simulation results.On this basis,it focuses on the Otsu algorithm of global threshold segmentation and the Sauvola algorithm of local threshold segmentation,and analyzes the shortcomings of Otsu algorithm and Sauvola algorithm and improvements.As for Otsu algorithm,one of the variance is large in the target and background variance,Otsu method threshold is high,resulting in poor segmentation problem,it proposes an improved Otsu algorithm.The improved algorithm is based on the proportion of the target and the background of the image,using the ratio of the target to the whole image and combining the intraclass variance to improve the threshold in the Otsu algorithm to select formula,the experiment result shows,we can effectively solve the problem that the segmentation effect is poor due to the high threshold when the background variance is large.Aiming at the problem of Sauvola algorithm,this paper proposes an improved Sauvola algorithm.When the local threshold is calculated,the improved algorithm replaces the pixel neighborhood standard deviation in the Sauvola algorithm by using the difference between the gray neighborhood mean of the pixel and the gray value of the central pixel.It uses the difference between the maximum of gray value of the pixel neighborhood and the minimum value to replace the standard deviation coefficient in the Sauvola algorithm,the experiment result shows it reduces the run time of the algorithm and improves the efficiency of segmentation when the segmentation effect is better.In this paper,an oblique correction algorithm based on the least squares method is proposed to solve the problem that detection time of the tilt angle of the image is too large for Hough transform and projection method.In this paper,the method isbased on the characteristics of the direction of the edge of the image to quickly detect the direction of the baseline angle,and then calculate the tilt of page,and finally do the rotation to correct the tilt image.Finally,it uses LabVIEW Vision visual development package to realize the related algorithm and the improved algorithm on the LabVIEW platform,and illustrates the effectiveness of the improved method in this paper combined with the experimental results.