Research On The Algorithm Of Charactor Recognition For Automobile Instrument Panel Detection System

With the development of information technology,the functions of automobile dashboards are no longer just displaying basic information such as vehicle speed,speed,fuel quantity,etc.The current dashboards can express a huge amount of information.At the same time,the more complicated the system,the greater the possibility of error and the higher the difficulty of detecting errors,so it is necessary to design a reasonable recognition and detection system.This article uses digital image processing technology to identify the digital character display part on the car dashboard,covering the image processing algorithm used in each link of image processing,specific theoretical methods and code implementation.Two character recognition methods,template matching method and BP neural network,were studied and compared respectively.First of all,in image preprocessing,the 24-bit true-color image is converted into an8-bit gray image by gray-scale transformation,and noise is removed by median filtering.The binary effect of the three binarization methods is compared,and the maximum between-class variance method is used for image binarization.Finally,the morphological processing of the open operation is used to eliminate the burrs of the characters and fill the small holes.Secondly,in the character recognition,the projection method of statistical gray values in the horizontal and vertical directions is used to distinguish the character area and the character interval,and an improvement is made in combination with the actual application.In the character normalization stage,a uniform size is set for the character picture,and a bilinear interpolation algorithm is used.In order to prevent character deformation,a proportional template is used to fill the background during scaling to produce a uniform template character.Finally,in the character recognition,the digital tube digital recognition adopts the threading method to recognize the digits first,and the display digits are recognized according to the structural characteristics of the digital tube digits;two characters are studied in the liquid crystal digital character recognition Recognition algorithm,one is the template matching method,extract each pixel as a feature value,establish a unified template library,use the algorithm to calculate similarity and dissimilarity for matching,and the other is BP neural network method The template training set is used to train the three-layer neural network,and the method of variable learning rate is added to improve the traditional BP neural network to achieve a good recognition effect.At the same time,a reusable C ++ image processing function was designed for the project.The several algorithms implemented in this paper have achieved a high recognitionrate,fast recognition speed,certain anti-noise ability,and high feasibility and effectiveness.The threading method recognition is very fast for the digital recognition of the digital tube,and the accuracy is sufficient to meet the requirements;the template matching method is simple to calculate,and the recognition speed is greatly affected by the template library,but designing a template library of suitable size can meet the speed requirements;The method storage is small,and the recognition accuracy is higher after the training meets the requirements,but the training process is relatively complicated.