Design And Implementation Of Pre-Processing Algorithm For Part Code Recognition Based On Android Platform

In recent years, people expect to experience more power and versatile functions on mobile phones as the rapid growing of the hardware capacity. Therefore image processing and character recognition technology based on Android have become a new application filed. However, the image processing capability of Android system is too limited to reach a satisfied pre-processing effect that traditional computers can get. As a consequence, it is hard for Android to achieve a better result in character recognition.This paper is focused on the design and implementation of pre-processing algorithm in part code recognition based on Android. Based on the basis of the analysis to raditional image processing algorithms, this paper designed an appropriate preprocessing algorithm for character image recognition with complex background. The algorithm consists of three major parts:preliminary processing, character segmentation and skeleton extraction. Firstly, the background noises in the image could be removed and the image edge will become more enhanced in preliminary processing; Secondly, complete morphological information of a single character could be extracted after character segmentation; And finally, the skeleton of a character could be extracted. In addition, the shape and topology of the character could be well preserved in this procedureIn this paper, by employing OpenCV and Android NDK packages, a satisfied result of pre-processing for character recognition on the Android platform has been achieved. Testing results show that the proposed algorithm could not only remove the background noise in the image, but also preserve the complete morphological information of the character, facilitating further character segmentation and skeleton extraction processes. With the support of our proposed pre-processing algorithm, the average recognition rate in the character recognition system could reach 90% or more.