Character Recognition Based On Embedded Linux

Optical character recognition belongs to the category of pattern recognition and artificial intelligence, related to pattern recognition, image processing, artificial intelligence, fuzzy mathematics, information theory and other technology, is a comprehensive discipline between theoretical and applied research. With the popularity and social needs of portable mobile terminal devices, ARM platform has become a hot in the current hardware platforms. So the study of implement of character recognition on ARM-Linux platform becomes a crossover hotspot. There are many development tools in the field of image processing. OpenCV has been widely used, because of its efficient, open source, cross-platform and other features. The paper takes the car license plate recognition as an example and illustrates character recognition implemented on an embedded Linux platform by using the C programming language and calling OpenCV library function.In the paper, license plate recognition system is divided into three modules: license plate location module, character segmentation module and license plate recognition module.The license plate location module uses edge detection and morphological processing algorithm to extract the license plate from the picture. The character segmentation module uses improved projection algorithm to split out individual characters from the license plate area. The character recognition module uses template matching algorithm to identify a single character.On this basis, the paper also describes the construction of embedded platforms, including how to build a cross-compilation environment and migration of the Linux operating system, and explains the steps of transplanting OpenCV library to embedded platforms.