Text detection in Nature Scene Images

Abstract

Text detection in natural scenes holds great importance in the field of research and still remains a challenge because of size, various fonts, line orientation, different illumination conditions, weak character and complex background in image. Here a nature Text detection from nature images can be processed through the Raspberry Pi device. Raspberry Pi device is one of the controller and it acts like CPU and the image processing code ca be done to this device which runs the code. In our proposed method, input image is captured through 2 ways: one through the Pi cam (real time image) and other through the system(dataset). The images maybe of different font and colour, different slant degree and are of different format like jpeg, jpg, png, etc. All kind of images are accepted as input. Both the images are trained and supervised and hence training and testing images are same. These processes completely rely on Raspberry Pi. The given input images boundary is identified and is converted from RGB format to gray scale image for further processing. A gray scale image is obtained then text regions are obtained based on the morphological features. The clustering the characters is performed. Finally forming of words and output is processed data displayed on terminal. All the process run through the terminal and output is also displayed on terminal.