AuroraPi uses a Raspberry Pi with a NoIR camera (for better low light imaging) in order to capture the image of the text in a document and uploads it to the google cloud where the vision (a machine learning based text detection API) recognizes the text, send the recognized text (En & FR ) to a file with different format (.text, .pdf, .docx … ) and store it
As only 60% of knowledge is available on the internet, the ultimate goal of this project is to make this technology available for everyone to digitize documents and make this digitized data accessible online in soft copies.
The project consist of two aspect, Electronic Engineering and Mechanical Engineering.
- In the first aspect, we would like to use Raspberry Pi zero instead of Raspberry Pi 3. Also implement the white and IR LED strips around the camera which gives it better focus and sharp images of the document.
- In the second aspect, we would like to design a frame to have an Auto page turning scanner. This frame would contains a plug where the AuroraPi could be detachable.
The project has been tested on several images and is successful in carrying out text detection and get the output in a .txt format.