Summary
This project demonstrates a method for solving electrical dc circuits from an image. Methods of computer vision are used throughout the programm.
The programm is at an initial stage and the only circuit images, which the programm can solve, are the ones in the directory ./images
However, this is a first demonstration of the way that a programm can analyse a circuit from an image. Imagine if you only could solve an electrical circuit by taking a photo with your smartphone, it would make the life of the electrical engineering student much easier, especially whenever he writes an exam on circuit analysis.
Structure
The directories are based on the stage of computer vision analysis. The main directories are listed bellow:
- recognize contains methods to recognize the patterns on the image
- ocr_manipulations is used to recognize characters on the image
- nodes_and_centers finds the topology, this is the way elements are connected
- node_analysis contains all the methods to solve the circuit
How to run it
Run the programm from the PROJECT.m script, change the variable cir_number to choose which circuit (1 to 8 because we have 8 circuit images) you want to solve. The programm can solve only the 5 of the 8 images (the 1, 2, 3, 5 and 6), this is 62.5% success on circuit solving.