Product Validation and Verification using Raspberry Pi

In this project, we build a system for identifying and counting beverage cans inside commercial fridges by placing a hidden camera into the fridge. Camera is connected to a Raspberry Pi

located at the top of the fridge. The idea is that if an undesired product is detected or the arrangement does not fit a planogram, Raspberry Pi will notify the irregularity to a server. We

use Bag-of-Words algorithm with SVM classifier. Our training data consists of Coca Cola, Sprite, Nestea, Redbull cans. We test the system with different arrangements of cans in a fridge and

present the results.