Electronic device for acquiring images of sardine cans

Abstract

In the field of artificial intelligence we will often have the problem of having good data to create an efficient model. Especially if the working environment does not help us to acquire data because of parasites (noise, speed, vibration, etc.), as is the case with canning factories. For this, data collection remains a challenge for developers of machine learning systems. In our case, we are going to create an electronic module connected to the Internet, to install it in the production line of a Sardine canning factory. In order to capture the images of the cans and send them to a server in the cloud. We run the learning machine in the server to ensure the speed of training. In this article, we will discuss the problems encountered in order to propose the solutions to acquire the images of the cans.