Experiment #207 Artificial Intelligence applied for anomalies detection – Preparation

Data preparation

In order to be able to detect the aforementioned use cases, for example, dirt on solar panels, we must be able to capture a battery of images with the different levels of dirt so that we can label them and show them to our models. Only then will we be able to train them them in a satisfactory way.

In addition to this, we must show them the exact location of the dirt/anomalies:

With this process we will be able to train a series of models for each of the use cases/detections we want to detect.


Once the information has been captured, it will be made available to various services that will be in charge of detecting the presence of anomalies in each image. In addition to customized models to improve and enrich our system, we will use custom vision:


Stay up to date!

Leave a comment