Intelligent Quarry Production Monitoring Risks and Quality by Artificial Intelligence

The presented work discusses the results of an industry project oriented on the development of a web platform aimed at monitoring the quality of the quarry production processes and related risks. The platform is oriented mainly on the predictive maintenance of quarry crusher and conveyor belts. Artificial Intelligence (AI) algorithms such as Multilayer Perceptron (MLP), Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM) are applied for brakeage prediction of crushers and the conveyor belt carpet. The dataset is obtained thanks to the several cameras and accelerometers installed in different check points of the quarry, sending data in real time to the platform. The system maps also risk conditions based on the combined analysis of the 3D quarry morphological reconstruction performed by photogrammetry and of the radargrams acquired by an Unmanned Aerial Vehicle (UAV) equipped with a Ground Penetrating Radar (GPR) antenna.