Machine Learning model applied to climbing for Lizcore

This project aims to develop a proof-of-concept based on machine learning techniques to identify, with the highest possible accuracy, hazardous situations before a top-rope climb begins. The solution aims to provide an aid tool that contributes to increasing the safety of climbing facilities by the preventative detection of potential hazardous situations.

To achieve this objective, computer vision models based on convolutional neural networks (CNNs) are used, capable of identifying, classifying and localising objects from the images captured by the system. These techniques allow the scene to be automatically analysed before the start of the activity and detect potential incidents related to the equipment or the climbers’ preparation.

The solution developed by inLab contributes to providing a support tool that increases the safety of climbing facilities through the preventative detection of potential hazardous situations.

You can see the system in action in this video: https://www.instagram.com/reel/DImhrPdypEf/