Sistema de reconhecimento de mapas de risco utilizando processamento de imagens.

Abstract

Inserting people with disabilities in the workplace require companies to prepare the environment taking into account a series of adaptive measures. In order to help the risk recognition process inherent in workplaces one could use the Risk Map, a document conceived by the Commission of Internal Accident Prevention (CIPA) that identifies the location and existence of physical, chemical, ergonomic, and biological hazards. It is often inaccessible for visual impaired personnel because it uses images to represent those risks. Our work aims to enhance accessibility in the workplace for people with disabilities aiming to preserve their physical integrity where we evaluate the tradeoffs of processing the risk maps and assess how this process would be conducted. So, we have developed a computational system that uses digital image processing techniques to read, process, and transform risk maps into a textual representation. The algorithm takes as input the risk map image, identifies and isolates the existing three ROIs (Region of Interest), analyses the color legends and intensities, locates objects within the layout, extracts and classifies attributes and then performs the conversion to its textual counterpart. Experimental results indicate that it is possible to computationally process visual risk maps by inserting markers in key areas. The process of converting visual characteristics of the risk map in other format allows accessibility for people with limited interpretation capabilities. The present work is a considerable step towards the development of an application that could use the information extracted from the risk maps to signal routes and present the associated risks in the path.