Industrial accidents are a main issue for companies with capital intensive infrastructure; one way to reduce this is avoiding the employees exposure to hazardous environments that affect their physical integrity and health. Inspection of gauges and manipulation of valves are important tasks that are typically performed in all industries, these industrial components are located in hard to access and dangerous places, for this reason it is necessary a system that can achieve this task without direct human intervention. In this report, the preliminary results of a teleoperated robotic system with computer vision capabilities is presented. In addition, a vision algorithm for analog gauge measurements was tested in different gauge models and the results show that a 1% error was achieved in less than 1 second.