Vision based attitude and altitude estimation for UAVs in dark environments

International audienceThis paper presents our novel approach to estimate the real time pose and altitude for unmanned aerial vehicle UAV under low light environment, such as on a cloudy day and night or indoor dark environment. This method utilises the maximum out of a calibrated projective camera and a laser pattern projector. The model is mathematically proven with the validation on simulation and later implemented for a real set of images. The estimated values from our method are compared with commercial sensors for their accuracy and correctness. The proposed solution with the system indicates it being suitable to use in dark environmental conditions especially in the night or dark indoor environment in the absence of GPS signals for positioning. The results prove its suitability for autonomous take off and landing, low altitude maneuver in the dark environment and provides room for additional payload due to use of lighter single camera system