Polarized Vision Navigation For UAVs Guidance In Non-GPS Environments

The drive towards utilizing small,cheap,Unmanned Aerial Vehicles（UAV）for military operations means that navigation systems that are robust in the GPS denied environments for both military and civilian operations must be employed.Vision based navigation is a satisfactory solution in terms of the weight,cost,and information it can provide.The visual Simultaneous Localization and Mapping（SLAM）is so important in the last years.The Visual navigation model that can be used in GPS-denied Environments is the main goal of the Ph.D.thesis.Outdoor scenes are usually affected by air particles and water vapor in the atmosphere.Haze is happened because of absorption and scattering in the atmosphere.Haze affects the visibility and causes a degradation in the image quality of outdoor scenes,which in turn affects visual navigation.This thesis highlights the importance of the polarization information for obtaining accurate skylight polarization parameters and restoring a high-quality image to be used in the calculation of the visual navigation measurements.This work inspects the difficulties in obtaining polarization parameters and reduce the aerosols particles effect in skylight-polarization navigation in various atmospheric conditions.A new comprehensive method based on skylight color mixing is presented to get the highly dependence effect of light polarization.In cloudy and hazy conditions,when skylight passes through the atmospheric particles causing a big distortion in polarization information leading to a very low or zero information of the sky neutral points.The experimental results show that this method improves the polarization patterns that used in obtaining accurate navigation parameters by eliminating the effect of noise in cloudy and hazy days.The Instant Dehaze method used polarized images to obtain a dehazed image and an estimated depth map of the scene.This estimated depth is misrepresented due to high degree of polarization and scene’s objects directly illuminated by the sun.In this thesis,a polarization guided auto-regressive model for depth recovery is presented.This proposed method restores the estimated depth map by incorporating polarized data to an adaptive auto-regressive（AR）model.First,a 90-degree polarized image is used in our polarization term of AR coefficient,then the Stokes vector component S₁ is used in our polarization guided depth map in the depth term of AR coefficient.The experimental results show that our method outperforms existing state-of-the-art schemes and improves the conventional polarization dehazing method.The scale invariant feature transform algorithm（SIFT）is applied to get invariant features of the scene images.But it is not sufficient in haze conditions.More edge information is required to enhance the SIFT matching process.Utilizing the polarization information expressed by the Stokes vector component S₁ with its edge information can improve the output of the keypoint localization in the matching process.In this thesis,a novel fusion method called Polarized Intensity-Hue-Saturation（Pol-IHS）is proposed that uses polarization and depth information by fusion of a polarized haze-removed image with the estimated depth and applying S₁.The Instant Dehazing method uses polarized images to obtain a haze-removed image and its estimated depth map.An optimization energy functional is used for adding spatial details,preserves the image quality,and guarantees the smoothness.The fused image has high spatial details required for enhancing the matching process.The experimental results show that the proposed method outperforms the existing conventional SIFT scheme and improves the conventional SIFT matching method.A polarized visual SLAM is proposed in our thesis.It combines the estimated recovered depth map,the proposed polarized matching,and then utilizing the accurate solar meridian to get the vehicle heading angle.These parameters will be integrated together to get an accurate navigation information for the vehicle in GPS-denied environments,which can be used for both military and civilian application at hazy and cloudy conditions.