Single and multi-frame motion deblurring for legged robots: characterization using a novel fd-aroc performance metric and a comprehensive motion-blur dataset

Download
2016
Gültekin, Gökhan Koray
Dexterous legged robots are agile platforms that can move on variable terrain at high speeds. The locomotion of these legged platforms causes oscillations of the robot body which become more severe depending on the surface and locomotion speed. Camera sensors mounted on such platforms experience the same disturbances, hence resulting in motion blur. This is a corruption of the image and results in loss of information which in turn causes degradation or loss of important image features. Most of the studies in the literature and the proposed performance metrics focus mainly on the visual quality of motion blurred images and its improvement. However, from the perspective of computer vision algorithms, feature detection performance is an essential factor that determines their performance. The aim of this study is to analyze and evaluate motion blur on a legged robot and the deblurring methods with a focus on feature detection. We propose a multi-frame motion deblurring method utilizing the variable motion blur in consecutive image frames captured from the camera on a legged mobile robot. For a comparison of blurred and deblurred images, we define a novel performance metric based on the feature detection accuracy. Noting that a suitable data set to evaluate the effects of motion blur and its compensation for legged platforms is lacking in the literature, we develop a comprehensive multi-sensor data set for that purpose. The data set consists of monocular image sequences collected in synchronization with a low cost MEMS gyroscope, an accurate fiber optic gyroscope and an externally measured ground truth motion data. We make use of this data set for an extensive benchmarking of prominent motion deblurring methods from the literature in terms of the proposed feature based metric. 
Citation Formats
G. K. Gültekin, “Single and multi-frame motion deblurring for legged robots: characterization using a novel fd-aroc performance metric and a comprehensive motion-blur dataset,” Ph.D. - Doctoral Program, Middle East Technical University, 2016.