DSO-Direct Sparse Odometry for Robot Vision & Vision applications
Today we are going to discuss a brief concept about how to deal with he Odometry values and in the robotics industry obtaining proper visual odometry values are being a big issue in order to reach the desired goal which could be done only when a machine obtains a proper odometrical data's.So,let's discuss in brief about a unique concept in Visual odometry which is Direct Sparse odometry which is to obtain a smooth visual odometry value.
Direct visual odometry and SLAM algorithms have demonstrated impressive levels of precision. However, they require a photometric camera calibration in order to achieve competitive results. Hence, the respective algorithm cannot be directly applied to an off-the-shelf-camera or to a video sequence acquired with an unknown camera
Recently a number of direct methods for visual odomertry and visual SLAM such as DSO or LSD-SLAM have been proposed, working only on pixel intensities.They all rely on the underlying assumption, that a scene point appears with constant brightness values across multiple images. However, when taking images with auto exposure video cameras, this assumption typically does not hold. The automatic adjustment of the exposure times, the photometric falloff of the pixel intensities to the sides of the image due to vignetting as well as an often nonlinear camera response function cause the observed pixel intensities to differ for the same scene point. It has been shown that prior photometric camera calibration can significantly enhance the performance of DSO
DSO is a novel direct and sparse formulation for Visual Odometry. It combines a fully direct probabilistic model (minimizing a photometric error) with consistent, joint optimization of all model parameters, including geometry - represented as inverse depth in a reference frame - and camera motion. This is achieved in real time by omitting the smoothness prior used in other direct methods and instead sampling pixels evenly throughout the images
DSO does not depend on keypoint detectors or descriptors, thus it can naturally sample pixels from across all image regions that have intensity gradient, including edges or smooth intensity variations on mostly white walls. The proposed model integrates a full photometric calibration, accounting for exposure time, lens vignetting, and non-linear response functions. We thoroughly evaluate our method on three different datasets comprising several hours of video. The experiments show that the presented approach significantly outperforms state-of-the-art direct and indirect methods in a variety of real-world settings, both in terms of tracking accuracy and robustness.
The applications of DSO are plenty especially in Robotics,Drones,SLAM research,Comuter vision applications,etc...