Abstract
Abstract. Environment-observing vehicle camera self-calibration using a structure from motion (SfM) algorithm allows calibration over vehicle lifetime without the need of special calibration objects being present in the calibration images. Scene-specific problems with feature-based correspondence search and reconstruction during the SfM pipeline might be caused by critical objects like moving objects, poor-texture objects or reflecting objects and might have negative influence on camera calibration. In this contribution, a method to use semantic road scene knowledge by means of semantic masks for a semantic-guided SfM algorithm is proposed to make the calibration more robust. Semantic masks are used to exclude image parts showing critical objects from feature extraction, whereby semantic knowledge is obtained by semantic segmentation of the road scene images. The proposed method is tested with an image sequence recorded in a suburban road scene. It has been shown that semantic guidance leads to smaller deviations of the estimated interior orientation and distortion parameters from reference values obtained by test field calibration compared to a standard SfM algorithm.
Highlights
Environment perception is one of the key enablers of advanced driver assistance systems in vehicles and especially on the way to autonomous driving
These scene-specific problems with correspondence search and reconstruction could subsequently have negative influence on vehicle camera self-calibration. As these problems can be related to certain groups of objects, a priori knowledge about object groups shown in road scene images could be used to avoid negative influences on camera calibration
A new method for on-board vehicle camera self-calibration by a semantic-guided structure from motion algorithm, whereby semantic knowledge about objects being present in images is applied to create semantic masks disabling feature extraction in image parts showing critical objects like moving objects in order to make calibration more robust
Summary
Environment perception is one of the key enablers of advanced driver assistance systems in vehicles and especially on the way to autonomous driving. As third group, moving objects like pedestrians or other vehicles or trees in the wind could cause a bad reconstruction as 2D image points of the same feature in different images might not be related to the same 3D road scene point even for correct matches. These scene-specific problems with correspondence search and reconstruction could subsequently have negative influence on vehicle camera self-calibration. An analysis of the effect of different semantic classes used for creating semantic masks on the structure from motion algorithm and on the consecutive global bundle adjustment
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
