Camera Calibration for the Surround-View System: a Benchmark and Dataset
The Visual Computer(2024)
Beijing Jiaotong University
Abstract
Surround-view system (SVS) is widely used in the Advanced Driver AssistanceSystem (ADAS). SVS uses four fisheye lenses to monitor real-time scenes aroundthe vehicle. However, accurate intrinsic and extrinsic parameter estimation isrequired for the proper functioning of the system. At present, the intrinsiccalibration can be pipeline by utilizing checkerboard algorithm, whileextrinsic calibration is still immature. Therefore, we proposed a specificcalibration pipeline to estimate extrinsic parameters robustly. This schemetakes a driving sequence of four cameras as input. It firstly utilizes laneline to roughly estimate each camera pose. Considering the environmentalcondition differences in each camera, we separately select strategies from twomethods to accurately estimate the extrinsic parameters. To achieve accurateestimates for both front and rear camera, we proposed a method that mutuallyiterating line detection and pose estimation. As for bilateral camera, weiteratively adjust the camera pose and position by minimizing texture and edgeerror between ground projections of adjacent cameras. After estimating theextrinsic parameters, the surround-view image can be synthesized byhomography-based transformation. The proposed pipeline can robustly estimatethe four SVS camera extrinsic parameters in real driving environments. Inaddition, to evaluate the proposed scheme, we build a surround-view fisheyedataset, which contains 40 videos with 32,000 frames, acquired from differentreal traffic scenarios. All the frames in each video are manually labeled withlane annotation, with its GT extrinsic parameters. Moreover, this surround-viewdataset could be used by other researchers to evaluate their performance. Thedataset will be available soon.
MoreTranslated text
Key words
Surround-view system,Advanced driver assistance system (ADAS),Automatic extrinsic calibration
PDF
View via Publisher
AI Read Science
Must-Reading Tree
Example

Generate MRT to find the research sequence of this paper
Data Disclaimer
The page data are from open Internet sources, cooperative publishers and automatic analysis results through AI technology. We do not make any commitments and guarantees for the validity, accuracy, correctness, reliability, completeness and timeliness of the page data. If you have any questions, please contact us by email: report@aminer.cn
Chat Paper
Summary is being generated by the instructions you defined