Performance Assessment of Predictive Lane Boundary Detection for Non-uniformly Illuminated Roadway Driving Assistance
| Autor: | H. Bryan Riley, Avishek Parajuli, Mehmet Celenk |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
050210 logistics & transportation
Matching (statistics) Accuracy and precision business.industry Machine vision Computer science 05 social sciences ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Glare (vision) Linear prediction 02 engineering and technology Salient 0502 economics and business Principal component analysis 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Computer vision Artificial intelligence business Focus (optics) ComputingMethodologies_COMPUTERGRAPHICS |
| Zdroj: | CRV |
| DOI: | 10.1109/crv.2016.52 |
| Popis: | In this paper, we focus on the performance of an earlier developed method for the detection of road lane markers which are minimally affected by non-uniform surface illumination (e.g., shadows and highlights) and road geometries. We investigate the detection performance and associated parameters of this approach. Experimental results show that the method yield accurate results in various situations such as broken or missing lane markings, the curved lane, the heavy shadows, the sun glare, and the occlusion of other vehicles. The accuracy and precision of the proposed model are presented and compared with existing methods as reported in the current literature. Specifically error curves are computed and presented for the actual verses predicted lane markers by blending salient features from the Gradient Spectrum Matching (GSM) and Principal Component Analysis (PCA) methods. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|