Semantic Segmentation for Road Surface Detection in Snowy Environment

Vachmanus, Sirawich; Ravankar, Ankit A.; Emaru, Takanori; Kobayashi, Yukinori

doi:10.23919/sice48898.2020.9240402

Cited by 14 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the lack of data, there have been efforts to develop drivable area detection in winter conditions with supervised learning [35], [36], [41]- [43]. However, caused by the lack of data, the training distribution is narrow, while winter driving conditions have high variation because the road network is very extensive and snow can cause the appearance of the drivable area to change rapidly.…”

Section: A Supervised Approachesmentioning

confidence: 99%

TADAP: Trajectory-Aided Drivable area Auto-labeling with Pretrained self-supervised features in winter driving conditions

Alamikkotervo,

Ojala,

Seppänen

et al. 2024

IEEE Trans. Intell. Veh.

View full text Add to dashboard Cite

Detection of the drivable area in all conditions is crucial for autonomous driving and advanced driver assistance systems. However, the amount of labeled data in adverse driving conditions is limited, especially in winter, and supervised methods generalize poorly to conditions outside the training distribution. For easy adaption to all conditions, the need for human annotation should be removed from the learning process. In this paper, Trajectory-Aided Drivable area Auto-labeling with Pretrained self-supervised features (TADAP) is presented for automated annotation of the drivable area in winter driving conditions. A sample of the drivable area is extracted based on the trajectory estimate from the global navigation satellite system. Similarity with the sample area is determined based on pre-trained selfsupervised visual features. Image areas similar to the sample area are considered to be drivable. These TADAP labels were evaluated with a novel winter-driving dataset, collected in varying driving scenes. A prediction model trained with the TADAP labels achieved a +9.6 improvement in intersection over union compared to the previous state-of-the-art of self-supervised drivable area detection.

show abstract

Section: A Supervised Approachesmentioning

confidence: 99%

TADAP: Trajectory-Aided Drivable area Auto-labeling with Pretrained self-supervised features in winter driving conditions

Alamikkotervo,

Ojala,

Seppänen

et al. 2024

IEEE Trans. Intell. Veh.

View full text Add to dashboard Cite

show abstract

“…With the development of fully convolutional networks for pixel-wise predictions [7], deep learning has resulted in neural networks that are capable of highly accurate dense semantic label predictions on elementary datasets [2], [8], [9], [10], [11], at least in ideal conditions. Achieving this level of accuracy in challenging domains, such as nighttime [4], [5] or snow [12], [13], remains a mostly unsolved research problem. Of these approaches, [4] is most conceptually similar to ours, except that our approach does not require any prior experience in the specific location.…”

Section: A Semantic Segmentationmentioning

confidence: 99%

Improving Road Segmentation in Challenging Domains Using Similar Place Priors

Malone

Garg

Mei

et al. 2022

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Road segmentation in challenging domains, such as night, snow or rain, is a difficult task. Most current approaches boost performance using fine-tuning, domain adaptation, style transfer, or by referencing previously acquired imagery. These approaches share one or more of three significant limitations: a reliance on large amounts of annotated training data that can be costly to obtain, both anticipation of and training data from the type of environmental conditions expected at inference time, and/or imagery captured from a previous visit to the location. In this research, we remove these restrictions by improving road segmentation based on similar places. We use Visual Place Recognition (VPR) to find similar but geographically distinct places, and fuse segmentations for query images and these similar place priors using a Bayesian approach and novel segmentation quality metric. Ablation studies show the need to re-evaluate notions of VPR utility for this task. We demonstrate the system achieving state-of-theart road segmentation performance across multiple challenging condition scenarios including night time and snow, without requiring any prior training or previous access to the same geographical locations. Furthermore, we show that this method is network agnostic, improves multiple baseline techniques and is competitive against methods specialised for road prediction.

show abstract

“…The combination of the published dataset and their snowy dataset can improve the result of segmentation. Our previous work [39] tried to improve the training process for segmentation in a snowy environment. The network is based on DeeplabV3 with extracted auxiliary outputs to enhance the training performance.…”

Section: Related Workmentioning

confidence: 99%