Road networks as collections of minimum cost paths

Wegner, Jan Dirk; Montoya-Zegarra, Javier A.; Schindler, Konrad

doi:10.1016/j.isprsjprs.2015.07.002

Cited by 83 publications

(94 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, modern approaches use combination of rules that support each other on their way towards a reliable road pixel extraction. Thinkable rules are nDSM, Normal Difference Vegetation Index (NDVI), color information from different color spaces , lines detected in images (Ünsalan and Sirmacek, 2012), stripes detected in nDSM for representing street canyons (Hinz, 2004), surface roughness (Hu et al, 2014), spatial signature measure employed by Jin and Davis (2005), filters, descriptors and textons employed by Montoya-Zegarra et al (2015); Poullis and You (2010); Wegner et al (2015), methods based on morphological profiles (Valero et al, 2010) up to the approaches based on Convolutional Neural Networks (Sherrah, 2016), where the problem is solved by applying very large numbers of features (neurons) and huge amounts of training data.…”

Section: Previous Workmentioning

confidence: 99%

“…Bottleneck: Variable selection Clearly, the main focus of the contribution of Wegner et al (2015) lay on minimum cost paths. However, we feel that the step of preliminary classification did not receive the attention and care its importance deserves.…”

Section: Previous Workmentioning

confidence: 99%

“…Redundant features are known to have only marginal detrimental effect on random forests (Genuer et al, 2010), but a dramatic degradation can result for other classifiers, such as logistic regression (Cox, 1958). As a consequence, any method, including the approach of Wegner et al (2015), could benefit greatly -both with respect to the performance and the computing time -from a careful selection of variables (features).…”

Section: Previous Workmentioning

confidence: 99%

“…Contribution In this work, we adopted from the superpixelwise classification framework and the features generated from filter banks Wegner et al (2015); that is, we omitted the postprocessing steps in order not to gloss over the raw classification results. However, we extend the variable set of Wegner et al (2015) by our variables and performed classification with two methods.…”

Section: Previous Workmentioning

confidence: 99%

“…

ABSTRACT:For extraction of road pixels from combined image and elevation data, Wegner et al (2015) proposed classification of superpixels into road and non-road, after which a refinement of the classification results using minimum cost paths and non-local optimization methods took place. We believed that the variable set used for classification was to a certain extent suboptimal, because many variables were redundant while several features known as useful in Photogrammetry and Remote Sensing are missed.

…”

mentioning

confidence: 99%

See 4 more Smart Citations

Variable Selection for Road Segmentation in Aerial Images

Warnke

Bulatov

2017

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:For extraction of road pixels from combined image and elevation data, Wegner et al. (2015) proposed classification of superpixels into road and non-road, after which a refinement of the classification results using minimum cost paths and non-local optimization methods took place. We believed that the variable set used for classification was to a certain extent suboptimal, because many variables were redundant while several features known as useful in Photogrammetry and Remote Sensing are missed. This motivated us to implement a variable selection approach which builds a model for classification using portions of training data and subsets of features, evaluates this model, updates the feature set, and terminates when a stopping criterion is satisfied. The choice of classifier is flexible; however, we tested the approach with Logistic Regression and Random Forests, and taylored the evaluation module to the chosen classifier. To guarantee a fair comparison, we kept the segment-based approach and most of the variables from the related work, but we extended them by additional, mostly higher-level features. Applying these superior features, removing the redundant ones, as well as using more accurately acquired 3D data allowed to keep stable or even to reduce the misclassification error in a challenging dataset.

show abstract

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

“…

…”

mentioning

confidence: 99%

See 3 more Smart Citations

Variable Selection for Road Segmentation in Aerial Images

Warnke

Bulatov

2017

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

Assessing an ecological security network for a rapid urbanization region in Eastern China

Zhou

Lin

et al. 2021

Land Degrad Dev

View full text Add to dashboard Cite

Rapid urbanization leads to fragmentation of landscape and destruction of the ecological security network. A sound ecological security network promotes connectivity between ecological sources and mitigates the degradation of an ecological system. However, when identifying the ecological sources of the ecological network, studies available in the current literature did not simultaneously consider the structures, functions, and values of ecological patches. Our study aimed to fill this research gap, and our objective was to improve the framework for assessing the ecological security network. We demonstrated the application of the proposed framework through a case-study of the urban agglomeration around Hangzhou Bay (UAHB), a rapid urbanization region in Eastern China. We improved the identification method of ecological sources by integrating the evaluations of ecosystem services value and ecological sensitivity, while we screened ecological sources by using the rank-size rule and the natural breaks method. Based on the screened ecological sources, the ecological corridors were reconstructed and optimized for the UAHB region. The results from this study showed that ecological network has actually improved in the UAHB in the last 20 years. However, the structure and function of the ecological security network were strongly influenced by human activities and urban sprawl. The ecological security network has deteriorated locally in eastern coastal areas of UAHB during the past 20 years with strong spatial variability in ecological security patterns. To maintain an environment with well-protected ecological quality, we proposed a set of five measures to improve the sustainable development of the ecological system.

show abstract