Horizon line detection using supervised learning and edge cues

Ahmad, Touqeer; Bebis, George; Nicolescu, Mircea; Nefian, Ara; Fong, Terry

doi:10.1016/j.cviu.2019.102879

Cited by 9 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If an edge node k is found within the search window, gap filling is performed by introducing dummy nodes between node i in stage j and node k within the search window j+tog and a high cost is associated with such dummy nodes. More details for Lie et al can be found in [38] or [36].…”

Section: A Skyline Detection As Graph Search Problemmentioning

confidence: 99%

“…In a series of work [30]- [36], Ahmad et al extended the skyline detection approach of [38] by incorporating supervised machine learning techniques. Interested readers are requested to please consult their original papers [30]- [36].…”

Section: B Skyline Detection Using Supervised Learning and Edgesmentioning

confidence: 99%

“…In the second set of their work [33], [35], [36], Ahmad et al investigated to exclude the edge-detection and generated a dense classification score image (DCSI) which reflects the likelihood of a pixel belonging to the skyline. The resultant DCSI is then used directly to initialize the nodal-cost:…”

Section: Skyline Detection Using Dense Classification Score Imagesmentioning

confidence: 99%

See 2 more Smart Citations

Resource Efficient Mountainous Skyline Extraction using Shallow Learning

Ahmad¹,

Emami²,

Čadík³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Skyline plays a pivotal role in mountainous visual geo-localization and localization/navigation of planetary rovers/UAVs and virtual/augmented reality applications. We present a novel mountainous skyline detection approach where we adapt a shallow learning approach to learn a set of filters to discriminate between edges belonging to sky-mountain boundary and others coming from different regions. Unlike earlier approaches, which either rely on extraction of explicit feature descriptors and their classification, or fine-tuning general scene parsing deep networks for sky segmentation, our approach learns linear filters based on local structure analysis. At test time, for every candidate edge pixel, a single filter is chosen from the set of learned filters based on pixel's structure tensor, and then applied to the patch around it. We then employ dynamic programming to solve the shortest path problem for the resultant multistage graph to get the sky-mountain boundary. The proposed approach is computationally faster than earlier methods while providing comparable performance and is more suitable for resource constrained platforms e.g., mobile devices, planetary rovers and UAVs. We compare our proposed approach against earlier skyline detection methods using four different data sets. Our code is available at https://github.com/TouqeerAhmad/skyline detection.

show abstract

Section: A Skyline Detection As Graph Search Problemmentioning

confidence: 99%

Section: B Skyline Detection Using Supervised Learning and Edgesmentioning

confidence: 99%

Section: Skyline Detection Using Dense Classification Score Imagesmentioning

confidence: 99%

See 1 more Smart Citation

Resource Efficient Mountainous Skyline Extraction using Shallow Learning

Ahmad¹,

Emami²,

Čadík³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In order to reduce the likelihood of coding different tilted versions of the same skyline, the normal vector of the curve was recorded at each sampled point of skyline (as illustrated in Figure 11), and 32-bit chain code was used for coding. Given the slope of Therefore, after extracting the skyline using supervised learning [27] and saving it as a list (as illustrated in Figure 12), a skyline comprises two Enhanced Angle Chain codes named curvature Enhanced Angle Chain code and normal Enhanced Angle Chain code to describe its features (see Table 2). Curvature enhanced chain code records the angle between the sampled point of the curve and the surrounding sampled points.…”

Section: Enhanced Angle Chain Codementioning

confidence: 99%

“…Therefore, after extracting the skyline using supervised learning [27] and saving it as a list (as illustrated in Figure 12), a skyline comprises two Enhanced Angle Chain codes named curvature Enhanced Angle Chain code and normal Enhanced Angle Chain code to describe its features (see Table 2). Therefore, after extracting the skyline using supervised learning [27] and saving it as a list (as illustrated in Figure 12), a skyline comprises two Enhanced Angle Chain codes named curvature Enhanced Angle Chain code and normal Enhanced Angle Chain code to describe its features (see Table 2). Curvature enhanced chain code records the angle between the sampled point of the curve and the surrounding sampled points.…”

Section: Enhanced Angle Chain Codementioning

confidence: 99%

Camera Geolocation Using Digital Elevation Models in Hilly Area

et al. 2020

View full text Add to dashboard Cite

The geolocation of skyline provides an important application in unmanned vehicles, unmanned aerial vehicles, and other fields. However, the existing methods are not effective in hilly areas. In this paper, we analyze the difficulties to locate in hilly areas and propose a new geolocation method. According to the vegetation in hilly area, two new skyline features, enhanced angle chain code and lapel point, are proposed. In order to deal with the skyline being close to the camera, we also propose a matching method which incorporates skyline distance heatmap and skyline pyramid. The experimental results show that the proposed method is highly effective in hilly area and has a robust performance against noise and rotation effects.

show abstract

Visual-based obstacle detection and tracking, and conflict detection for small UAS sense and avoid

Opromolla

Fasano

2021

Aerospace Science and Technology

View full text Add to dashboard Cite

Horizon line detection using supervised learning and edge cues

Cited by 9 publications

References 30 publications

Resource Efficient Mountainous Skyline Extraction using Shallow Learning

Resource Efficient Mountainous Skyline Extraction using Shallow Learning

Camera Geolocation Using Digital Elevation Models in Hilly Area

Visual-based obstacle detection and tracking, and conflict detection for small UAS sense and avoid

Contact Info

Product

Resources

About