Sky pixel detection in outdoor imagery using an adaptive algorithm and machine learning

  title={Sky pixel detection in outdoor imagery using an adaptive algorithm and machine learning},
  author={Kerry A. Nice and Jasper S. Wijnands and Ariane Middel and Jingcheng Wang and Yiming Qiu and Nan Zhao and Jason Thompson and Gideon D.P.A. Aschwanden and Haifeng Zhao and Mark R. Stevenson},
Computer vision techniques allow automated detection of sky pixels in outdoor imagery. Multiple applications exist for this information across a large number of research areas. In urban climate, sky detection is an important first step in gathering information about urban morphology and sky view factors. However, capturing accurate results remains challenging and becomes even more complex using imagery captured under a variety of lighting and weather conditions. To address this problem, we… 
Sky and Ground Segmentation in the Navigation Visions of the Planetary Rovers
The proposed framework successfully fills the gap between the laboratory results (with rich idea data) and the practical application (in the wild) and can provide essential semantic information (sky and ground) for the rover navigation vision.
Street view imagery in urban analytics and GIS: A review
Abstract Street view imagery has rapidly ascended as an important data source for geospatial data collection and urban analytics, deriving insights and supporting informed decisions. Such surge has


Sky segmentation in the wild: An empirical study
A deep learning based variant of an ensemble solution that outperforms the methods tested, in some cases achieving above 50% relative reduction in misclassified pixels is proposed.
A physical model-based approach to detecting sky in photographic images
A model-based approach consisting of color classification, region extraction, and physics-motivated sky signature validation is proposed, which computes the sky belief of the region by the percentage of traces that fit the physics-based sky trace model.
Edge-based detection of sky regions in images for solar exposure prediction
This work is concerned with developing a technique which can do this for images taken under different weather conditions and the general approach is to use the Canny edge detector and the morphology closing algorithm to find the regions in the image.
Sky Region Detection in a Single Image for Autonomous Ground Robot Navigation
A sky region detection algorithm within a single image based on gradient information and energy function optimization that is applicable to both colour and greyscale images is proposed.
A Novel Sky Region Detection Algorithm Based On Border Points
Sky detection is an important problem in computer vision. Generally, the color and gradient of the pixel in sky border change greatly. So, sky in the scene can be detected by measuring the sky border
An efficient sky detection algorithm based on hybrid probability model
This work detects a rough sky-ground boundary and calculates the parameters related to appearance of sky to construct a hybrid probability model that indicates how possible a pixel belongs to sky.
Improved blue sky detection using polynomial model fit
A novel algorithm is proposed based on a two-dimensional polynomial model of the image of blue sky to establish high-confidence blue sky regions and correctly increased the belief values on 6 regions found by the initial sky detector.
Sky is not the limit
An automatic background replacement algorithm that can generate realistic, artifact-free images with a diverse styles of skies by utilizing visual semantics to guide the entire process including sky segmentation, search and replacement is proposed.
An Effective Method for Sky Region Detection
An effective method for sky region detection with color information and gradient is proposed. In the proposed algorithm, multi-border points in each column of image can be detected to complete
Classification of weather situations on single color images
This work presents an approach that is able to distinguish between multiple weather situations based on the classification of single monocular color images, without any additional assumptions or prior knowledge.