A Parametric Algorithm for Skyline Extraction

Abstract

International audienceThis paper is dedicated to the problem of automatic skyline extraction in digital images. The study is motivated by the needs, expressed by urbanists, to describe in terms of geometrical features, the global shape created by man-made buildings in urban areas. Skyline extraction has been widely studied for navigation of Unmanned Aerial Vehicles (drones) or for geolocalization, both in natural and urban contexts. In most of these studies, the skyline is defined by the limit between sky and ground objects, and can thus be resumed to the sky segmentation problem in images. In our context, we need a more generic definition of skyline, which makes its extraction more complex and even variable. The skyline can be extracted for different depths, depending on the interest of the user (far horizon, intermediate buildings, near constructions , ...), and thus requires a human interaction. The main steps of our method are as follows: we use a Canny filter to extract edges and allow the user to interact with filter's parameters. With a high sensitivity , all the edges will be detected, whereas with lower values, only most contrasted contours will be kept by the filter. From the obtained edge map, an upper envelope is extracted, which is a disconnected approximation of the skyline. A graph is then constructed and a shortest path algorithm is used to link discontinuities. Our approach has been tested on several public domain urban and natural databases, and have proven to give better results that previously published methods

Detection of Land in Marine Images

“…The idea of the line-based methods is to identify pixels which are likely to belong to the HL, and then to combine some of them to form a true line. Extraction of HL pixels is performed through application of edge filters (the most popular is Canny filter) [11][12][13] or edge filters combined with machine learning techniques (e.g., support vector machine, convolutional neural network) [14][15][16]. In order to build the HL based on the extracted pixels, Ayadi et al [11] connect pixels of the highest contrast which form the shortest path between two edges of the image.…”

“…In order to build the HL based on the extracted pixels, Ayadi et al [11] connect pixels of the highest contrast which form the shortest path between two edges of the image. In [12], a line with the highest skyline measure is considered to be the HL. The skyline measure characterizes each line by the length, contrast, and homogeneity.…”

The Skyline as a Marker for Augmented Reality in Urban Context

“…Our method ( Figure 1) takes as input a live video stream from the smartphone's camera and a point cloud of the neighborhood urban scene. We propose a two-step registration method to refine the estimated camera's pose, knowing the camera's intrinsic parameters: first, a live video-stream is acquired with the smartphone's camera from which a real skyline (1) is extracted using a skyline extraction algorithm [15]. Then, a camera pose is estimated in world coordinate system with its 6 degrees of freedom due to the combination of smartphone's embedded sensors, as explained in section 4.2 : magnetic compass, gyroscope, accelerometer and barometer.…”

Sky-CNN: A CNN-based Learning Approach for Skyline Scene Understanding

“…Skyline scenes are very particular scenes that have not been well-handled on image understanding. These scenes were scientifically interesting for geography and urbanism domains [2,3,4,5,6] since they are considered as a particular dimension of some cities. Our skyline scenes understanding mission was lead in a multidisciplinary research project named SKYLINE 1 .…”