Automated search for livestock enclosures of rectangular shape in remotely sensed imagery

Abstract

We introduce an approach for the detection of approximately rectangular structures in gray scale images. Our research is motivated by the Silvretta Historica project that aims at automated detection of remains of livestock enclosures in remotely sensed images of alpine regions. The approach allows detection of enclosures with linear sides of various sizes and proportions. It is robust to incomplete or fragmented rectangles and tolerates deviations from a perfect rectangular shape. Morphological operators are used to extract linear features. They are grouped into parameterized linear segments by means of a local Hough transform. To identify appropriate configurations of linear segments we define convexity and angle constraints. Configurations meeting these constraints are rated by a proposed rectangularity measure that discards overly fragmented configurations and configurations with more than one side completely missing. The search for appropriate configurations is efficiently performed on a graph. Its nodes represent linear segments and edges encode the above constraints. We tested our approach on a set of aerial and GeoEye-1 satellite images of 0.5m resolution that contain ruined livestock enclosures of approximately rectangular shape. The approach showed encouraging results in finding configurations of linear segments originating from the objects of our interest.

Detection of incomplete enclosures of rectangular shape in remotely sensed images

“…2.3). In contrast to [38], the new rectangularity measure does not rely on a heuristic partitioning of the set of linear segments into four subsets. Hard decisions are softened.…”

“…Using remotely sensed imagery in Sec. 3, we, for the first time, evaluate the performance of the normalized maximal rectangularity (NMR) measure [38] and the introduced here rectangularity features for enclosure detection. We also compare them with the GODF-based feature from [1] used for building detection.…”

“…Our approach relies on the basic detection scheme, which includes localization of candidate points, we presented in [38]. A binary map of edges accompanied by angle information is computed first.…”

“…Examples are detection of buildings in remotely sensed images [24,19,14,3,12,15,1,31,30,23,26], traffic signs [21,13,22], and particles of a rectangular shape in cryo-electron microscopy images [36,35]. The methods used were based on Markov Random Fields [15,21], Marked Point Processes [1,26], search on a graph [14,38], Hough Transform and other voting schemes [3,12,36,13,22], template matching [25], aggregation of local features [1,23,31], and heuristic rules [19].…”

A morphological approach for distinguishing texture and individual features in images

“…6 (left) contain rectangular structures composed of nearly linear walls that were used as livestock enclosures. This type of structures is needed to be detected within the archaeological project (Lambers and Zingman, 2012, in press;Zingman et al, 2013a). The white top-hat transform is commonly used to remove background and emphasize small bright structures in images.…”