Detection of Incomplete Enclosures of Rectangular Shape in Remotely Sensed Images

Igor Zingman, Dietmar Saupe, Karsten Lambers; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2015, pp. 87-96

We develop an approach for detection of ruins of livestock enclosures in alpine areas captured by high-resolution remotely sensed images. These structures are usually of approximately rectangular shape and appear in images as faint fragmented contours in complex background. We address this problem by introducing a new rectangularity feature that quantifies the degree of alignment of an optimal subset of extracted linear segments with a contour of rectangular shape. The rectangularity feature has high values not only for perfect enclosures, but also for broken ones with distorted angles, fragmented walls, or even a completely missing wall. However, it has zero value for spurious structures with less than three sides of a perceivable rectangle. Performance analysis using large imagery of an alpine environment is provided. We show how the detection performance can be improved by learning from only a few representative examples and a large number of negatives.