A Statistical Model for Recreational Trails in Aerial Images

Abstract

We present a statistical model of aerial images of recreational trails, and a method to infer trail routes in such images. We learn a set of textons describing the images, and use them to divide the image into super-pixels represented by their texton. We then learn, for each texton, the frequency of generating on-trail and off-trail pixels, and the direction of trail through on-trail pixels. From these, we derive an image likelihood function. We combine that with a prior model of trail length and smoothness, yielding a posterior distribution for trails, given an image. We search for good values of this posterior using a novel stochastic variation of Dijkstra's algorithm. Our experiments, on trail images and groundtruth collected in the western continental USA, show substantial improvement over those of the previous best trail-finding method.

Global Message Passing in Networks via Task-Driven Random Walks for Semantic Segmentation of Remote Sensing Images

“…There is a long tradition of leveraging computer vision techniques for aerial scene parsing. Earlier works (Liu, Liu, 2014, Blaschke et al, 2004, Predoehl et al, 2013 mainly lie on exploring effective visual features and semantic modeling approaches. Recently, deep CNNs have been widely explored in this field and taken a giant leap (Marcos et al, 2018a, Sun et al, 2019, Azimi et al, 2019, Kellenberger et al, 2019, Cheng et al, 2019, Marcos et al, 2018a, Mou et al, 2019.…”

Let’s Take a Walk on Superpixels Graphs: Deformable Linear Objects Segmentation and Model Estimation

“…Curvature Likelihood p C (π i,vn |π i ) is concerned with estimating the most likely configuration of a DLO's curvature. Following the intuitions of Predoehl et al [32], for each new node v n we can assume that the object's curvature changes smoothly along the walk. To quantify this smoothness criterion we exploit the product of the von Mises distributions of the angles between two successive vertices.…”

Marked Point Process Model for Curvilinear Structures Extraction

“…Sampling techniques with geometric priors have been exploited to detect multiple line segments in a scene [2,16,18,22,24,28]. The marked point process (MPP) framework [6,7,20,26] is helpful to enforce high level constraints on shape prior.…”