Paper
7 November 2007 Detection of mine-like objects in high-resolution aerial images
Author Affiliations +
Abstract
A system for mine detection in aerial images is considered as an interactive system in which the operator is responsible for making iterative queries to the database of images and analyzing the results. Preliminary, each image undergoes formal decomposition into a set of feature vectors. Each feature vector is calculated for every irregularity found at a scale-invariant salient point detection stage where a blob detector is used. Assuming that every small object in the image can be described by a single invariant feature vector calculated on a patch around the salient point, formalization of search algorithms is feasible. While the template-based search is straightforward in the terms of one object - one feature vector, we focus on another important option when searching for mines - similar object search. For similar object search a hierarchical clustering algorithm is considered. The mentioned steps of image processing as well as similar object search are illustrated by performing on aerial mine field images taken by an electronic camera from a height of 27 meters. Encouraging preliminary results lead to the formulated plan for future research. The developed algorithms are planned to be used in an image search engine allowing the operator to interactively search for mines in a database of aerial images in humanitarian de-mining operations.
© (2007) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Andrey V. Zhuravlev, Alexander S. Bugaev, Sergey I. Ivashov, and Vladimir V. Razevig "Detection of mine-like objects in high-resolution aerial images", Proc. SPIE 6739, Electro-Optical Remote Sensing, Detection, and Photonic Technologies and Their Applications, 673905 (7 November 2007); https://doi.org/10.1117/12.735988
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Mining

Genetic algorithms

Databases

Detection and tracking algorithms

Feature extraction

Image processing

Algorithm development

Back to Top