Vehicle detection from very-high-resolution (VHR) aerial imagery using attribute belief propagation (ABP)

Yanli Wang; Ying Li; Li Zhang; Yuchun Huang

doi:10.1117/12.2241295

26 October 2016 Vehicle detection from very-high-resolution (VHR) aerial imagery using attribute belief propagation (ABP)

Yanli Wang, Ying Li, Li Zhang, Yuchun Huang

Author Affiliations +

Proceedings Volume 10008, Remote Sensing Technologies and Applications in Urban Environments; 100080Y (2016) https://doi.org/10.1117/12.2241295
Event: SPIE Remote Sensing, 2016, Edinburgh, United Kingdom

Abstract

With the popularity of very-high-resolution (VHR) aerial imagery, the shape, color, and context attribute of vehicles are better characterized. Due to the various road surroundings and imaging conditions, vehicle attributes could be adversely affected so that vehicle is mistakenly detected or missed. This paper is motivated to robustly extract the rich attribute feature for detecting the vehicles of VHR imagery under different scenarios. Based on the hierarchical component tree of vehicle context, attribute belief propagation (ABP) is proposed to detect salient vehicles from the statistical perspective. With the Max-tree data structure, the multi-level component tree around the road network is efficiently created. The spatial relationship between vehicle and its belonging context is established with the belief definition of vehicle attribute. To effectively correct single-level belief error, the inter-level belief linkages enforce consistency of belief assignment between corresponding components at different levels. ABP starts from an initial set of vehicle belief calculated by vehicle attribute, and then iterates through each component by applying inter-level belief passing until convergence. The optimal value of vehicle belief of each component is obtained via minimizing its belief function iteratively. The proposed algorithm is tested on a diverse set of VHR imagery acquired in the city and inter-city areas of the West and South China. Experimental results show that the proposed algorithm can detect vehicle efficiently and suppress the erroneous effectively. The proposed ABP framework is promising to robustly classify the vehicles from VHR Aerial imagery.

Conference Presentation

Citation Download Citation

Yanli Wang, Ying Li, Li Zhang, and Yuchun Huang "Vehicle detection from very-high-resolution (VHR) aerial imagery using attribute belief propagation (ABP)", Proc. SPIE 10008, Remote Sensing Technologies and Applications in Urban Environments, 100080Y (26 October 2016); https://doi.org/10.1117/12.2241295

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available