Jones, Gwynfor David; (2003) Fusion of data from visual and low-spatial-resolution thermal cameras for surveillance. Masters thesis (M.Phil), UCL (University College London).

Text Fusion_of_data_from_visual_and.pdf Download (16MB)

Abstract

This thesis investigates whether the segmentation of surveillance images can be improved by fusing low-spatial resolution thermal data with high-spatial resolution visual information. The context of this investigation is the surveillance of sterile zones where an alarm is required should a person enter the zone and at no other time. The aim is to reduce false alarms due to wildlife movement or changes in environmental conditions. Current work on the fusion of visual and thermal data employs expensive high-spatial resolution cameras which preclude many applications; however the advent of low-cost, low-spatial resolution thermal cameras is providing new opportunities. This thesis explores the design issues and advantages in using such an array in conjunction with a visual camera. A calibration methodology has been designed to allow the mapping of a point in one image onto the equivalent point in the other despite the operational constraints of different imaging modalities, resolutions and lens systems. Once they are calibrated it is possible to extract areas of interest from the visual camera using the information from the thermal. Therefore only the relevant part of the visual image is processed allowing a more sophisticated segmentation algorithm to be used. Processing in the visual domain is initiated by a method of background representation or change detection. Segmentation of the highlighted object is achieved using Markov Random Fields. An implementation has been developed which extends previous research by fusing data from the visual and thermal cameras both for static and temporal image sequences. A Hidden Markov Model has been developed to classify pixels in the thermal image for superior extraction of objects of interest over a wide range of operating conditions. This improves the isolation of the object in the thermal image compared with using a threshold strategy.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item