Search results for " Change Detection"

Unsupervised change detection with kernels

2012

In this paper an unsupervised approach to change detection relying on kernels is introduced. Kernel based clustering is used to partition a selected subset of pixels representing both changed and unchanged areas. Once the optimal clustering is obtained the estimated representatives (centroids) of each group are used to assign the class membership to all others pixels composing the multitemporal scenes. Different approaches of considering the multitemporal information are considered with accent on the computation of the difference image directly in the feature spaces. For this purpose a difference kernel approach is successfully adopted. Finally an effective way to cope with the estimation o…

Lost in space? Using geo-narratives to interpret land use changes in a rural landscape of inner Sicily

2021

Intercropping landscapes characterised by the presence of certain plant features (i.e. old-century olive trees) are usually considered traditional landscapes, extremely important for their biocultural heritage. Olive agroforestry systems were widespread in the past throughout Sicily. Recent evolution processes involving intensification have switched to olive grove monocultures. Here we present ongoing work on the application of geo-narratives to interpret transformation dynamics of land use practices in a rural landscape of inner Sicily. Based on the assumption that spatial patterns are the expression of transformation processes, where the spatial variation of human activities is a signific…

A study about DDoS attacks in SIP environments

2007

Modelling Recurrent Events for Improving Online Change Detection

2016

The task of online change point detection in sensor data streams is often complicated due to presence of noise that can be mistaken for real changes and therefore affecting performance of change detectors. Most of the existing change detection methods assume that changes are independent from each other and occur at random in time. In this paper we study how performance of detectors can be improved in case of recurrent changes. We analytically demonstrate under which conditions and for how long recurrence information is useful for improving the detection accuracy. We propose a simple computationally efficient message passing procedure for calculating a predictive probability distribution of …