Page 73 - Book of Abstracts 2023

73
A review of algorithms and methods for detecting anomalies in in hyperspectral spectral spectral images EEE-1-3
By: Nir Rosh nir rosh21@gmail com Advisor: Dr Isaac August Shamoon College of Engineering Beer-Sheva
In In recent years using spectral-imaging remote remote remote sensing sensing data has been a a a a a a a a a a a a hot issue In In the the process of of remote remote remote hyperspectral sensing sensing information is is is is is extracted from from objects that are remote remote remote from from the the the system In this process there is is is is is is exposure to to a a a a a a a a a a a a a a a a a a a a huge amount of of of data data and and as a a a a a a a a a a a a a a a a a a a a result it it is is is is is is difficult to to to distinguish between normal normal normal normal and and and abnormal abnormal data data With the the the help of of the the the anomaly detection method it it is is is is is possible to to conclude which which target is is is is is normal normal normal normal and and which which is is is is is abnormal abnormal Identifying anomalies in in in in in in in in hyperspectral hyperspectral imaging involves identifying unusual unusual data data points or or or or unusual unusual cases in in in in in in in in the the data data set captured by by the the hyperspectral hyperspectral system by by using three common and and useful algorithms: Isolation Forest Support Vector Machine (SVM) and and Local Outlier Factor (LOF) Keywords: anomalies anomaly detection data hyperspectral sensing Implementation of a a a a a a a computational liquid-crystal hyperspectral imaging system EEE-1-4
By: Erez Abadyan erezab1@ac sce sce ac ac ac ac il il Netanel Zigdon netznzi1@ac sce sce ac ac ac ac ac il il Advisor: Dr Isaac August Shamoon College of Engineering Beer-Sheva
There are are different methods methods both classical and and computational for acquiring hyperspectral images Classical methods methods methods directly measure data without algorithmic reconstruction and and and are are subject subject to to the the the the the the sampling theory theory theory of Nyquist Nyquist Advanced computational methods methods on on on on on the the the the the the other hand may be based on on on on on compressed sensing theory theory theory which is is is not subject subject to to Nyquist's theory theory theory Hyperspectral imaging systems
that rely on on on on on on these computational methods have greater data compression capabilities making image
acquisition easier However the the the drawback is is is the the the long time required to recover each spectral spectral cube To deal with this problem we we developed a a a a a a a a a a a a a a a a a hyperspectral hyperspectral imaging system that combines neural neural network network reconstruction algorithms Keywords: compressive sensing hyperspectral hyperspectral images liquid crystal neural neural network network Book of Abstracts | 2023