Decision Support System for Image Analysts of the Royal Armed Forces of the Netherlands

Abstract

One of the tasks of a human image analyst of the Royal Armed Forces of the Netherlands is to assess the threat level of vehicles. Technical improvements provide means to cover larger areas in more detail resulting in higher workload for the human analysts. In order to circumvent cognitive lock up or overload, supporting the human analyst in their task is a necessity. This study presents a reasoning system which can evaluate whether a vehicle poses a threat given a certain situation in order to support the human analyst. Multiple models were trained, compared and reviewed in order to investigate which elements are useful for such a reasoning system. All models were trained using participant test data that were gathered from average Dutch civilians performing a simpli?ed version of the image analyst task. The results showed that the amount of predictors, the combination of predictors and training of the models are important elements for the reasoning system in order to properly asses threats. However, future research should indicate whether these models perform well enough in real world situations.