Publication Type:Journal Article
Source:Journal of Defence & Security Technologies, Volume 5, Issue 3, Number 3, p.44-57 (2022)
Keywords:AI, Border surveillance, machine learning, occlusions, sensor fusion, visual sensors
In this paper we introduce AIT’s Smart Sensor Technology platform, developed within the FOLDOUT project (Through-foliage detection, including in the outermost regions of the EU). The novel platform is part of FOLDOUT’s solution for ground based border surveillance, and in particular detection of foliage penetration. The platform combines a high resolution (4K) RGB camera, a thermal camera, as well as a 30x zoom (2K) NIR low-light camera. It is designed to work day and night under strongly varying weather conditions, it is transportable and can be deployed in remote areas, as well as being self-sufficient if operated by battery. The typical sensor range is about 100-200m which can be adapted per specific usage. The platforming thobe used for surveillance of dedicated areas or by providing an open interface to be integrated into whole surveillance systems in which ground-based sensors are combined with high altitude sensors to provide both close and wide area coverages. The platform provides state-of-the-art AI-based detection algorithms for object detection and classification, for both RGB and thermal images. The individual detections are further fused by establishing coincidence in time and space between the individual detectors, where sensor is geo-localized for accurate target localization and visualization on a map of the area to be monitored.