Omvärldsanalys inom signalbehandling för EO/IR-sensorer

Authors:

  • Fredrik Näsström
  • Fredrik Bissmarck
  • Viktor Deleskog
  • David Gustafsson
  • Jörgen Karlholm
  • Jonas Nordlöf
  • Jonas Nygårds

Publish date: 2016-08-31

Report number: FOI-R--4287--SE

Pages: 64

Written in: Swedish

Keywords:

  • detection
  • recognition
  • target tracking
  • positioning and sensor control

Abstract

The project Intelligent reconnaissance functions has the task to conduct research with the aim to make future sensor systems more intelligent. By being more intelligent, future sensor systems provide sensor operators with better situational awareness, higher system confidence, reduced risk of human errors (eng. Human error) and low mental workload. This report aims at giving an overview of current research in signal processing for EO/IR sensors. In this work, we have analyzed the research in automatic target detection, target recognition, target tracking, positioning and sensor control. The research within target detection and target recognition has been very strong in recent years. A relatively new approach of machine learning called Deep learning has produced dramatic advances in several image processing applications to levels that were previously unrealistic and far into the future. Research within the positioning area is still very active. Especially SLAM (simultaneous localization and mapping) is a very active research area. The goal with SLAM is with a sensor platform in motion at the same time build a map and locate the platform in the map that is being expanded when the platform move. This can be made by finding and following stable and well defined points in images from different sensors, called landmarks, which then can be used to calculate an improved position and orientation estimate. More advanced SLAM methods are becoming more common where all landmarks are used in the same system of equations to calculate the whole trajectory (i.e. not just the last position). In the sensor control area, the research is still very active with many different applications. Research in this area is constantly advancing with new methods and algorithms applied to solving the planning problem. For sensor control of multiple sensor implies new questions about how and what information that should be exchanged between the sensor systems so that they could cooperate. Research on multi-robot systems (cooperative robots) is emerging quickly in research with advanced applications in various robot competitions. Cooperative sensors have become increasingly attractive in recent years to use all available sensors in a better way.

Share page on social media