Models used in assessment of optical sensors

Authors:

  • Sten Nyberg

Publish date: 2007-08-14

Report number: FOI-R--2290--SE

Pages: 27

Written in: English

Keywords:

  • optical sensor
  • synthetic imagery
  • performance model
  • assessment

Abstract

This report contains a description of three models that can be used in the assessment of an optical sensor. The focus is on methods assessing sensor performance, not the technical details of the sensor. The complete chain from target to observer is taken into account. From a lot of parameters an output in terms of recognition probability is obtained. Here recognition is used as a common notion for detection, classification and identification. The first model is a simple model called the Bailey model which assumes an observer is looking at the target. He might be using a binocular or an electro-optical sensor to aid him. The second model is one of latest versions of the well known Night Vision Lab model named NVTherm. It has a very detailed description of the sensor. The third model differs from the preceding ones since it starts with an image of the scene taken with a relevant choice of target, background, atmosphere and sensor. A small experiment is made where synthetic images generated with the Cameosim computer code are compared to real images captured with an IR sensor. The result may give some guidance concerning important aspects in generating synthetic imagery.

Share page on social media