SPIE Startup Challenge 2015 Founding Partner - JENOPTIK Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:

SPIE Photonics West 2017 | Register Today

SPIE Defense + Commercial Sensing 2017 | Call for Papers

Get Down (loaded) - SPIE Journals OPEN ACCESS


Print PageEmail PageView PDF

Remote Sensing

The importance of characterizing a versatile digital camera system

For optimal data acquisition and the ability to choose the optical configuration that closely fits the user's needs, good knowledge of the versatile multi-spectral airborne system being used is required.
27 April 2006, SPIE Newsroom. DOI: 10.1117/2.1200603.0170

Airborne remote sensing is a very flexible way to map and monitor the land surface. It provides digital image data with very fine spatial resolutions on a precise target area. However, to extract the best information, providers must achieve an optimal optical configuration and a suitable geometric and radiometric correction of the raw data acquired. With the recent development of large visible-spectrum CCDs, snapshot images become feasible with either narrow or large spectral bandwidth. This makes possible many optical configurations to suit a large panel of experiments. However, snapshot images need some calibration. To keep the versatility that is the main advantage of airborne systems, rigorous but light calibration is necessary.

A technical agreement between Office National d'Études et de Recherche Aérospatiales (ONERA), Institut Géographiques National (IGN), and Centre National d'Études Spatiales (CNES) led, in 2004, to the development1, calibration2, and use3 of a versatile multi-spectral airborne sensor based on a 4096 × 4096 pixel CCD detector called PELICAN (Plateforme Et Logiciels Informatiques de Caméras Aéroportées Numériques). The main features of this airborne system are summarized in Table 1. and Table 2. The goal of PELICAN is to deliver radiometrically and geometrically well-calibrated images. Several accurate and exhaustive guides4,5 for calibration are available. But the question remains: what is the best configuration and the fastest way to calibrate a versatile system?

Table 1. Shown are the main features of PELICAN.

Number of spectral bands 4 (up to 8 synchronous bands)
Spectral range 400 to 950nm
Spectral bandwidth 300 to 20nm
Digital output 12 bits (4096 levels)
Detector size (for each spectral band) 4096×4096 pixels (9×9μm)
Focal length 50 to 150mm
Field of view (FOV) ±7° to ± 17°
Instantaneous field of view (IFOV) 60 to 150 μrad

Table 2. Shown are the features of the airborne system.

Aircraft Beechcraft 200 — King Air
Altitude Up to 10,000m (30,000ft)
Speed About 370km/h (200kts)
Airborne standard support Wild
Frame rate Snapshot or 0.25Hz
Swath width 200 to 4000m
Spatial resolution 5cm to 1m
Image overlapping Up to 70%
Forward moving compensation GPS driven TDI mode
Storage capacity 4500 multi-spectral images

The PELICAN operational process is shown on the flow chart in Figure 1. The first step is to express the customer's need in terms of a suitable sensor configuration. Using previous characterization and calibration results may contribute to cost reduction. However, experience of and knowledge about our system make the radiometric-calibration steps easier. It is often more efficient to make the calibration measurement after the airborne campaign, in order to calibrate the configurations that were actually used.

Figure 1. Shown is the flow chart for operating PELICAN.

Our calibration process consists of four main steps: modeling, characterization, acquiring sensitivity coefficients, and implementing corrections. In order to reduce the calibration measurements, we first determine a radiometric model. By taking into account the relationship and behavior of all the terms of the model, we can reduce the number of system configurations to be radiometrically corrected. The characterization specifies the validity domain of our model and gives a clear indication of new steps to be investigated for correction.

The radiometric model gives us the relationship between digital count and equivalent radiance at pupil entrance level. The model points out that the signal of each pixel is spatially dependent (due to lens effects), spectrally dependent, time dependent (through integration time and darkness current), and configuration dependent (through the optical f-number and electronic gain).

All these items need some investigation to check how they behave. The relationship between digital count and equivalent radiance at the pupil entrance can be considered as linear in the first approximation. From these investigations, we can deduce the following calibration equation:

Leq i,j is the equivalent radiance at the pixel (i,j),

Acal is the absolute sensitivity, including both the spectral dependence and the constants found in the radiometric equation, and achieved with an f-number Ncal, and an integration time of ti_cal,

Ei,j is the inter-pixel sensitivity (flat field matrix),

Xi,j the raw value at pixel (i,j), and

Oi,j is the darkness level.

This model is interesting because it shows that, with very few measurements, we can have a complete set of correction coefficients. But this apparent ease hides strong hypotheses, namely, the independence of Acal and Ei,j from the f-number and integration time, and the effective linearity. These hypotheses are checked during the characterization and calibration phases.

The light source used for calibration is the integrating sphere from CNES, already used for Satellite Probatoire d'Observation de la Terre (SPOT) calibrations (see Figure 2). This sphere provides homogeneous and calibrated spectric brightness. We use it at the same time to acquire a flat field matrix. Acquisitions are carried out for each camera.

Figure 2. The photograph shows the calibration of PELICAN in front of 160cm diameter integrating sphere (CNES/ASTRIUM facility).

PELICAN is now ready for use by the scientific community for a large range of experiments. This multi-spectral airborne digital camera combines high-spatial with high-spectral resolution. As a versatile facility, it can be placed in optical configurations that closely fit the user's need. We can provide several levels of products, at radiometrically and geometrically corrected levels, enabling fusion and composition of products for geographic information systems. The fact that the same team controls the operations of ground preparation, in-flight acquisitions, and image processing improves the reliability of the data quality.

Philippe Déliot
Philippe Déliot is a physicist and engineer who worked at the Theoretical and Applied Optics department. The main work of his unit concerns quality image measurement for visible and infrared remote sensors.
Joel Duffaut
Joel Duffaut is the project manager of the airborne multi-spectral digital sensor called PELICAN. He is a specialist in airborne sensor operation.