APEX is an airborne (dispersive push broom) imaging spectrometer developed by a Swiss-Belgian consortium on behalf of ESA. It is intended as a simulator and a calibration and validation device for future spaceborne hyperspectral imagers. Furthermore, APEX is an advanced scientific instrument for the European remote sensing community, recording hyperspectral data in approximately 300 bands in the wavelength range between 400 nm and 2500 nm and at a spatial ground resolution of 2 m to 5 m.
Imaging spectroscopy greatly extends the scope of traditional remote sensing. It is based on the detection of many narrow, contiguous spectral bands. This presents opportunities for more precise identification of surface materials than is possible with broadband multispectral sensors.
For more information visit the APEX-website : http://www.apex-esa.org
Hyperspectral imaging is a powerful technique for monitoring the earth in a variety of ways through environmental assessments, vegetation species monitoring, health and land use management studies, etc.. This is done through image acquisition in many narrow spectral bands to detect narrow band features and subtle variations in the reflectance spectra of natural objects. Although conceptually very powerful, acquiring narrow spectral bands from a moving platform faces an unavoidable trade-off:
- high spatial resolution to avoid intermixing of spectral signatures of small scale features
- high signal to guarantee that the spectral signature can be effectively discriminated from background noise
Hyperspectral instruments for earth observation are often built using prism or grating based pushbroom spectrometers making use of a complex optical system. They also rely on a stabilised environment and high quality inertial navigation systems to achieve precise georeferencing of their line-based image data. Both aspects limit the use of these instruments on mass and volume constrained platforms.
Therefore we propose a novel camera concept which we call “geospectral camera”. Within the volume of a digital camera, the instrument allows to generate geometrically precise hyperspectral image data combined with high spatial resolution images. In the field of earth observation this camera is therefore ideally suited for mass and volume constrained platforms like compact remotely piloted aircraft systems (RPAS) or very small satellites.
The unique characteristics of the geospectral camera originate from the innovative detector configuration. It consists of two 2-dimensional sensor elements on a single image sensor chip. Since the two sensor elements have been processed on one single die with lithographic accuracies, they are perfectly aligned with respect to each other and their relative position is known very accurately. Further more, the spectral sensor element of the camera, the “spectral sensor”, is equipped with optical filters positioned just in front of the detector. These filters typically have a 2-dimensional structure consisting of narrow spectral bandpass filters, of which the central wavelength is uniform in one direction and variable in the other direction. The second sensor element is a panchromatic or spectrally broadband frame sensor which is called the “geometric sensor”. These characteristics allow a simple and compact optical configuration for the camera.
It is important to note that with this filter concept the complete reflectance spectrum of one spatial point on the ground can only be reconstructed by combining the spectral line information in different images acquired while scanning the camera over this point. It is clear that this process requires a very accurate determination of the camera exterior orientation. The concept of the geo-spectral camera is by design able to tackle this challenge by using the fact that both sensor elements are tightly coupled on a single die.
The narrow spectral band filter limits the amount of light falling on the spectral sensor element. Therefore it needs to be configured for longer exposure times and/or pixel binning modes. The geometric sensor element is operated with much shorter exposure times and captures the geometric information of the scene at high spatial resolution. The frame images acquired by the geometric sensor are used to derive the exterior camera parameters. Due to the strong geometric link between the two sensors on a single chip also the orientation of the spectral sensor can be derived very precisely. This approach relaxes the specifications of the needed inertial navigation system to the level of typical devices used in handheld mobile devices. In addition, both high spatial (geometric sensor element) and hyperspectral image data (spectral sensor element) are being acquired of the target region that can be exploited to perform image fusion or spatially sharpen the hyperspectral image product.
VITO has developed a breadboard version of the geospectral camera concept. It makes use of the unique MEDUSA CMOS image sensor which contains two sensor elements of each 10000 x 1200 pixels with 5.5 µm pitch on a single imager chip: one sensor element is equipped with an RGB Bayer filter, one sensor element is panchromatic. The current version of the geospectral camera is equipped with an optical filter covering the range 450-900 nm in more than 40 bands with a spectral resolution of about 15 nm. The filter is deposited on a glass substrate which is placed in front of the panchromatic sensor element.
The COSI imager, originating from the concept of the geospectral camera and suitable for small remotely controlled aircraft systems (RPAS), has been recently developed at VITO with co-funding of the EC FP7 Airbeam security project. To capture spectral data the COSI imager employs step line interference filter deposited directly on the sensor surface. Neighbouring image rows correspond to a different spectral band as well as a different location on the ground. Therefore the scanning motion is required to cover the scene, and to retrieve the complete spectrum for every spatial location.
COSI camera developed at VITO
72 narrow (FWHM: 5nm to 10 nm) spectral bands of the filter cover the spectral range of 600-900 nm. Such spectral information is highly favourable for vegetation studies, since the main chlorophyll absorption feature centred around 680nm is measured, as well as, the red-edge region (680nm to 730nm) which is often linked to plant stress. The NIR region furthermore reflects the internal plant structure, and is often linked to leaf area index and plant biomass.
The payload is compact (6cm x 7cm x 11.6cm) and lightweight, with the total mass of 500g including: an embedded computer, power distribution unit, data storage and optics (330g without optics).
The imager captures very high spatial resolution data, i.e. images captured with a 9mm lens at 40m altitude cover the swath of ~40m and geometrically correct (orthorectified) hyperspectral data can be reconstructed with a ground sampling distance of ~4cm.
COSI datasheet can be accessed here.
Strawberry field near Sint Truiden, Belgium. Hyperspectral data reconstructed with ground sampling distance of 4cm. False colour composite (R=801.7nm, G=672.6nm, B=604.0nm).
The acquired images are processed into a conventional hypercube using a dedicated processing chain developed at our institute (COSI-Cube). Radiometric and spectral correction are applied in order to derive reflectance values of the imaged area. The COSI-Map modulewas designed to provide various information maps representing the status of the vegetation health. Based on this information management decisions can be optimized within the agricultural fields.
Auxiliary data, such as geolocation of the images or ground control points (GCPs) are not required, although their presence improves data scaling and georeferencing.
Although the imager is used similarly to line scanner, it captures a series of traditional 2D perspective images and therefore allows for extraction of 3D information such as digital surface model, required to produce geometrically correct, orthorectified hypercube. Another advantage of the COSI camera is the lack of moving parts used in imagers based on the Fabry–Pérot interferometer.
COSI system - hyperspectral imager and data processing modules
More information on the COSI system as well as examples of the acquired datasets and data analysis can be found in the following papers:
Sima, A. A., Baeck, P., Nuyts, D., Delalieux, S., Livens, S., Blommaert, J., Delauré, B., and Boonen, M., 2016: COmpact hyperSpectral Imaging system (cosi) for small remotely piloted aircraft systems (RPAS) – system overview and first performance evaluation results, Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLI-B1, 1157-1164,http://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLI-B1/1157/2016/
Baeck, P., Blommaert, J., Delalieux, S., Delauré, B., Livens, S., Nuyts, D., Sima, A., Jacquemin, G., Goffart, J.P., 2016: High resolution vegetation mapping with a novel compact hyperspectral camera system. Proceedings of the 13th International Conference on Precise Agriculture, July 31-August 4, 2016, St. Luis, USA. https://www.ispag.org/media/abstracts/2121_Full_paper_1464734865_Abstract%202121%20Pieter-Jan%20Baeck.docx
The LiCrIS (Liquid Crystal Imaging Spectrometer) system is a small, light hyperspectral camera, developed by VITO operating in the 400-720 wavelength range to provide information on agriculture, sand transport and water quality.
Schematic overview of the LICRIS system
LiCrIS consists of four basic parts: the VariSpec Liquid Crystal Tunable Filter (PerkinElmer Inc.), the optical system, the Basler scA1400-30gm camera (Basler AG) and an embedded computer.
Licris setup on a fixed pole at the border of a water reservoir