A Sensor Independent Concept for the Characterisation of Imaging Spectrometers

Abstract

This thesis describes an operational concept and the implementation of an automated process chain for the characterisation of hyperspectral sensors. The process chain includes the definition of sensor independent measurement procedures, the conduction of the measurements and the analysis of the recorded data to determine sensor parameters. The design driver for the setup of the Calibration Home Base (CHB) was the imaging spectrometer APEX (Airborne Prism EXperiment). But the facility can also be used for the characterisation of other optical sensors like the DLR sensors ROSIS (Reflective Optics System Imaging Spectrometer), ARES (Airborne Reflective Emissive Spectrometer) and AISA (Airborne Imaging Spectrometer for different Applications). The first part of this thesis describes the setup of the laboratory and its devices (chapter 2) as well as the operational concept for the automatic conduction of characterisation measurements (chapter 3). The concept divides the measurement process into three functional modules: a generic measurement control module (Master), a module for the control of the laboratory devices (Slave) and a sensor specific control module (Sensor) per sensor. The single modules interact via TCP/IP on the intranet by exchanging commands using well-defined XML formatted data. The implementation of the concept is described in Chapter 4. Utilising the new Slave software the laboratory can be operated fully automated. The operator in the laboratory does not need a broad knowledge of the operation of the individual devices. The necessary manual settings are interactively reported to the operator by the Slave module. Furthermore the Slave module monitors a multitude of possible errors of the single laboratory devices. In addition the Slave enables external users to use the laboratory needing nothing more than the description of the interface. The Master software is the central superordinate module which controls the Slave and the respective Sensor module. The Master hosts the various characterisation measurement procedures and executes the detailed instructions of the actual procedure(s). The Master offers many different options for the execution of measurements. For example, it is possible to use the emulate mode, or to control only the laboratory devices or the sensor, or to transfer data, or to carry out a data analysis. A time estimation function enables the evaluation of the estimated time for the execution of a measurement series. The Master module has comprehensive monitoring functions to ensure a safe operation of the laboratory, to intercept arising errors, to interrupt the measurement process and to inform the operator immediately. Besides the monitoring functions, the Master stores the actual device settings in measurement logs which are necessary for the analysis software. The current measurement process can be monitored using an internal web site. The operator has to access the laboratory only in case of a necessary manual interaction. This is beneficial for working safety aspects (dark room) and cost reasons. Each sensor needs its individual Sensor module for the communication with the Master module, due to different internal devices which have to be controlled (sensor mirrors, internal calibration devices, shutter, etc.). The sensor control software has to provide an input interface for the communication with the Sensor module. If this is missing the sensor control software has to be modified. For the implementation of a new sensor, a Sensor module is needed and a new sensor interface and a GUI to capture the sensor parameter values have to be implemented in the Master module. The implementations of the two sensors AISA and ROSIS are described in chapter 4.4. With the newly developed software modules larger measurement series can be performed automatically in less time. The second part of the thesis was the development of generic characterisation measurement and analysis methods for imaging spectrometers in the wavelength range from 0.4 μm to 2.5 μm (chapter 5). The three main types are spectral, geometric and radiometric characterisation. Each of the defined measurement methods describes the setup of the necessary laboratory devices and explains the measurement principle. For the essential device parameters general selection criteria were defined. The developed IDL analysis procedures are generic except for the conversion of the raw data into the defined generic data format (ENVI cube format). Each analysis procedure has quality and error checks. The first and second part of this work have been verified independently using the sensors AISA and ROSIS (chapter 6). The third part of the thesis exemplifies characterisation measurements and data analysis performed for the two above mentioned sensors intending to test the methods and to draw some conclusions from the results.