[SmallSat] Onboard Data Processing Pipeline and Software/Hardware Implementation of Extended Multiplicative Signal Correction

Abstract

This thesis concerns the data processing pipeline from anCMV2000v3ImageSensor to memory in a Hyper-spectral imaging Application. ALow Voltage Dif-ferential Signal(LVDS) Receiver was implemented, consisting of aDeserializerModuleand aPixel Order Alignment Moduleto receive and organize the pixeldata from the Image Sensor to the rest of the pipeline. Further aBinning Mod-ulewas implemented to make simple binning operations on the pixel data toreduce minor observation errors as well as reducing the data size. Additionallythe thesis presents an algorithm calledExtended Multiplicative Signal Correc-tion[8], and a proposal to how this can be implemented in a Software/HardwareCo-design executing on the Zynq-7000 platform. A pure software implementa-tion was made to map which parts of the algorithm would benefit the mostfrom being executed in hardware based on time consumption. Then a SW/HWco-design was implemented with the chosen parts accelerated in hardware. Aspeedup of 4.36 was achieved by the combined SW/HW implementation to thepure software version.