• Simulink
• MATLAB
• KCU105
• image processing
• ILA
• hyperspectral image
• high-level synthesis
• HDL Coder
• FPGA
• SoC
• VHDL
• Vitis
• Vivado

This study aims to design a block using MATLAB/Simulink capable of taking two images as input, performing an algorithm, and calculating the Normalized Difference Vegetation Index (NDVI) as output. The HDL Coder tool is used to convert the Simulink model into HDL, followed by the creation of a System-on-Chip (SoC) around the IP generated and implementation of the architecture on the KCU105 FPGA.
The validity of the design was tested both in software using Simulink and ModelSim and in hardware using the Integrated Logic Analyzer (ILA), as well as through direct implementation on the board.
Eight different Simulink models were developed to compare them in terms of estimated resource usage and critical path. After verifying the consistency of results in Simulink and ModelSim, two of the models were chosen for hardware implementation. Through the Vitis console, the expected indices can be observed.