ImageResize

The project aims at RESIZING a still image, first using Processing System (PS) using Python on an ARM9 processor, and the other executed in hardware using a kernel contained within a hardware overlay in the Programmable Logic (PL).

Technical Description

PS Implementation (Python on ARM9):

• The RESIZE function is executed on the ARM9 processor within the Processing System (PS).
• The ARM9 processor accesses the input image and performs the RESIZE operation using Python code.
• The RESIZE operation involves resampling the input image and applying filtering/interpolation techniques to generate the output image.
• The execution time of the RESIZE function is measured on the ARM9 processor.

Hardware Implementation (PL with Hardware Overlay):

• The RESIZE function is implemented in hardware using a kernel contained within a hardware overlay in the Programmable Logic (PL).
• The input image is passed from the ARM9 processor to the DDR Memory (DRAM).
• The Resize IP block is configured to execute the RESIZE function within the hardware overlay.
• The DMA controller is set up to transfer source pixels from DRAM to the Resize IP block and to transfer destination pixels from the Resize IP block back to DRAM.
• The ARM9 processor can access the resulting resized image stored in DRAM after the hardware acceleration.

Observation:

• FPGA implementation is X4 times faster than Software Implementation.

---