FPGA-Based Soft-Core Processors for Image Processing Applications.

Area of Science:

• Computer Engineering
• Digital Signal Processing
• Embedded Systems

Background:

• Increasing demand for efficient image data processing in security and surveillance.
• Field-Programmable Gate Arrays (FPGAs) are key but face challenges in design time and verification.
• Conventional FPGA design is time-consuming and requires extensive verification and debugging.

Purpose of the Study:

• To propose an optimized FPGA-based soft-core processor approach for efficient image processing.
• To reduce the design time and effort associated with FPGA implementations.
• To achieve high-quality performance comparable to dedicated FPGA designs.

Main Methods:

• Utilizing optimized FPGA-based soft-core processors.
• Exploiting task and data level parallelism.
• Developing a streamlined design flow for programming the processor structure.

Main Results:

• Achieved high performance for image processing tasks.
• Demonstrated a Histogram of Gradients algorithm implementation at 328 frames per second (fps).
• Successfully avoided the lengthy design, verification, and debugging cycles of traditional FPGA methods.

Conclusions:

• Optimized FPGA-based soft-core processors offer a viable solution for efficient image processing in surveillance.
• This approach balances high performance with reduced design complexity and time.
• The methodology enables faster development cycles for embedded image processing applications.