High speed FPGA implementation of hough transform for real-time applications

Hough Transform (HT) is a popular line detection algorithm in image processing and machine vision applications, favored for its tolerance to noise and partial occlusion. However, due to its computational complexity, software and hardware implementations for real-time video processing are usually limited to low resolutions and frame rates. We propose a novel architecture that exploits modern FPGA inherent parallelism capabilities, combined with efficient resource utilization and deep pipelining to enable real-time processing of high resolution, high frame rate videos.