Fisheye lens distortion correction on multicore and hardware accelerator platforms

摘要: Wide-angle (fisheye) lenses are often used in virtual reality and computer vision applications to widen the field of view conventional cameras. Those lenses, however, distort images. For most real-world video stream needs be transformed, at real-time (20 frames/sec or better), back natural-looking, central perspective space. This paper presents implementation, optimization characterization a fisheye lens distortion correction application on three platforms: conventional, homogeneous multicore processor by Intel, heterogeneous (Cell BE), an FPGA implementing automatically generated streaming accelerator. We evaluate interaction with those architectures using both high- low-level performance metrics. In macroscopic terms, we find that todays mainstream multicores not effective supporting correction, least currently commercially available core counts. Architectures, such as Cell BE FPGAs, offer necessary computational power scalability, expense significantly higher development effort. Among these platforms, only fully optimized version code running can provide realtime processing speed. general, FPGAs meet expectations performance, flexibility, low overhead. General purpose are, other hand, much easier program.