Peregrin is a machine vision engine. Unlike software-based approaches, Peregrin uniquely merges customizable hardware accelleration with application software. Peregrin provides a complete Open-CV deployment environment in which hardware-accelerated computation is made available to software simply by reading from a device file.
Peregrin includes options for hardware-based real-time image pre-processing, including pixel defect correction, noise reduction, resampling, windowing, Bayer pattern processing and color space conversion.
Feature detection is application dependent. Dimensional measurement, color sorting, pose determination, and character recognition each have different algorithm requirements. Peregrin can leverage the Xilinx reVision stack or the Pixel Machines Algorithm Graph Compiler (available 1Q 2018) to accelerate feature detection computation by 10-20x. Users with pre-existing Zynq-based accellerators will find it straightforward to port their work to Peregrin.
Feature extraction is the process of converting information in the image domain (pixels) into information in the symbolic domain (lists) for processing by the CPU. The advantage of hardware-accelerated feature extraction is that the CPU is not involved in the time-consuming and high latency process of connected component analysis and pixel coordinate extraction. Instead, data structures are read directly from a hardware device.