Embedded system implementation of an evolutionary algorithm for circle detection on programmable devices

Programmable devices combine powerful processing systems with a rich infrastructure of general-purpose and specific logic blocks, making it possible the efficient implementation of embedded systems to perform complex tasks by facilitating hardware acceleration of critical stages to improve their performance. Based on these characteristics, a hardware implementation of a genetic algorithm for circle detection in digital images is described in this paper. The detection system has been designed for Xilinx Zynq-7000 and Zynq UltraScale+ family devices and implemented on two low-cost development boards that reach acceleration factors of 33.12 and 37.3, respectively, when compared to the fully software implementation. Detection results from both development boards have been compared using synthetic and real images from different scenarios. The accuracy and performance achieved demonstrate the suitability of this proposal to design embedded systems with restricted size, resources and energy consumption for applications in Internet of Things, Industry 4.0 and other related paradigms.