Jerzy Stefanowicz
supervisor: Krzysztof Poźniak
The industrial Ph.D. program in which Jerzy Stefanowicz participates is conducted in cooperation between the Finnish-Polish synthetic aperture radar (SAR) satellite manufacturer ICEYE and the Warsaw University of Technology. The Ph.D. candidate's thesis subject revolves around the idea of putting artificial intelligence onboard ICEYE'S satellites. The main idea explored by Jerzy is utilizing the powerful Field-Programmable Gate Array (FPGA) chip used for digital signal processing in ICEYE's SAR to automatically detect ships in radar imagery. Historically, this task was performed by adaptive thresholding algorithms, which utilize local statistics of the SAR image to calculate the detection threshold. Currently, following the development of efficient deep learning methods for optical image object recognition, convolutional neural networks are the state-of-the-art method of ship detection in SAR imagery. Additionally, FPGAs are very well suited for the implementation of such data processing architectures.
During the presentation, the candidate will explain the topic in a broader context, share the progress of the work, and discuss the next steps to take to achieve the final goal.