The speed-up of detection of space debris using "InterP" and "FLOPS2D"

Abstract

A new analysis method which can detect faint geostationary (GEO) objects has been proposed. The stacking method, which uses numerous CCD frames to detect objects below the background noise level, is widely used and works well, but has the drawback that detecting objects whose movements are unknown is extremely time-consuming. To overcome this, we developed a new algorithm which uses binarization of CCD images and calculates sum values instead of medians. A hardware realization of the algorithm was developed on an emulator "InterP" in preparation for being implemented in the "FLOPS2D" modular field programmable gate array (FPGA) system. When implemented in an FPGA, the algorithm reduces the analysis time by a factor of a fifty over a software implementation of the stacking method. We propose that FPGA acceleration is very effective when used in conjunction with modification of existing method to render them suitable for FPGA implementation.