The aim of the thesis was to design and implement a preprocessing system for tomographic data in FPGA devices. The system has been designed with high parallelism of calculations in mind, thanks to which it works in real-time. Additionally, applications were prepared for live-visualization of the results (i.e. Lines of Response). The paper describes the architecture of the CM-PET scanner (with which the implemented system has been integrated) and the technologies used. In the paper results of tests with the use of four CM-PET?s modules are presented. The estimated maximum system throughput is around 14 MSps.