The maintenance of heavy haul railway lines causes considerable cost and in many cases difficulties to rail traffic. Railway infrastructure management in Finland is important because in many locations the same railway lines are used by both freight and passenger trains. Along with the mixed traffic, the cold climate presents challenges in keeping the track in good condition. There is a great demand for economical and non-destructive methods that provide continuous condition information of track structure. In recent years, the ground penetrating radar (GPR) technology for structure inspection has improved to faster systems and better data quality. This article presents the main results from a research programme that was aimed to identify different GPR-based railway ballast degradation classes and to develop a preventative maintenance planning system. An extensive reference sampling and laboratory analysis has been performed to aid in developing the GPR-based classification method for qualifying the ballast fouling. Classification is made by a fouling index, which is calculated from the frequency contents of the GPR signal. This quality classification can be utilized in the planning of undercutting programmes.