Using spatially calibrated images of fast visible cameras, a database was established containing pellet cloud images and the related pellet- and plasma parameters. Using this database, two scalings were derived for the cloud size along the magnetic field lines as a function of pellet speed and ablation rate (first case) and pellet speed, pellet volume, plasma temperature and plasma density (second case). Using the images—based on the number of radiation maxima—the four main cloud shapes were also categorized. The isotope effect (the effect of hydrogen pellets in hydrogen or helium plasma) was also investigated with particular attention devoted to the cloud characteristics. Finally, a synthetic diagnostic—which simulates the measurement system and produces a synthetic pellet cloud image based on the output of the pellet cloud simulation—was developed to reveal the underlying density- and temperature distributions of the observed pellet cloud images. Using this synthetic diagnostic, one of the main identified cloud shapes was reconstructed. Our goal is to derive a scaling law for the toroidal extension of the pellet cloud at different pellet- and plasma conditions, to give a more reliable input for the pellet ELM triggering simulations and using these two results—a better understanding of the pellet-caused pressure perturbation.