We have characterised the Perseid meteoroid stream from data acquired in a series of observing campaigns between 2010 and 2016. The data presented in this work were obtained by the Smart Panoramic Optical Sensor Head (SPOSH), an all-sky camera system designed to image faint transient noctilucent phenomena on dark planetary hemispheres. For the data reduction, a sophisticated software package was developed that utilises the high geometric and photometric quality of images obtained by the camera system. We identify 934 meteors as Perseids, observed over a long period between late July (∼124 • ) and mid-to-late August (∼147 • ). The maximum meteor activity contributing to the annual shower was found at λ = 140. • 08 ± 0. • 07. The radiant of the shower was estimated at RA = 47. • 2 and Dec = 57. • 5 with a median error of 0. • 6 and 0. • 2, respectively. The mean population index of the shower between solar longitudes of 120. • 68 and 145. • 19 was r = 2.36 ± 0.05, showing strong temporal variation. A predicted outburst in shower activity for the night of August 11-12, 2016 was confirmed, with a peak observed 12.75 hr before the annual maximum at 23:30 ± 15 UT. We measure a peak flux of 6.1 × 10 −4 km −2 hr −1 for meteoroids of mass 1.6 × 10 −2 g or more, appearing in the time period between 23:00 and 00:00 UT. We estimate the measured flux of the outburst meteoroids to be approximately twice as high as the annual meteoroid flux of the same mass. The population index of r = 2.19 ± 0.08, computed from the outburst Perseids in 2016, is higher than the value of r = 1.92 ± 0.06 derived from meteors observed in 2015 belonging to the annual Perseid shower which was active near the time of the outburst. A dust trail with an unusually high population index of r = 3.58 ± 0.24 was encountered in 2013 between solar longitudes 136. • 261 and 137. • 442. The relatively high r-value implies an encounter with a dust trail rich in low-mass particles.