SO₂ camera systems are increasingly being used to image volcanic gas plumes. The ability to derive SO₂ emission rates directly from the acquired imagery at high time resolution allows volcanic process studies that incorporate other high time-resolution datasets. Though the general principles behind the SO₂ camera have remained the same for a number of years, recent advances in CCD technology and an improved understanding of the physics behind the measurements have driven a continuous evolution of the camera systems. Here we present an intercomparison of seven different SO₂ cameras. In the first part of the experiment, the various technical designs are compared and the advantages and drawbacks of individual design options are considered. Though the ideal design was found to be dependent on the specific application, a number of general recommendations are made. Next, a time series of images recorded by all instruments at Stromboli Volcano (Italy) is compared. All instruments were easily able to capture SO₂ clouds emitted from the summit vents. Quantitative comparison of the SO₂ load in an individual cloud yielded an intra-instrument precision of about 12%. From the imagery, emission rates were then derived according to each group's standard retrieval process. A daily average SO₂ emission rate of 61 ± 10 t/d was calculated. Due to differences in spatial integration methods and plume velocity determination, the time-dependent progression of SO₂ emissions varied significantly among the individual systems. However, integration over distinct degassing events yielded comparable SO₂ masses. Based on the intercomparison data, we find an approximate 1-sigma precision of 20% for the emission rates derived from the various SO₂ cameras. Though it may still be improved in the future, this is currently within the typical accuracy of the measurement and is considered sufficient for most applications.