Machine vision systems are used in diverse kinds of industries. They are employed in automobiles for the detection and protection of pedestrians, in visual inspection systems to ensure the quality of produced goods
etc. The parameters of such systems result in many degrees of freedom. Determining the optimal setting of these parameters usually represents a time-consuming task that has to be performed empirically. This article introduces
a software framework consisting of multiple plugins for the physically based rendering suite Mitsuba. By means of combining the single framework components, the sensor signals of arbitrary machine vision systems
can be simulated for synthetic scenes consisting of objects having complex 3D geometries and the design process of the whole system can be streamlined.