Cutoff due to pointwise degradations in color images

Many studies analyze resolution limits in single-channel, pan-chromatic systems. However, color imaging is popular. Thus, there is a need for its modeling in terms of resolving capacity under noise. This work analyzes the probability of resolving details as a function of spatial frequency in color imaging. The analysis introduces theoretical bounds for performance, using optimal linear filtering and fusion operations. The work focuses on resolution loss caused strictly by noise, without the presence of imaging blur. It applies to full-field color systems, which do not compromise resolution by spatial multiplexing. The framework allows us to assess and optimize the ability of an imaging system to distinguish an object of given size and color under image noise.