Imaging　sensors　as　a　critical　technology　are　greatly　significant　in　numerous　light　detection　applications　including　security　monitor,　smart　mobile　devices,　machine　vision,　navigation,　self-driving,　drones　and　remote　sensing　etc.　In　particular,　spectral　imaging　technology　such　as　multispectral　and　hyperspectral　imaging　are　attracting　more　and　more　attention　due　to　the　available　spectral　and　spatial　information,　which　provide　an　opportunity　for　deeper　cognition　of　the　detected　object.　However,　conventional　strategy　for　wavelength　sensing　by　complicated　optical　system　has　obstacle　in　achieving　lightweight,　low　cost,　high　quality　image　and　high　pixel　densities.　Therefore,　new　emerging　strategies　based　on　the　nanomaterials　and　nanostructures　involving　filter-assisted　and　filter-free　light　detection　are　explored　to　achieve　precise　spectral　discrimination.　Here,　we　review　the　latest　advances　of　these　emerging　approaches　for　spectral　discrimination.　Furthermore,　every　filter-assisted　and　filter-free　strategy　was　discussed　and　compared　in　detail　by　analyzing　their　fundamental　physical　principles.