A promising method for detection of parasites in whitefish fillets has been developed. By use of imaging spectroscopy it is possible to record both spectral and spatial information from an object. In this work it is shown that by applying a white light transmission setup and imaging spectroscopy to cod (Gadus morhua) fillets, it is possible to make spectral images containing information to differentiate between fish muscle and parasites. The spectral images are analyzed by discriminant partial least square regression as well as image-filtering techniques. The method identifies parasites on the surface of the fillets as well as embedded parasites. One parasite was detected at 0.8 cm below the fillet surface, which is 2 to 3 mm deeper than what can be found by manual inspection of fish fillets. The method is nonintrusive and should thus be feasible for industrial purposes.