A spectral image acquiring apparatus includes an optical filter on which light is incident; an image sensor including a two-dimensionally disposed pixel array for detecting the light via the optical filter; and a signal processing unit generating a difference-value image based on a detection signal from the image sensor. The optical filter includes a diffraction grating having a lattice pattern corresponding to one or more pixels on the image sensor. The signal processing unit calculates a difference value in an amount of received light between two adjacent pixels based on the detection signal from the image sensor, and generates the difference-value image based on the difference value. The difference value between the two adjacent pixels is varied depending on a difference in an interference point on the image sensor corresponding to a diffraction angle of the light that has passed through the diffraction grating.