Atmospheric correction based on hyperspectral image itself was performed on the new spaceborne imaging spectrometer CHRIS image by using the popular radiance transfer code ACORN(Atmospheric CORrection Now) and empirical algorithm as well, in this way, the calibration performance of CHRIS was evaluated by the retrieved spectra of vegetation and soil. It turned out, the vegetation reflectance spectra inverted by ACORN could characterize vegetation reflectance in the range of 498~760nm, but gave a fairly large deviation beyond 760nm, showing the deficiency of spectral calibration beyond 760nm. The ACORN derived soil reflectance decreased after 760nm, which is quite inconsistent with common sense, showing that the calibration accuracy couldn't meet requirements of ACORN for spectral and radiometric calibration at certain spectral range. In addition, the stripes on the retrieved water vapor content map indicated that the radiance calibration performance needs to be improved. On the contrary, AVIRIS was validated to have better calibration performance so that more precise spectra could be retrieved by ACORN5.