Surface observation apparatus and surface observation method

A surface observation apparatus is achieved, which enables even a beginner to easily select an optimal evaluation indicator for each of various patterns to be evaluated without a trial and error approach. A plurality of images to be evaluated are input from an image processing unit (114) to an evaluation image input unit (113a) (in step 901). The input images to be evaluated are displayed on a display (115). A user rearranges the images in accordance with an evaluation criterion of the user while referencing the display (115), and defines an evaluation criterion (in step 902). Evaluation values are calculated for the input images (to be evaluated) using a plurality of evaluation indicators (in step 903). The evaluation values for each of the evaluation indicators are compared with the evaluation criterion defined by the user, and correlation coefficients are then calculated (in step 904). An evaluation indicator having the maximum absolute value of a correlation coefficient is automatically selected as an evaluation criterion that is closest to the evaluation criterion defined by the user (in step 905). The images are rearranged in order of correlation coefficient and a list of the images is displayed on the display (115) so that the images are arranged in order of evaluation value (in step 906).