A three-dimensional image processing apparatus includes: an image capturing part for acquiring reflected light to capture a plurality of pattern projected images; a distance image generating part capable of generating a distance image based on the plurality of pattern projected images; a tone conversion part for tone-converting the distance image generated in the distance image generating part to a low-tone distance image that has a lower number of tones than the number of tones of the distance image and is obtained by replacing height information in the distance image with a shade value of the image; and a tone conversion condition automatic setting part for automatically setting, based on the height information in the distance image, a tone conversion parameter for prescribing a tone conversion condition at the time of tone-converting the distance image to the low-tone distance image in the tone conversion part.