An average concentration value is detected within a measurement area (about 10 dots) corresponding to the pitch of a detection pattern in the total color matching adjustment range (100 dots), and from a value detected as a local maximum value, a candidate for correction value for the exposure timing of an adjustment image is acquired within the color matching adjustment range. This process is carried out for two types of detection patterns having different pitches thereby to acquire two candidates. A correction value for the exposure timing is determined from a candidate shared by the two detection patterns. At the time of color matching adjustment, the wasteful consumption of the developer or the like is suppressed, while carrying out the color matching adjustment efficiently within a short time.