We studied a new analytical method for determining self-excited chatter vibration at an end mill processing site. When chatter vibration occurs, cutting marks with periodic features called “chatter marks” are generated on the processed surface. We focused on the relationship between the vibration generated during cutting and the roughness of the processed surface. We developed a method to analyze such vibrations, based on the roughness of the finished surface. Instead of using a roughness meter and laser-measuring instrument, which are often used for measuring the contours of a processed surface, we presumed the roughness of the processed surface from image data obtained using a digital camera. As a comparative experiment to verify the proposed method, side cutting with an end mill was performed. Chatter vibrations occurring during the cutting were measured using a displacement sensor and microphone, and compared with the vibration of the tool edge as determined from the processed surface. The chatter frequency read from the processed surface was close to that obtained from the displacement sensor and microphone. Thus, the proposed method can be used to analyze the information of chatter vibration inexpensively and easily. It can serve as a substitute for displacement and acceleration sensors.