PURPOSE: To provide a picture discrimination device for automatically discriminating the kind of pictures.

CONSTITUTION: The average picture element value Ym(p) of the picture element values Y(p) of the picture elements of the respective coordinates (p) {(p)=1, 2,...} of one line in picture data stored in a storage part 11 and the picture element values of the picture elements in front and behind of the picture elements of the respective coordinates (p) is calculated in an average value calculation part 12 and the absolute value (d) of a difference between the average picture element value Ym(p) and the picture element values Y(p) of the picture elements of the respective coordinates (p) is calculated in an adder 13. The absolute value is compared with the comparison value (e) of a ROM 16 in a second comparison part 15, the respective picture element values Y(p) are compared with the upper limit value Ymax and lower limit value Ymin of the ROM 16 in a first comparison part 14 and the kind of the pictures indicated by the picture data is discriminated in a discrimination part 17 from the compared results.