Discrimination of Strawberry Class Using 3D Image Processing

Three-dimensional shape features were examined for the distinction of the strawberry shape. First, the three-dimensional coordinates were calculated using a three-dimensional shape entry machine. From the coordinates, eight parameters were proposed. The classification was not adequate when using only one parameter. So, several parameters were selected using step wise decision. The Bayes' estimate and neural network were performed according to the selected parameters. As a result, the agreement percentage with the judgement by the act of man exceeded 90%. [Keywords] strawberry, image processing, 3D image, canonical discrimination, Bayes' theorem, neural network