The Study of Adaptive Multi Threshold Segmentation Method for Apple Fruit Based on the Fractal Characteristics

According to the characteristics of the fruit and vegetable mechanization picking working environment, a segmentation method for apple fruit based on fractal characteristics is proposed in the paper. Firstly, the image is converted from RGB color space to I₁I₂I₃ color space. Secondly, the image is divided into several regions and the texture features of every region are described by using the fractal theory. Finally, the regional segmentation threshold is determined according to the fractal dimension. Thus the fractal theory and the segmentation method of adaptive multi threshold have been combined to realize the accurate segmentation in the natural working environment of the apple fruit and its background. The experimental verification showed that the method has own stronger anti-noise ability. The method can be applied to precisely and effectively segment fruits from the background in the natural working environment. By comparing with the method of traditional local threshold segmentation and threshold iteration segmentation, the method proposed in the paper has higher processing speed and better segmentation effect at the background with different complexity, and can meet the real-time request of mechanization picking operations.