Double yolk nondestructive identification system based on Raspberry Pi and computer vision

The nutritional and commercial value of double yolk eggs is higher than that of single yolk eggs, but the traditional manual identification of double yolk eggs is costly, inefficient and has high error. For rapid and accurate identification of double yolk eggs, a portable double yolk nondestructive identification system and an online double yolk nondestructive identification system are proposed in this paper. Two systems are based on Raspberry Pi, a low-cost embedded computer. First, Raspberry Pi controls camera to capture translucent image of duck eggs, then image processing algorithm is used to remove image noise and enhance image features. The trained convolutional neural network model is used to identify duck egg images. Identification results can be preserved, displayed, and prompted by Raspberry Pi. The model identifies double yolk eggs with 100% accuracy over a calculation time of 0.07 s.