“…Experimental process: image data collection, image pre‐processing, establishment of neural network algorithm model, train neural network, and accuracy verification of example experiments. Among them, image pre‐processing uses Wiener filtering to suppress image noise (Jamaludin et al, 2021; Park et al, 2020; Brown et al, 2019), and blur enhancement to improve image contrast (Li et al, 2020; Golestan et al, 2014; Cao et al, 2008; Li et al, 2011), and then extract texture and fractal features from the pre‐processed image, and finally input artificial model. The segmentation algorithm function was used to train the image neural network model (Saha et al, 2011; Kemnitz & Eckstein, 2017).…”