Lightweight contour detection method for simulating parallel grading processing mechanism of visual system

The invention aims to provide a lightweight contour detection method for simulating a parallel hierarchical processing mechanism of a visual system, and the method comprises the following steps: A, constructing a neural network structure which specifically comprises a coding network and a decoding network; b, inputting the original image into a neural network, performing 1 * 1 convolution, and inputting the original image into the X-type cell model and the Y-type cell model respectively; respectively carrying out 1 * 1 convolution on output results of the X-type cell model and the Y-type cell model, then carrying out addition fusion, and respectively inputting addition fusion results into a simple cell model and a decoding network; the output result of the simple cell model is input into the complex cell model and the decoding network after being subjected to 1 * 1 convolution; the output result of the complex cell model is input into a decoding network after being subjected to 1 * 1 convolution; and the fina