Construction of Robot Computer Image Segmentation Model Based on Partial Differential Equation

Abstract

The application of partial differential equation in robot computer image smoothing and image segmentation based on curve evolution is studied to achieve the breakthrough and innovation of classical methods. First of all, the image is interpolated and enlarged, usually with double primary interpolation and double secondary interpolation. The robot computer image obtained by interpolation has a better effect on the smooth part of the image, but the edge becomes blurred. Then, the model is used to enhance the amplified image and remove the noise introduced during interpolation. In the process of PDE image segmentation, using the level set method, natural continuation cannot guarantee that the embedded function is always a signed distance function in the evolution process. The results show that compared with the classical fourth-order model, this model can well retain the details of the image, smooth bevels, and edges of the image. The proposed new model is better than the second-order model and the classical fourth-order model both in PSNR value and MAE value and in visual effect. Secondly, the model based on piecewise smooth Mumford-Shah not only expresses the original image better but also solves the problem of gradual image segmentation with gray value. Therefore, the robot computer image segmentation model based on partial differential equation can segment the image better.