Electronic Information Image Processing Based on Convolutional Neural Networks

In order to improve the accuracy and efficiency of mechanical part recognition, the author proposes a research on electronic information image processing based on convolutional neural networks. The author first performs saturation based grayscale processing on the image; Then, the binary image is obtained through significance enhancement, binarization using the Maximum Between Class Variance (OTSU) method, and morphological closure operation; Extract the part area using an improved seed filling method; Finally, the parts are identified by combining Scale Invariant Feature Transform (SIFT) features of image keypoints with Convolutional Neural Network (CNN) models. The experimental results show that the accuracy of the part recognition algorithm can reach 98.84%, and the recognition speed is about 5fps. Conclusion: Through experimental comparison and analysis, it has been proven that this method is fast and effective, with high accuracy and good robustness.