Mechanical part industrial quality inspection method based on machine learning

The invention discloses a mechanical part industrial quality inspection method based on machine learning, and belongs to the technical field of image processing, and the method comprises the following steps: S1, data collection and labeling; s2, data preprocessing; s3, data expansion; s4, extracting data features; s5, training a classifier; through a data expansion algorithm, an artificial ant colony algorithm, a graph neural network algorithm and an improved classifier algorithm, the model can better adapt to different types of parts and quality inspection scenes, and key features such as shapes and textures of the parts are accurately captured. Accurate input is provided for a subsequent classifier, and a negative feedback algorithm is introduced, so that a classification boundary can be adaptively adjusted, and the robustness and adaptability of the classifier are improved. 本发明公开了一种基于机器学习的机械零件工业质检方法，属于图像处理技术领域，包括以下步骤：S1、数据采集与标注；S2、数据预处理；S3、数据扩充；S4、数据特征提取；S5、训练分类器；本发明通过数据扩充算法、人工蚁群算法和图神经网络算法以及改进的分类器算法，使模型能够更