Surface Defect Measurement System of Die Castings Based on Deep Learning

Abstract

The invention discloses a deep learning-based die casting surface defect measurement system, which relates to the field of deep learning technology. The system includes a data storage module, a graphics processing flow module and a defect location module; the graphics processing flow module uses an industrial camera to monitor the die casting Scanning, obtaining two-dimensional images of different sections to establish a three-dimensional model, and judging whether there are defects on the surface of the die-casting; Compare the data in the module to determine whether the size and position of the defect on the surface of the die-casting part affect the rear process work, and when the degree of influence is high, give an early warning; the data storage module sets a number of training set data and updates the training set in real time; The deep learning algorithm is used to train the data to judge the defect degree of die castings, which improves the detection rate and improves the accuracy of judging the defects of die castings.

Description

technical field [0001] The invention relates to the field of deep learning technology, in particular to a deep learning-based surface defect measurement system for die castings. Background technique [0002] Die casting is a kind of part, which is manufactured as a mechanical die casting machine through the pressure of the mold. Through a series of heating of the metal, it is poured into the inlet to form parts of different shapes. This part is also called die casting; but after the processing is completed Different types of defects will be formed on the die castings, and it is necessary to measure and judge the defects to determine whether they are within the error range; [0003] In the existing technology, the defect measurement is judged by manual measurement. The method used to judge whether the die casting is qualified is relatively subjective, and the judgment speed is relatively slow. The speed cannot be improved and the labor cost is increased; the method of machine...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, the defect measurement is judged by manual measurement. The method used to judge whether the die casting is qualified is relatively subjective, and the judgment speed is relatively slow. The speed cannot be increased and the labor cost is increased; the method of machine learning is used to judge the quality of the die casting. , Compared with the use of manual labor, the speed of judging the quality of die castings is improved, but because machine learning takes a long time, it is not as efficient as using deep learning;

[0004] The current market uses 3D models to judge defects on die castings. Through 3D models, only a few faces of die castings can be identified to contain defects, but they do not know the size of the defects on different faces and whether they are within the standard error range. Therefore, the real situation of the defect cannot be known only by the judgment of the 3D model, and it is impossible to further analyze whether the die casting can still be used in the next process

Images