An automobile apparent mass detection method based on a generative adversarial network

Abstract

The invention relates to an automobile apparent quality detection method based on a generative adversarial network. The method comprises the following steps: acquiring apparent defect image data suchas surface pit packages, scratches, geometric dimension gaps and surface differences of offline vehicles; in consideration of the existence of shelters such as mud points, rain points and oil stains on the surface of the actual defect, performing random binary mask equivalent processing on the acquired defect data to complete an image restoration task based on unsupervised learning; the repaired defect data and the unprocessed defect data are used for training the generative adversarial network, and a defect recognition and classification task based on semi-supervised learning is completed; and optimizing the weight parameters in the generative adversarial network training process by using an interval optimization algorithm. The network structure of the generative adversarial network provided by the invention is used for unsupervised authenticity discrimination to assist in completing an image restoration task. And finally, defect shielding object removal can be carried out simultaneously to complete image restoration and automobile apparent defect identification and classification multitask.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to an automobile appearance quality detection method based on a generative confrontation network. Background technique [0002] With the improvement of living standards, people's criteria for judging whether a car is good or bad is no longer just about functional use, but also about an aesthetic feeling of appearance. Whether the appearance is qualified or not directly affects consumers' desire to buy. More and more car manufacturers are trying to improve the quality of their car's exterior surface. Appearance defects mainly include the following types: vehicle surface pits, scratches, apparent geometric gaps or surface differences, etc. Once these defects appear, the designer must re-finish or return the mold to the manufacturer for repair. Either way, it will cause a huge waste of time and cost. However, in practical applications, the collected defect images ...

AI Technical Summary

Problems solved by technology

Appearance defects mainly include the following types: vehicle surface pits, scratches, apparent geometric gaps or surface differences, etc. Once these defects appear, the designer must re-finish or return the mold to the manufacturer for repair. method, will cause a huge waste of time and cost

However, in practical applications, the collected defect images often have local obstacles such as mud spots, oil stains, and raindrops, which interfere with feature extraction and affect the accuracy of discrimination.

Images