Detection of Defects in Adhesive Coating Based on Machine Vision
- 25 December 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 24 (4) , 5172-5185
Abstract
The integrity of adhesive coating between propellant and insulation rubber significantly affects the stability of the functional interface and the service reliability of solid engines. The high aspect ratio and unstable lighting condition of the engine make it very difficult to inspect the coating integrity with the naked eye. More and more research is focused on using machine vision to detect defects. Limited by the size of the dataset, the current machine vision accuracy can not reach very high, and only limited to one application scenario. It is challenging to adapt to other scenarios of defect detection. In this paper, a machine vision detection method for lining spraying process adhesive coating defects is proposed, which realizes the accurate identification of defects in the coating integrity. The deflection compensation of the detection device and the adaptability design between the camera and the lens are fully considered, and a transfer learning model based on residual network is built with TensorFlow to achieve high-precision classification and recognition of defects. The model is trained by incrementally reducing the freezing layer of the model and the learning rate several times. Results show that the recognition accuracy can reach 100% on the training dataset and 96.15% on the validation. This vision method provides a fast and accurate method for engine liner coating defect detection, and it is about to be used in real industrial sites. The code and other supplementary materials will be available at: https://github.com/BeihangHuiye/Detection-of-Defects-in-Adhesive-Coating-Based-on-Machine-Vision.Keywords
This publication has 48 references indexed in Scilit:
- Surface crack detection of the abradable seal coating by laser bidirectional scanning thermographyInfrared Physics & Technology, 2022
- An effective data enhancement method of deep learning for small weld data defect identificationMeasurement, 2022
- A new approach for detecting and grading blistering defect of coatings using a machine vision systemMeasurement, 2022
- Use of La2O3 with 8YSZ as thermal barrier coating and its effect on thermal cycle behavior, microstructure, mechanical properties and performance of diesel engine operated by hydrogen-algae biodiesel blendInternational Journal of Hydrogen Energy, 2022
- Self-lubricating epoxy-based composite abradable seal coating eliminating adhesive transfer via hierarchical designJournal of Materials Science & Technology, 2022
- A dissection and enhancement technique for combined damage characterisation in composite laminates using laser-line scanning thermographyComposite Structures, 2021
- RETRACTED ARTICLE: Localization and segmentation of metal cracks using deep learningJournal of Ambient Intelligence and Humanized Computing, 2020
- A fast button surface defect detection method based on Siamese network with imbalanced samplesMultimedia Tools and Applications, 2019
- RevHashNet: Perceptually de-hashing real-valued image hashes for similarity retrievalSignal Processing: Image Communication, 2018
- Surface defect saliency of magnetic tileThe Visual Computer, 2018