| 陈克琼,卓士虎,赵晨曦,傅立涛,王家铭,李帷韬.融合深度迁移学习和改进ThunderNet的 瓷砖表面缺陷检测[J].电子测量与仪器学报,2024,38(3):208-218 |
| 融合深度迁移学习和改进ThunderNet的 瓷砖表面缺陷检测 |
| Integrated deep transfer learning and improved ThunderNetin tile surface defect detection |
| |
| DOI： |
| 中文关键词: 瓷砖表面缺陷检测 可切换空洞卷积 迁移学习 通道注意力 特征融合 小样本 |
| 英文关键词:tile surface defect detection switchable atrous convolution transfer learning channel attention feature fusion few-shot |
| 基金项目:国家自然科学基金（62173120）、安徽省自然科学基金青年项目（1908085QF270）、安徽省高等学校科学研究项目（自然科学类）重点项目（2022AH051793）资助 |
|
|
| Author | Institution |
| Chen Keqiong | School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, China |
| Zhuo Shihu | School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, China |
| Zhao Chenxi | School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, China |
| Fu Litao | School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, China |
| Wang Jiaming | School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, China |
| Li Weitao | School of Electrical Engineering and Automation, Hefei University of Technology, Hefei 230009, China |
|
| 摘要点击次数: 386 |
| 全文下载次数: 2226 |
| 中文摘要: |
| 瓷砖生产过程中由于环境的复杂性和随机性导致缺陷特性各异，实际中要构建大规模、高质量的瓷砖表面缺陷数据样本非常困难，而小样本条件下的可分特征信息不足对瓷砖表面缺陷检测的精度有较大影响。针对这一问题，探索了一种融合深度迁移学习和改进两阶段ThunderNet网络的瓷砖表面缺陷检测方法。首先，提出了一种基于改进ThunderNet网络的瓷砖表面缺陷检测模型，阐述了模型的结构与功能特点；其次，构造了瓷砖表面缺陷深度特征空间参数迁移决策机制，以有效提升样本特征表征能力；第三，基于可切换空洞卷积（switchable atrous convolution, SAC）优化ShuffleNet骨干网络，增强模型对缺陷形状变化的学习能力；第四，提出了基于多尺度映射和通道注意力（squeeze and excitation, SE）的特征融合算法，实现有限特征层次中瓷砖表面缺陷特征信息多层次差异化表征；最后，给出了融合深度迁移学习和改进ThunderNet网络的瓷砖表面缺陷检测算法。实验数据表明，在相同的瓷砖表面缺陷测试集上，本文方法对于小样本条件下瓷砖表面缺陷检测具有优越的性能，模型平均精度、平均召回率和平均检测速度分别达到87.22%、93.69%、61.6 ms/img，与传统ThunderNet模型相比，平均精度和平均召回率分别提高了9.30%、4.16%，其中，基于SAC最优空洞率组合{1,2}，模型精度提高了5.51%，基于SE的最优压缩率24，模型精度提高了6.16%，基于本文迁移机制，模型精度提高了3.86%，同时加速了网络收敛。本文方法相比于传统ThunderNet网络和其他主流检测模型，通过迁移机制知识共享提高小样本对象特征表达能力，通过引入SAC和SE在控制模型规模的前提下实现对象特征的层次化表征，有效提升了模型的实时性和可靠性。 |
| 英文摘要: |
| Due to the complexity and randomness of the environment in the production process of ceramic tiles, it is very difficult to construct large-scale and high-quality ceramic tile surface defect data samples, and the insufficient distinguishable feature information under few-shot conditions has a great impact on the accuracy of ceramic tile surface defect detection. To solve this problem, a tile surface defect detection method based on deep transfer learning and improved two-stage ThunderNet network is explored. Firstly, a tile surface defect detection model based on the improved ThunderNet network is proposed, and the structure and functional characteristics of the model are elaborated. Secondly, the decision-making mechanism for spatial parameter transfer of tile surface defect depth features is constructed to effectively improve the characterization ability of sample feature. Third, the ShuffleNet backbone network is optimized based on Switchable Atrous Convolution (SAC) to enhance the model’s learning ability to the changeable shape of defects. Fourth, a feature fusion algorithm based on multi-scale mapping and squeeze and excitation (SE) is proposed to realize the multi-level differentiated characterization of tile surface defect feature information in a limited feature level. Finally, a tile surface defect detection algorithm for fusion deep transfer learning and improved ThunderNet network is given. The experimental data show that on the same tile surface defect test dataset, the proposed method has superior performance for the detection of tile surface defects under few-shot conditions, and the average accuracy, average recall and average detection speed of the model reach 87.22%, 93.69% and
61.6 ms/img, respectively, compared with the traditional ThunderNet model, the average accuracy and average recall are improved by 9.30% and 4.16%, respectively, among which, based on the SAC optimal atrous ratio combination {1,2}, The model accuracy is improved by 5.51%, the model accuracy is improved by 6.16% based on the optimal compression ratio of SE 24, and the model accuracy is improved by 3.86% based on the transfer mechanism in this paper, and the network convergence is accelerated. Compared with the traditional ThunderNet network and other mainstream detection models, the proposed method improves the expression ability of few-shot object features through knowledge sharing of transfer mechanism, and realizes hierarchical representation of object features by introducing SAC and SE under the premise of controlling the scale of the model, which effectively improves the real-time reliability and reliability of the model. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
