GDDS: A Single Domain Generalized Defect Detection Frame of Open World Scenario using Gather and Distribute Domain-shift Suppression Network

TL;DR

GDDS is a novel single-domain generalized defect detection network for photovoltaic modules that effectively handles distribution shifts in open-world scenarios, balancing accuracy and speed.

Contribution

The paper introduces GDDS, a one-stage network with modules for contextual information and global feature interaction, independent of test samples, improving defect detection under domain shifts.

Findings

GDDS outperforms state-of-the-art methods on photovoltaic defect datasets.

It achieves faster adaptation to open-world scenarios.

The model balances detection accuracy and computational efficiency.

Abstract

Efficient and intelligent surface defect detection of photovoltaic modules is crucial for improving the quality of photovoltaic modules and ensuring the reliable operation of large-scale infrastructure. However, the scenario characteristics of data distribution deviation make the construction of defect detection models for open world scenarios such as photovoltaic manufacturing and power plant inspections a challenge. Therefore, we propose the Gather and Distribute Domain shift Suppression Network (GDDS). It adopts a single domain generalized method that is completely independent of the test samples to address the problem of distribution shift. Using a one-stage network as the baseline network breaks through the limitations of traditional domain generalization methods that typically use two-stage networks. It not only balances detection accuracy and speed but also simplifies the model deployment and application process. The GDDS includes two modules: DeepSpine Module and Gather and Distribute Module. Specifically, the DeepSpine Module applies a wider range of contextual information and suppresses background style shift by acquiring and concatenating multi-scale features. The Gather and Distribute Module collects and distributes global information to achieve cross layer interactive learning of multi-scale channel features and suppress defect instance shift. Furthermore, the GDDS utilizes normalized Wasserstein distance for similarity measurement, reducing measurement errors caused by bounding box position deviations. We conducted a comprehensive evaluation of GDDS on the EL endogenous shift dataset and Photovoltaic inspection infrared image dataset. The experimental results showed that GDDS can adapt to defect detection in open world scenarios faster and better than other state-of-the-art methods.