Low-Contrast Coating Surface Microcrack Detection Using an Improved U-Net Network Based on Probability Map Fusion

TL;DR

A new U-Net-based method improves microcrack detection on low-contrast surfaces, achieving high accuracy and robustness in aerospace and energy applications.

Contribution

The novel ImF-Att-Do-U-net method combines preprocessing and network improvements to detect microcracks as small as 7 μm with high reliability.

Findings

The method achieved a Dice coefficient of 0.884, SSIM of 0.893, and accuracy of 0.911, outperforming existing models.

It detected cracks as small as 7 μm with a recognition rate >85% and had a 98% extraction rate for cracks ≥10 μm.

Multi-stage preprocessing and optimized network modules improved anti-interference ability under noise and blur.

Abstract

What are the main findings?
The proposed ImF-Att-Do-U-net method achieved excellent detection performance—a Dice coefficient of 0.884, SSIM of 0.893, and accuracy of 0.911—outperforming comparative models. It detected cracks as small as 7 μm (recognition rate >85%) and had a 98% extraction rate for cracks ≥10 μm.The multi-stage preprocessing (contrast enhanced by over four times) and optimized network modules (CBAM, DO-Conv, and Leaky ReLU) enhanced anti-interference ability, maintaining stable performance under noise and blur.

The proposed ImF-Att-Do-U-net method achieved excellent detection performance—a Dice coefficient of 0.884, SSIM of 0.893, and accuracy of 0.911—outperforming comparative models. It detected cracks as small as 7 μm (recognition rate >85%) and had a 98% extraction rate for cracks ≥10 μm.

The multi-stage preprocessing (contrast enhanced by over four times) and optimized network modules (CBAM, DO-Conv, and Leaky ReLU) enhanced anti-interference ability, maintaining stable performance under noise and blur.

What are the implications of the main findings?
Provides a reliable solution for coating microcrack detection in aerospace and new energy fields.Innovations in preprocessing, network, and loss function inform low-contrast microdefect detection.

Provides a reliable solution for coating microcrack detection in aerospace and new energy fields.

Innovations in preprocessing, network, and loss function inform low-contrast microdefect detection.

To address challenges such as low contrast, complex backgrounds, and discontinuous crack distribution in coating surface microcrack detection, a detection method combining circular neighborhood features with an improved U-net is proposed. In the preprocessing stage, a background template is constructed via median filtering, and crack contrast is enhanced through a combination of difference operations and Gaussian smoothing. Based on the spatial aggregation and directionality of crack pixels, multi-scale and multi-directional circular scanning filters were constructed to generate neighborhood difference maps for quantifying the crack distribution probability. The ImF-Att-DO-U-net was designed by utilizing a dual-channel input consisting of the original image and the crack probability map. The encoder embeds lightweight CBAMs to strengthen crack features, while the decoder introduces DO-Conv and Leaky ReLU to enhance detail capture capabilities. A hybrid loss function combining Binary Cross-Entropy and Dice loss was employed to optimize class imbalance. Algorithm testing results demonstrate that the proposed method achieved a Dice coefficient of 0.884, an SSIM of 0.893, and an accuracy of 0.911, outperforming comparative models such as DO-U-net. The extraction rate for cracks ≥10 μm reached 98%, with a minimum detectable crack size at the 7 μm level. The method exhibited excellent robustness under noise and blur testing, demonstrating superior environmental adaptability.

Figures