# Research on Synthetic Data Methods and Detection Models for Micro-Cracks

## TL;DR

This paper introduces a method to generate realistic micro-crack images and a robust detection model that works well in challenging conditions.

## Contribution

The novel contribution is a combined approach of realistic data synthesis and a complex-scene-tolerant YOLO-based model for efficient micro-crack detection.

## Key findings

- A Poisson image editing-based synthesis pipeline generates visually consistent micro-crack samples.
- CST-YOLO achieves 0.926 mAP@0.5:0.95 at 139 FPS on real images with ablation studies confirming component effectiveness.
- The method enables accurate and efficient micro-crack detection in complex scenes with real-time performance.

## Abstract

What are the main findings?

A Poisson image editing-based synthesis pipeline produces visually consistent micro-crack samples and improves data availability without degrading realism.

The proposed CST-YOLO (YOLOv10 + LAPM + SCS-Former + SOFB) achieves 0.926 mAP@0.5:0.95 at 139 FPS on real-image evaluation, with complementary gains verified by ablations.

What are the implications of the main findings?

The combined “data realism + architecture robustness” strategy enables accurate and efficient micro-crack detection under challenging illumination, shadow, and clutter conditions.

The real-time performance supports practical deployment in large-scale infrastructure inspection scenarios that require both precision and throughput.

Micro-crack detection on concrete surfaces is challenging because labeled micro-crack data are scarce, crack cues are extremely weak (often only a few pixels wide), and complex backgrounds (e.g., non-uniform illumination, shadows, and stains) degrade feature extraction; this study aims to improve both data availability and detection robustness for practical inspection. A Poisson image editing-based synthesis strategy is developed to generate visually coherent micro-crack samples via gradient-domain blending, and a Complex-Scene-Tolerant YOLO (CST-YOLO) detector is proposed on top of YOLOv10, following an “lighting decoupling–global perception–micro-feature enhancement” design. CST-YOLO integrates an Lighting-Adaptive Preprocessing Module (LAPM) to suppress illumination/shadow perturbations, a Spatial–Channel Sparse Transformer (SCS-Former) to model long-range crack topology efficiently, and a Small Object Focus Block (SOFB) to enhance micro-scale cues under cluttered backgrounds. Experiments are conducted on a 650-image dataset (200 real and 450 synthesized), in which synthesized samples are used only for training, and the validation/test sets contain only real images, with a 7:2:1 split. CST-YOLO achieves 0.990 mAP@0.5 and 0.926 mAP@0.5:0.95 at 139 FPS, and ablation results indicate complementary contributions from LAPM, SCS-Former, and SOFB. These results support the effectiveness of combining realistic synthesis and architecture-level robustness for real-time micro-crack detection in complex scenes.