On-Device Crack Segmentation for Edge Structural Health Monitoring

TL;DR

This paper develops lightweight U-Net models optimized for TinyML to enable efficient crack segmentation on resource-limited edge devices, advancing autonomous structural health monitoring.

Contribution

It introduces tailored lightweight U-Net architectures with optimization strategies suitable for TinyML, balancing accuracy and resource constraints in crack segmentation.

Findings

Significant reduction in memory and inference time with minimal accuracy loss.

Effective use of depthwise separable convolutions for resource efficiency.

Achieved a practical compromise between performance and resource consumption.

Abstract

Crack segmentation can play a critical role in Structural Health Monitoring (SHM) by enabling accurate identification of crack size and location, which allows to monitor structural damages over time. However, deploying deep learning models for crack segmentation on resource-constrained microcontrollers presents significant challenges due to limited memory, computational power, and energy resources. To address these challenges, this study explores lightweight U-Net architectures tailored for TinyML applications, focusing on three optimization strategies: filter number reduction, network depth reduction, and the use of Depthwise Separable Convolutions (DWConv2D). Our results demonstrate that reducing convolution kernels and network depth significantly reduces RAM and Flash requirement, and inference times, albeit with some accuracy trade-offs. Specifically, by reducing the filer number to 25%, the network depth to four blocks, and utilizing depthwise convolutions, a good compromise between segmentation performance and resource consumption is achieved. This makes the network particularly suitable for low-power TinyML applications. This study not only advances TinyML-based crack segmentation but also provides the possibility for energy-autonomous edge SHM systems.