# POLD-YOLO: A Lightweight YOLO11-Based Algorithm for Insulator Defect Detection in UAV Aerial Images

## TL;DR

POLD-YOLO is a lightweight algorithm that improves the detection of small insulator defects in UAV images, offering high accuracy with low computational cost for real-time power line inspections.

## Contribution

POLD-YOLO introduces novel modules and a loss function to enhance small target detection in UAV imagery with reduced model size and complexity.

## Key findings

- POLD-YOLO achieves a state-of-the-art mAP@0.5 of 92.4% on a UAV insulator defect dataset.
- The model outperforms YOLOv5n, YOLOv8n, YOLOv10n, and YOLO11n by 3.6%, 1.6%, 1.4%, and 1.6%, respectively.
- It maintains high accuracy with only 1.55 million parameters and 3.8 GFLOPs, making it suitable for resource-constrained UAV platforms.

## Abstract

What are the main findings?
Small target detection capability is significantly enhanced in the proposed YOLO-based method through the integration of several key innovations: depthwise and full-dimensional dynamic convolutions within a C3K2-PFCGLU module, adaptive downsampling via an OD-ADown module, a lightweight shared convolutional detection head (LSCD-Head) employing global average pooling, and a Focaler-MPDIoU loss that introduces the minimum point distance to focus on different regression samples.The proposed method achieves a reduction in both parameter count and computational complexity while improving detection accuracy and robustness, particularly for small targets.

Small target detection capability is significantly enhanced in the proposed YOLO-based method through the integration of several key innovations: depthwise and full-dimensional dynamic convolutions within a C3K2-PFCGLU module, adaptive downsampling via an OD-ADown module, a lightweight shared convolutional detection head (LSCD-Head) employing global average pooling, and a Focaler-MPDIoU loss that introduces the minimum point distance to focus on different regression samples.

The proposed method achieves a reduction in both parameter count and computational complexity while improving detection accuracy and robustness, particularly for small targets.

What are the implications of the main findings?
The proposed framework is designed for detecting small targets in UAV-captured visible-light images, specifically aiming at insulator defect detection. It thus offers a practical and effective solution to a critical task in power line inspection.

The proposed framework is designed for detecting small targets in UAV-captured visible-light images, specifically aiming at insulator defect detection. It thus offers a practical and effective solution to a critical task in power line inspection.

Detecting small insulator defects in unmanned aerial vehicle (UAV) imagery remains challenging due to low resolution, complex backgrounds and scale variation, which degrade the performance of existing detectors. This study aims to develop a highly efficient and accurate model for real-time insulator defect inspection on resource-constrained UAV platforms. This paper proposes POLD-YOLO, a novel lightweight object detector based on YOLO11. The key innovations include: (1) A backbone enhanced by a PoolingFormer module and Channel-wise Gated Linear Units (CGLUs) to boost feature extraction efficiency; (2) An Omni-Dimensional Adaptive Downsampling (OD-ADown) module for multi-scale feature extraction with low complexity; (3) A Lightweight Shared Convolutional Detection Head (LSCD-Head) to minimize the number of parameters; (4) A Focaler-MPDIoU loss function to improve bounding box regression. Extensive experiments conducted on a self-built UAV insulator defect dataset show that POLD-YOLO achieves a state-of-the-art mAP@0.5 of 92.4%, outperforming YOLOv5n, YOLOv8n, YOLOv10n, and YOLO11n by 3.6%, 1.6%, 1.4%, and 1.6%, respectively. Notably, it attains this superior accuracy with only 1.55 million parameters and 3.8 GFLOPs. POLD-YOLO establishes a new Pareto front for accuracy-efficiency for onboard defect detection. It demonstrates great potential for automated power line inspection and can be extended to other real-time aerial vision tasks.