Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Detection Fire in Camera RGB-NIR

Nguyen Truong Khai, Luong Duc Vinh

TL;DR

This work tackles nighttime fire detection using RGB-NIR imagery by addressing data scarcity and false positives from bright lights. It introduces a two-stage detection pipeline (YOLOv11n for detection and EfficientNetV2-B0 for classification) augmented with Laplacian edge enhancement, and a patch-based Patched-YOLO to improve small/distant fire detection. Two new datasets are developed: Dataset Detection NIR for detection and Datasets Classify Sobel for edge-based classification, with augmentation via Fusion-GAN and grayscale conversion to expand training data. The approach achieves improved night-time fire detection accuracy and reduced false positives, demonstrating practical potential for safer surveillance and rapid response, while identifying remaining challenges for distant, small fires and blurred imagery.

Abstract

Improving the accuracy of fire detection using infrared night vision cameras remains a challenging task. Previous studies have reported strong performance with popular detection models. For example, YOLOv7 achieved an mAP50-95 of 0.51 using an input image size of 640 x 1280, RT-DETR reached an mAP50-95 of 0.65 with an image size of 640 x 640, and YOLOv9 obtained an mAP50-95 of 0.598 at the same resolution. Despite these results, limitations in dataset construction continue to cause issues, particularly the frequent misclassification of bright artificial lights as fire. This report presents three main contributions: an additional NIR dataset, a two-stage detection model, and Patched-YOLO. First, to address data scarcity, we explore and apply various data augmentation strategies for both the NIR dataset and the classification dataset. Second, to improve night-time fire detection accuracy while reducing false positives caused by artificial lights, we propose a two-stage pipeline combining YOLOv11 and EfficientNetV2-B0. The proposed approach achieves higher detection accuracy compared to previous methods, particularly for night-time fire detection. Third, to improve fire detection in RGB images, especially for small and distant objects, we introduce Patched-YOLO, which enhances the model's detection capability through patch-based processing. Further details of these contributions are discussed in the following sections.

Detection Fire in Camera RGB-NIR

TL;DR

This work tackles nighttime fire detection using RGB-NIR imagery by addressing data scarcity and false positives from bright lights. It introduces a two-stage detection pipeline (YOLOv11n for detection and EfficientNetV2-B0 for classification) augmented with Laplacian edge enhancement, and a patch-based Patched-YOLO to improve small/distant fire detection. Two new datasets are developed: Dataset Detection NIR for detection and Datasets Classify Sobel for edge-based classification, with augmentation via Fusion-GAN and grayscale conversion to expand training data. The approach achieves improved night-time fire detection accuracy and reduced false positives, demonstrating practical potential for safer surveillance and rapid response, while identifying remaining challenges for distant, small fires and blurred imagery.

Abstract

Improving the accuracy of fire detection using infrared night vision cameras remains a challenging task. Previous studies have reported strong performance with popular detection models. For example, YOLOv7 achieved an mAP50-95 of 0.51 using an input image size of 640 x 1280, RT-DETR reached an mAP50-95 of 0.65 with an image size of 640 x 640, and YOLOv9 obtained an mAP50-95 of 0.598 at the same resolution. Despite these results, limitations in dataset construction continue to cause issues, particularly the frequent misclassification of bright artificial lights as fire. This report presents three main contributions: an additional NIR dataset, a two-stage detection model, and Patched-YOLO. First, to address data scarcity, we explore and apply various data augmentation strategies for both the NIR dataset and the classification dataset. Second, to improve night-time fire detection accuracy while reducing false positives caused by artificial lights, we propose a two-stage pipeline combining YOLOv11 and EfficientNetV2-B0. The proposed approach achieves higher detection accuracy compared to previous methods, particularly for night-time fire detection. Third, to improve fire detection in RGB images, especially for small and distant objects, we introduce Patched-YOLO, which enhances the model's detection capability through patch-based processing. Further details of these contributions are discussed in the following sections.
Paper Structure (24 sections, 9 equations, 5 figures, 4 tables, 1 algorithm)

This paper contains 24 sections, 9 equations, 5 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background & Motivation
  3. Challenges in Nighttime Fire Detection and RGB images
  4. Our Contributions
  5. Related Work
  6. Dataset
  7. Dataset Detection NIR
  8. Datasets Classify Sobel
  9. Method
  10. Enhancing the Dataset
  11. Dataset Detection NIR
  12. Datasets Classify Sobel
  13. Model Two-stage
  14. Yolov11
  15. Efficientnetv2
  16. ...and 9 more sections

Figures (5)

  • Figure 1: Conversion of RGB images to NIR using FIRE-GAN.
  • Figure 2: NIR same as image grayscale.
  • Figure 3: (Follow processing images) This is the image transformation process I used the sequential transformations applied to the input image to enhance edge details for object.
  • Figure 4: value loss in training process.
  • Figure 5: value loss in training process.