MWIRSTD: A MWIR Small Target Detection Dataset

Abstract

This paper presents a novel mid-wave infrared (MWIR) small target detection dataset (MWIRSTD) comprising 14 video sequences containing approximately 1053 images with annotated targets of three distinct classes of small objects. Captured using cooled MWIR imagers, the dataset offers a unique opportunity for researchers to develop and evaluate state-of-the-art methods for small object detection in realistic MWIR scenes. Unlike existing datasets, which primarily consist of uncooled thermal images or synthetic data with targets superimposed onto the background or vice versa, MWIRSTD provides authentic MWIR data with diverse targets and environments. Extensive experiments on various traditional methods and deep learning-based techniques for small target detection are performed on the proposed dataset, providing valuable insights into their efficacy. The dataset and code are available at https://github.com/avinres/MWIRSTD.

Figures (5)

• Figure 1: A typical frame of the proposed dataset and its mask.
• Figure 2: The ground truth of the considered frame shown in Fig. \ref{['img:frm']} with annotations for three distinct classes. Class 1 refers to cracker rockets (shown in blue), Class 2 represents debris from cracker rockets (shown in green), and Class 3 includes other small moving targets (shown in yellow).
• Figure 3: Visual results obtained using Deep learning-based methods.
• Figure 4: Visual results obtained using traditional computational methods.
• Figure 5: Failed cases for Deep learning-based models.