Low Resolution Person Detection with a Moving Thermal Infrared Camera by Hot Spot Classification

In many visual surveillance applications the task of person detection and localization can be solved easier by using thermal long-wave infrared (LWIR) cameras which are less affected by changing illumination or background texture than visual-optical cameras. Especially in outdoor scenes where usually only few hot spots appear in thermal infrared imagery, humans can be detected more reliably due to their prominent infrared signature. We propose a two-stage person recognition approach for LWIR images: (1) the application of Maximally Stable Extremal Regions (MSER) to detect hot spots instead of background subtraction or sliding window and (2) the verification of the detected hot spots using a Discrete Cosine Transform (DCT) based descriptor and a modified Random Naïve Bayes (RNB) classifier. The main contributions are the novel modified RNB classifier and the generality of our method. We achieve high detection rates for several different LWIR datasets with low resolution videos in real-time. While many papers in this topic are dealing with strong constraints such as considering only one dataset, assuming a stationary camera, or detecting only moving persons, we aim at avoiding such constraints to make our approach applicable with moving platforms such as Unmanned Ground Vehicles (UGV).

1. Introduction

Person detection and localization is an important part of many camera-based safety and security applications such as search and rescue, surveillance, reconnaissance, or driver assistance. However, achieving a high rate of correct detections with only few false positives or negatives at the same time is still a challenge due to low resolution, changing background, or runtime requirements. Furthermore, when visual-optical cameras are used, strong variation in illumi-nation, background, and human appearance complicate the problem even more, which leads to complex solutions using very high dimensional feature spaces to find and select the few discriminative features among them [4], [19], 3[4]. Thermal long-wave infrared (LWIR) cameras can provide a better fundament for person detection especially in complex outdoor scenarios with masking background texture or lack of illumination. In such scenarios the thermal signature of persons is more prominent compared to the visual-optical signature [16]. An example coming from the Object Tracking and Classification Beyond the Visible Spectrum (OTCBVS) dataset OSU Color-Thermal Database [9] is shown in Fig. 1. Although there is some variation in thermal signatures, too, this variation is smaller even across different cameras and datasets compared to visual-optical images. Gradient-based methods such as HOGs [7] try to normalize visual-optical signatures but this leads again to complex approaches when aiming to detect humans reliably in low resolution [4], [34].