Pedestrian Detection in Infrared Thermal Images Based on Raised Cosine Distribution

Abstract

We propose a simple and fast image thresholding approach for detecting pedestrians in an infrared thermal image. The approach uses the raised cosine distribution function as an analogous function to the one-dimensional histogram of the infrared thermal image. Experiments are conducted on the infrared thermal images gathered from the standard infrared image datasets to describe the performance of the proposed method. The performance is evaluated by comparing the results with state-of-art methods.