Detection and tracking of dim objects in infrared (IR) images using Support Vector Machine

Detecting and tracking dim small targets in infrared images and videos are very important in the field of image processing, such as video surveillance and infrared imaging precise guidance. Recently, Support Vector Machine (SVM) based algorithms have been proposed to detect and track the infrared dim small targets. In general, SVM are based on the concept of decision planes that define boundaries between two object categories. A decision plane is one that separates between a set of objects having different class memberships. This paper presents a novel method of detection and tracking of dim objects in infrared (IR) images using SVM. The method starts with a Retinex filtering to remove the noise and improve the image quality. In the second step, the object in infrared image is detected through the support vector machine classifier from the background. Discrete Wavelet Transform (DWT), mean, entropy, and variance are four important features used by SVM classifier to separate the objects from the background. Experimental results of infrared images with different types of backgrounds show the efficiency and accuracy of the proposed method in detecting and tracking the dim small targets.