Application of Offset Averaging Filters to enhance the tracking information for small moving targets

Object tracking has become one of significant tasks in the area of computer vision, it has been widely applied for robot navigation and localization, automated surveillance, traffic monitoring, etc. vision sensors have been a favored choice for these tasks because of the simplicity of installation, cost effectiveness and the ability to operate in unstructured environments. However, obtaining accurate tracking information via vision sensors is not an easy task. In this paper, we deal with two types of tracking problems. The first is when a low resolution camera is used as a tracking system, and the second when the target object has different motion patterns, therefore, in order to cope with the aforementioned problems, a super-resolution technique, namely an Offset Averaging Filter (OAF) will be applied for improving the spatial resolution of images, it has previously proved its capability to enhance the spatial resolution up to four times. In this paper, an OAF technique will be used for enhancing the visual information about a target object moving in different motion patterns which are: translation, rotation and rotation plus translation. Moreover, a feature extraction algorithm, namely Two-Stage Circular Hough Transform (CHT) will be applied on the reconstructed images in order to detect and track the moving object. The results showed the capability of OAFs in improving the spatial resolution for the low resolution captured frames, and thus enhancing the visual tracking information for the small moving target objects with the three different motion scenarios.