Intelligent multimodal and hyperspectral sensing for real-time moving target tracking

Abstract

Abstract-Real time moving target tracking and identification with hyperspectral imagery is still very challenging with conventional sensors and algorithms. The increased information content of hyperspectral imaging has enabled improved classification and quantification of targets of interest. However, recording hyperspectral data for target classification is very time consuming. We design a sensor platform with multi-modalities, consisting of a dual-panoramic peripheral vision system and a narrow field-of-view hyperspectral fovea. Thus, we only need to capture hyperspectal images in regions of interest. This design is inspired by the human vision system where the periphery vision of the retina is used to detect motion and the fovea of the retina is used to recognize objects.The proposed intelligent sensors design is also supported by real-time algorithms for target detection, tracking and identification. Only hyperspectral data for areas of interest are captured for target classification and recognition. Important issue such as multimodal sensing component cooperation, region of interest extraction, target tracking, hyperspectral image analyzing and target signature identification are discussed.

Recent Advances on Video Coding

“…Further, there are many other recent tracking methods, such as: (Mehmood, 2009) implements kernel tracking of density-based appearance models for real-time object tracking applications; discloses a wireless, embedded smart camera system for cooperative object tracking and event detection; (Sun, Z. & Sun, J., 2008) presents an approach for detecting and tracking dynamic objects with complex topology from image sequences based on intensive restraint topology adaptive snake mode; (Wang & Zhu, 2008) presents a sensor platform with multi-modalities, consisting of a dual-panoramic peripheral vision system and a narrow field-of-view hyperspectral fovea; thus, only hyperspectal images in the ROI should be captured; (Liu et al, 2006) presents a new method that addresses several challenges in automatic detection of ROI of neurosurgical video for ROI coding, which is used for neurophysiological intraoperative monitoring (IOM) system. According to (Liu et al, 2006), the method is based on an object tracking technique with multivariate density estimation theory, combined with the shape information of the object, thereby by defining the ROIs for neurosurgical video, this method produces a smooth and convex emphasis region, within which surgical procedures are performed.…”