Multispectral detection and tracking of multiple moving targets in cluttered urban environments

Abstract

This paper presents an algorithm for target detection and tracking by fusion of multispectral imagery. In all spectral bands, we build a background model of the pixel intensities using a Gaussian mixture model, and pixels not belonging to the model are classified as foreground pixels. Foreground pixels from the spectral bands are weighted and summed into a single foreground map and filtered to give the fused foreground map. Foreground pixels are grouped into target candidates and associated with targets from a tracking database by matching features from the scale-invariant feature transform. The performance of our algorithm was evaluated with a synthetically generated data set of visible, near-infrared, midwave infrared, and long-wave infrared video sequences. With a fused combination of the spectral bands, the proposed algorithm lowers the false alarm rate while maintaining high detection rates. All 12 vehicles were tracked throughout the sequence, with one instance of a lost track that was later recovered.

Target Localization and Tracking by Fusing Doppler Differentials from Cellular Emanations with a Multi-Spectral Video Tracker

“…Doppler shifts for a cellular frequency were generated from the ground truth radial velocity with varying RMS levels of white Gaussian noise added to the Doppler shifts to model measurement uncertainty. Centroid location information was extracted from a multi-spectral video set that was developed with the DIRSIG software tool [ 34 ] with the detection and tracking algorithm presented in [ 31 ]. The results are the correct association rates over the video sequence for 100 Monte Carlo simulations.…”

“…We present an overview of the algorithm to fuse multi-spectral video data to detect and track moving targets in a cluttered urban environment [ 31 ]. The algorithm was developed and tested using a sythetically generated dataset produced using the DIRSIG toolset in visible, NIR, MWIR, and LWIR spectral bands [ 34 ].…”

“…The block diagram of the fusion algorithm described here is shown in Figure 3 . The multi-spectral video tracker fuses images for detection of moving foreground objects which are then tracked through a video sequence [ 31 ], giving two dimensional time history of centroid locations of multiple moving targets. Radial velocity estimates are computed from the tracker outputs and used to calculate the theoretical DD which would have been observed at the cell phone frequency for each tracked target.…”

Target detection, tracking, and localization using multi-spectral image fusion and RF Doppler differentials

“…I present an overview of the detection and tracking algorithm for moving targets in the multi-spectral video sequence to produce track sequences in which the theoretical DD is calculated. This algorithm has been used successfully to detect and track targets in cluttered urban environments [70]. Spectral bands used were visible, NIR, MWIR, and LWIR and was developed using a synthetically generated dataset from the DIRSIG toolset [18].…”

“…3.3. The multi-spectral video tracker fuses images for detection of moving foreground objects which are then tracked through a video sequence [70], giving two dimensional time history of (x, y) centroid locations of multiple moving targets. Radial velocity estimates are computed from the tracker outputs and used to calculate the theoretical This algorithm was developed and evaluated using synthetically generated datasets.…”