Object Tracking in Satellite Videos Based on a Multiframe Optical Flow Tracker

Abstract

Object tracking is a hot topic in computer vision. Thanks to the booming of the very high resolution (VHR) remote sensing techniques, it is now possible to track targets of interests in satellite videos. However, since the targets in the satellite videos are usually too small compared with the entire image, and too similar with the background, most state-of-the-art algorithms failed to track the target in satellite videos with a satisfactory accuracy. Due to the fact that optical flow shows the great potential to detect even the slight movement of the targets, we proposed a multi-frame optical flow tracker (MOFT) for object tracking in satellite videos. The Lucas-Kanade optical flow method was fused with the HS V color system and integral image to track the targets in the satellite videos, while multi-frame difference method was utilized in the optical flow tracker for a better interpretation. The experiments with three VHR remote sensing satellite video datasets indicate that compared with state-of-the-art object tracking algorithms, the proposed method can track the target more accurately.

A Lightweight CNN Model for Refining Moving Vehicle Detection From Satellite Videos

“…Ao [2] proposed a local exponential probability distribution noise model to differentiate potential targets from noise patterns. To track moving vehicles from satellite videos, Du [8] proposed a multi-frame optical flow tracker.…”

Detecting and Tracking Moving Airplanes from Space Based on Normalized Frame Difference Labeling and Improved Similarity Measures

“…Despite the great success in tracking with ordinary videos, algorithms from computer science may encounter new problems with satellite video data [2,8], such as low-resolution targets, similar background, and extensive geographic coverage. Du et al [6] fused Lucas-Kanade optical flow with the HSV color system and the integral image to track targets in satellite videos. Hu et al [20] proposed a tracking method incorporating the regression model and convolutional layers.…”

“…In another study by Shao et al [8], a hybrid kernel correlation filter was proposed, which employs optical flow and histogram of oriented gradient features in a ridge regression framework. While the approaches proposed in [2,6,8,20] are concentrated on single object tracking, Guo et al [5] developed a multi-object tracking approach with a high-speed correlation filter and a Kalman filter (CFKF). CFKF can deliver stable tracking under various conditions, including background similarity, occlusion, motion blur, etc.…”

“…In the last decade, new satellite sensors capable of capturing videos have been developed and launched [1][2][3][4][5][6][7][8][9][10][11][12]. Unlike satellite images, the temporal resolution of a satellite video is determined by its frame rate.…”

“…Ideally, moving object detection and tracking algorithms should be integrated into a unified method so that the outcomes of detection can automatically provide the initial locations of moving objects immediately for tracking. However, most studies in remote sensing utilize standalone tracking algorithms with the initial locations of moving objects being manually defined [2,5,6,8,9,20]. Moreover, the concentration of most tracking algorithms is on suppressing model drift, which happens when the tracked location of an object deviates from the correct one.…”