A Novel Robust and High Reliability for Lucas-Kanade Optical Flow Algorithm Using Median Filter and Confidence Based Technique

Optical flow is an important technique for motion estimation to identify the density velocity in a level of pixel based. But in the real world, most of the sequences are interfered by many conditions that cause noises over the sequences. In order to against noise, this paper, a novel robust Lucas-Kanade optical flow algorithm based on the robust estimation and effective confidence technique using bidirectional symmetry of forward and backward flow is proposed. Comprehensive evaluations demonstrate the effectiveness results of our proposed algorithm under the Additive White Gaussian Noise (AWGN) at several noise power levels (such as AWGN at 25 dB, AWGN at 20 dB, and AWGN at 15 dB respectively) on several standard sequences such as AKIYO, COASTGUARD, CONTAINER, and FOREMAN that have differences in foreground and background movement and speed in characteristic. Peak Signal to Noise Ratio (PSNR) is used as the performance indicator in our experiment.