Time-of-flight based imaging in strong scattering underwater environments

Time-of-flight (TOF) based underwater imaging is of great importance in practical applications due to its high image quality. Existing works separate scattered and ballistic photons in the time and space domains to recover objects in weakly scattered underwater scenes. However, in turbid underwater environments, absorption and strong anisotropic scattering cause weak ballistic light tightly coupled with forward-scattered and backward-scattered photons. The difficulty in isolating scattered light significantly limits the imaging capabilities of the existing methods. To tackle the problem, a forward-backward-distinctive imaging model is proposed, which models the spatial distribution of forward scattered illumination by point spread function (PSF) of the turbid water while modeling the backward scattered field by diffusion equation (DE) to describe the anisotropic scattering in the water accurately. Based on this, the underwater boundary migration model (WBMM) is derived, an explicit mapping relationship between the scene and the measurements is established, and a reconstruction algorithm utilizing time-of-flight information in the turbid water is realized. Experiments on a real scattering imaging system are conducted to demonstrate the effectiveness of the proposed method. Experimental results show that the proposed method outperforms the existing methods in terms of reconstruction accuracy and imaging limit subjectively and objectively. Even though the signal photons are highly scattered in turbid water, and the spatial distribution of the reflected light are greatly changed, the proposed method can reconstruct an object with a one-way scattering length of 9.5 mean transmission free-range (TMFPs), corresponding to a round-trip scattering length of 19 TMFPs, which is very favorable for dealing with underwater scattering imaging problems.

Algorithm Pipeline

Results

Related Works