Cleaning up the world’s oceans with underwater laser imaging

Abstract

It is estimated that there are 5.25 trillion pieces of plastic waste in the oceans. Of this plastic, roughly 70% is not visible because it is underwater. This plastic is harmful to animals as they can get entangled in the litter and ingestion of plastic toxins can cause disease. Although cameras and hyperspectral imagers have been used to find plastics in the ocean, their efficacy is limited to visible garbage on the ocean surface. Ocean cleaning up requires retrieval of not just visible garbage, but underwater garbage as well. For this task, both sonar and Light Detection and Ranging (LIDAR) can be utilized; however, sonar can be harmful to marine animals by disrupting their echolocation capabilities. LIDAR generates high resolution point clouds with which recognition of small objects may be difficult. Therefore, we investigated the ability of lasers to image and identify underwater objects. Laser imaging can penetrate underwater, providing informative images at depths where there is no ambient light. This study examined the influence of various factors on underwater laser image quality: polarization, depth, turbidity, and object material. We conducted experiments by illuminating objects in water with a laser and recording illuminated objects using a camera. We utilized various image processing techniques to enhance image quality. Our results show that deconvolution was a more effective method than alternatives for reducing blurriness and that laser imaging is a viable method for detecting underwater garbage.