# US 2021/0279895 A1 · Real-time Marine Snow Noise Removal from Underwater Video

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0279895 A1
- Document ID: US20210279895A1
- Status: Published application
- Kind: A1
- Application number: 17/184,076
- Filed: 2021-02-24
- Published: 2021-09-09
- Earliest claimed: 2020-02-25
- Inventors: JACK WADE
- Web page: https://patentgrep.com/patent/US20210279895A1

## CPC

- G06T 7/254: involving subtraction of images
- G06T 7/269: using gradient-based methods
- G06T 5/70: Denoising; Smoothing
- G06T 5/77: Retouching; Inpainting; Scratch removal
- G06T 7/215: Motion-based segmentation
- G06T 2207/10016: Video; Image sequence
- G06T 2207/10024: Color image
- G06T 2207/20224: Image subtraction
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Optical flow refers to the pattern of apparent motion of objects, surfaces, and edges in a visual scene caused by the relative motion between an observer and a scene. Optical flow algorithms can be used to detect and delineate independently moving objects, even in the presence of camera motion. The present invention uses optical-flow algorithms to detect and remove marine snow particles from live video. Portions of an image scene which are identified as marine snow are reconstructed in a manner intended to reveal underwater scenery which had been occluded by the marine snow. Pixel locations within the regions of marine snow are replaced with new pixel values that are determined based on either historical data for each pixel or a mathematical operation, such as one which uses data from neighboring pixels.

## Claims

### Claim 1

An optical flow sensor to detect the movement of water while immersed under the surface of said water comprising:
a camera for detecting optical flow events within said water;
an optical flow comparison process wherein independently moving objects in the presence of said water are isolated against a backdrop of water flow not associated with said moving objects and wherein movement of said camera is subtracted from movement associated with said object;
a process performed by a computer wherein video signals produced by said camera wherein marine snow noise is identified and removed from an overall underwater image, so that said overall underwater image consists of said moving objects and wherein said overall underwater image is relatively clear for human observation, and
wherein underwater scenery otherwise occluded by marine snow noise become more clearly visible for human observation.

### Claim 2 (depends on 1)

An optical flow sensor according to claim 1 wherein pixel locations within the regions of marine snow noise are replaced with new pixel values that are determined based on either historical data for each pixel or a mathematical operation, such as one which uses data from neighboring pixels.

### Claim 3

A method for an optical flow sensor to detect the movement of water while immersed under the surface of said water and producing a corresponding data output comprising:
a camera for detecting optical flow events within said water;
an optical flow comparison process wherein independently moving objects in the presence of said water are isolated against a backdrop of water flow not associated with said moving objects and wherein movement of said camera is subtracted from movement associated with said object;
a process performed by a computer wherein video signals produced by said camera wherein marine snow noise is identified and removed from an overall underwater image, so that said overall underwater image consists of said moving objects and wherein said overall underwater image is relatively clear for human observation, and
wherein underwater scenery otherwise occluded by marine snow noise become more clearly visible for human observation.

### Claim 4 (depends on 3)

A method according to claim 3 wherein pixel locations within the regions of marine snow noise are replaced with new pixel values that are determined based on either historical data for each pixel or a mathematical operation, such as one which uses data from neighboring pixels.

### Claim 5

A system for providing clear underwater visualization using an optical flow sensor to detect the movement of water while immersed under the surface of said water comprising:
a camera for detecting optical flow events within said water;
an optical flow comparison process wherein independently moving objects in the presence of said water are isolated against a backdrop of water flow not associated with said moving objects and wherein movement of said camera is subtracted from movement associated with said object;
a process performed by a computer wherein video signals produced by said camera wherein marine snow noise is identified and removed from an overall underwater image, so that said overall underwater image consists of said moving objects and wherein said overall underwater image is relatively clear for human observation, and wherein underwater scenery otherwise occluded by marine snow noise become more clearly visible for human observation, and
wherein pixel locations within the regions of marine snow noise are replaced with new pixel values that are determined based on either historical data for each pixel or a mathematical operation, such as one which uses data from neighboring pixels.

### Claim 6 (depends on 5)

A system according to claim 5 wherein said neighboring pixels include at least said historical data for each of said pixels and wherein said historical data is shared with other third party users of said system.
