PropellerSafety.com

Detecting People Near Boats More Reliably Using Optical Image Stabilization: An Invention

Our Detecting People Near Boats More Reliably Using Optical Image Stabilization invention disclosure below was posted on 19 September 2012. We will make no changes to the text below the line following this paragraph except to correct misspellings, punctuation, and to update html/computer codes and links. Any updates to the invention will be posted below the invention disclosure.


Detecting People Near Boats More Reliably Using Optical Image Stabilization Invention

This invention has to do with improving the reliability of virtual propeller guard systems designed to detect people in the water near boats that may be in danger of being struck by the boat’s propeller.

Background Information and Current State of the Art

We have long promoted the concept of virtual propeller guards: detecting people in the water, then using that knowledge to sound an alarm, kill the motors/ marine drives, or take other appropriate actions to prevent boat propeller accidents.

Brunswick has made some efforts in this area (U.S. Patent 6,354,892, U.S. Patent 7,105,800, U.S. Patent 7,476,862, and U.S. Patent 7,511,276).

More recently, Volvo Penta has been issued two patents in this field (U.S. Patent 8,195,381 and U.S. Patent 8,271,155).

Volvo Penta Patent 8,195,381

Volvo Penta U.S. Patent 8,195,381 sketch

Most of the work to date by Brunswick and Volvo Penta utilizes infrared sensors to detect people in the water. Much of their efforts are focused on trying to more reliably detect the presence of people in the water, and especially to prevent false positives (to prevent taking action when no one is really in the water near the propeller).

We too previously posted an invention to reduce Virtual Propeller Guard false alarms on 2 February 2009 on our list of pre 2011 Propeller Guard Technologies.

Detecting humans in the water near a boat is challenging. Waves, wind, sunlight, water temperatures in the range of human temperatures, glare, the potential for non human objects near a boat (boat docks, etc.) all add to the challenge.

The two recent Volvo Penta patents teach the use of moving averages to help prevent false positives. If you look at the same scene as imaged by a camera or an infrared sensor/detector over time as sequence of images, although things may be a little bouncy, you will begin to see “detections” in the same pixels or nearby pixels.

Rather than “sounding the alarm” if a few pixels are detected once, the Volvo Penta system averages the detections for pixels over time to improve decision making. By making judgements on the moving average they are able to reduce false positives.

Similarly, the Volvo Penta patents teach mounting one or more of the sensors (infrared detectors) on gyroscopically angularly stabilized servo-platforms to compensate for the angular movement of the vessel relative to the water (rocking of the boat combined with waves).

Our invention uses Optical Image Stabilization (OIS or IS) techniques instead of mounting the sensors on “angularly stabilized servo-platforms” as described by Volvo Penta.

Optical image stabilization techniques could be used to create a more constant position of the overlay of subsequent images/scans allowing moving average techniques to be more reliable at lower cost than using angularly stabilized servo-platforms to mount the sensors.

Image stabilization systems are well known in the small handheld cameras of today. They compensate for the pan and tilt (yaw and pitch) during exposure to reduce blurring. Somewhat similar techniques or used in handheld movie cameras.

For a review of some of these techniques, see:

  • An Overview of Image / Video Stabilization Techniques. Kao and Lin. Single-Sensor Imaging Methods and Applications for Digital Cameras. Edited by Lukac. CRC Press. 2008. Pgs. 535-561.
  • Review of Motion Estimation and Video Stabilization Techniques for Hand Held Mobile Video. Rawat and Singhal. Signal & Image Processing: An International Journal. Vol.2. No.2. June 2011.
  • Digital Video Stabilization Through Curve Warping Techniques. Bosco, Bruna, Battiato, Bella an Puglisi. IEEE Transactions on Consumer Electronics. Vol.54. No.2. (May 2008) Pgs. 220-224.

Some work has been done more specifically in terms of stabilizing images from moving platforms such as UAVs (Unmanned Aerial Vehicle) and from the masts of Navy ships. For a review of these techniques, see:

  • Real-time Video Stabilization For Moving Platforms. Vermeulen. 21st Bristol UAV Systems Conference. April 2007.
  • Flying Sword: A Real Time Motion Detection Video Registration, Stabilization, Mosaicing and Moving Object Tracking System. Yang and Zhang. Asian Conference on Computer Vision (ACCV). China. 2009.
  • Evaluation of the AN/SAY-1 Thermal Imaging Sensor System. Smith and Middlebrook. 5th International Symposium on Technology and Mine Problem. 2002.
  • Video systems currently being attached to naval vessel masts (visual and infrared).

Once the image has been stabilized, blob detection and blob tracking techniques can be used to improve decision making. For a review of these techniques see:

  • Real Time Tracking of Multiple Blobs With a Possibly Moving Camera. Antonis Argyros. YouTube.com. Uploaded October 11, 2011.
  • Motion-based Background Modeling for Object Detection on Moving Platforms. Shih, Chang, Fu, and Huang. IEEE. 2007. This reference discusses the use of blob detection for moving object detection from moving platforms (like detecting moving people in the water from a moving boat).
  • Blob Detection. Wikipedia.
  • Simultaneous Video Stabilization and Moving Object Detection in Turbulence. Oreifej, Li, and Shah. IEEE. TPAMI 2012.

Our Claims

We (Polson Enterprises and the Propeller Guard Information Center) claim the the propeller safety invention described above and now make a public disclosure of it and place all parts of it not previously patented by others into the public domain for use by all who wish to manufacture or use them. The world is certainly welcome to patent improvements to this invention but the basic concepts not previously patented by others are open to all.

More specifically we claim:

1. The use of Optical Image Stabilization (OIS / IS) techniques with a sensor (infrared, visual, or other band) to detect people in the water near a boat.

2. The use of Optical Image Stabilization (OIS /IS) techniques with a sensor (infrared, visual, or other band) to detect people in the water near a boat that also integrates the use of blob detection and /or blob tracking techniques as part of a virtual propeller guard system.

3. The use of Optical Image Stabilization (OIS / IS) techniques with a sensor (infrared, visual, or other band) to detect people in the water near a boat that also integrates moving average techniques to reduce false positives as described in the two Volvo Penta patents (U.S. Patent 8,195,381 and U.S. Patent 8,271,155) as part of a virtual propeller guard system.

4. A virtual propeller guard system of the nature of that described in the two Volvo Penta patents (U.S. Patent 8,195,381 and U.S. Patent 8,271,155) that also incorporates Optical Image Stabilization. For example, a virtual propeller guard system that uses “angularly stabilized servo-platforms” as mountings for one or more sensors PLUS also uses Optical Image Stabilization to even better stabilize those images/scans.

5. The use of Optical Image Stabilization techniques as described in Claim #1 in conjunction with one or more of the techniques identified in our previous invention to reduce Virtual Propeller Guard false alarms posted on 2 February 2009 on our list of pre 2011 Propeller Guard Technologies.

Where a virtual propeller guard system is defined as shown on our Virtual Propeller Guard Terminology page.

Additional Reference Materials

  • Makohead a camera mount designed for use on boats.

END OF INVENTION DISCLOSURE


Comments

We welcome your comments on this invention.

Leave a Reply