DRAFT: This module has unpublished changes.

Applied Physics Lab: Summer Internship:

Underwater Mine Identification via Sonar

The sources listed here contributed to understanding and progression of that mini-research project, but not the collection of research (regarding The Story Template) presented this year.  The results of this learning experience were not nor will be presented in any other form except on this page, and under the Journals tab under "summer work".


Color Imaging Sonar User’s Manual Imaginex Model 855 (Monograph). (n.d.). Imaginex Technology Corporation. 

This source explained some of the basic sonar terminology, principles of side-scan sonar, and included figures supplementing the text.  The sonar system used in this experiment utilizes color interpretation of the sea floor.  The red spectrum indicates the target’s strong reflective quality, while blue represents an absorbent tendency of the target.  


Sonar operates by sending an audio signal and measuring the time it takes for the sound to return as an echo.  This time is divided by 2 (going and coming distance) and divided by the speed of sound under water (roughly a mile a second).  If the audio signal encounters nothing in its way to reflect it, the signal is lost.  When an object obstructs the signal’s path, the signal bounces back to the receiver.  


There will be two side-scan units on either side of the prototypes torso.  These broadcast the auditory signals in a fan shape hitting the sea floor.  The width of the fan is small (1.7 degrees) but the vertical dimension of the fan is greater (roughly 40-60 degrees).  


The side-scan sonar system also uses a system of narrow pencil-shaped beams; each of these beams containing many thousands of points corresponding to the visual representation sent back to the viewer.



Debugging Reference Clock Drivers. (2007, November 24). Retrieved July 8, 2009, from  www.cis.udel.edu...‌~mills/‌ntp/‌html/‌rdebug.html



Dobeck, G. J., Hyland, J. C., & Smedley, L. (n.d.). Automated Detection/Classification of Sea Mines in Sonar Imagery (Monograph). Coastal Systems Station Panama City, Florida: Naval Surface Warfare Center, Dahlgren Division. 

This report details the specifications for a previous project similar to the current one.  It describes the algorithm developed for the automated detection and classification of mines underwater using side-scan sonar.  


There are four fundamental stages of this algorithm: image enhancement, detection, optimal feature selection, and classification.  Matrices are used to indicate and describe the needed image inputs and outputs.


The first stage, image enhancement normalizes the image, removing as much extraneous detail as possible  The gain (viewed color spectrum) is adjusted relative to the given and pre-set mean gains of the image.


Some environmental conditions make it more difficult to identify targets than others: gravel is more reflective, like the target, than mud which tends to show targets with good contrast.


Targets indicators are the presence of the four following zones: pre-target, highlight, dead zone, and shadow/‌post target.  The pre-target area is the sea floor surface (closest to the sonar craft) before the target.  The highlight area is the bright spot that is most apparent in a sonar visualization; it is the highest point of the object (vertically closest to the sonar).  The dead zone is the point vertically in line with the target which is angled away from the sonar wave.  After the highest point of the highlight zone, the downward sloping remainder of the target would not be ‘seen’ by sonar rays.  Thus their signal is lost, and the area where this occurs is called the dead zone.  The final area of consequence in terms of detecting targets is the shadowed zone.  Like an object will cast an optical shadow from an incoming beam of light, the undersea targets also cast an acoustic shadow from the incoming sonar beam.  The area (of sea floor) farthest from the sonar craft after the target is grouped together with the area of shadow.


The length of a target’s acoustic shadow is relative to the height of the object which subsequently can be calculated through the trigonometric ratios.  


Now that the general areas of a potential target have been acquired, the next stage refines these findings by comparing ratios of the magnitudes of the four zones relative to the pre-programmed standard of how a mine would appear.  A mask (chosen by the user according to the target and setting) is fitted over potential targets to compare the proportions of the four zones.  This is done using a list of 18 characteristics that include the relations of size and strength of the return signals.  


To make the system more applicable in various conditions, the algorithm allows several options (selected by the user) to describe the setting and goal in terms of target mine type and primary sea floor environment type.


The whole point of this study is to devise a process that a machine can use to differentiate mines from a distracting background of clutter.  The way to do this is to articulate the acoustic differences between the target and all other features that will likely be encountered.



Fish, J. P. (n.d.). Acoustics and Sonar Primer. Retrieved July 1, 2009, from The Institute for Marine Acoustics website:  www.instituteformarineacoustics.org‌SonarPrimer/‌SideScanSonar.htm 

This source helped to confirm what I’ve already learned.  It explained why it is important to keep the ship going in a straight path while measuring sonar (because turns cause blurring and make non-targets resemble targets.  This source was also useful for the example sonar images it provided that allowed the viewer to guess and check his/‌her analyses.  In addition, this page explained the older grayscale system with sonar.  The most popular version appears as a negative of the sea floor (near items are black and background white).  Its review on the importance of shadows was also helpful.



How to configure an authoritative time server in Windows XP. (n.d.). Retrieved July 10, 2009, from support.microsoft.com...‌kb/‌314054



Introduction to NTP. (n.d.). Retrieved July 8, 2009, from  www.akadia.com‌services/‌ntp_synchronize.html



Lombardi, M. (n.d.). Computer Time Synchronization (Monograph). Retrieved from National Institute of Standards and Technology: Time and Frequency Division website: tf.nist.gov/‌service/‌pdf/‌computertime.pdf



Manufacturers of Time and Frequency Receivers. (2009, April). Retrieved July 8, 2009, from National Institute of Standards and Technology: Time and Frequency Division website: tf.nist.gov...‌general/‌receiverlist.htm



The Network Time Protocol (NTP) Distribution. (2009, April 8). Retrieved July 8, 2009, from  www.cis.udel.edu‌~mills/‌ntp/‌html/‌index.html#info

R ̈ mer, K., Blum, P., & Meier, L. (n.d.). Time Synchronization and Calibration in Wireless Sensor Networks (Monograph). Retrieved from  www.site.uottawa.ca...‌~casteig/‌files/‌csi5140-jaynesh-doshi.ppt



Setting up your SportScan for use (Monograph). (n.d.). 

This source was not particularly helpful for my purposes; it was a manual describing how to assemble and use the SportScan.  It was helpful to realize the need for an on-board GPS to counteract the effects of motion in terms of distorting the visualization of the sonar image.  This source may of more help later when I need more specific information about the device (signal’s range of frequency, frequency signal is transmitted, wiring, etc.).


A problem that I may be asked to review later is how to know that the craft has returned to the point of origin.  Clutter on the sea floor changes as a result of currents.  Is GPS specific enough to know exactly where it is?  Depending on the mine size (crab pots for the purposes of this experiment), a few meters in either direction may throw off the system’s ability to re-find the targets again.


In my autonomous tomato harvester, I found this to be the case.  I planned on solving the problem by placing reference beacons around the area and having the robot triangulate its relative position.  It is unlikely that a similar system would work in this case due to the rough conditions, cost, sheer size of the ocean, and the probable immediate need of the military for this purpose.



Tucker, J. D. (2009). Coherence-based Underwater Target Detection for Side-Scan Sonar Imagery (Doctoral dissertation).

DRAFT: This module has unpublished changes.