Side Imaging in 360° ?

A "look down and sweep outwards" array formation, currently used by Humminbird 797/997/1197 units, takes a slice of the full circle possible.
So the whole transducer must be rotated to get the full view, with the speed of rotation being a function of distance viewed (ping return rate) and refresh/ping rate.
In order to have a constantly rotating radar/sonar array with good connection from rotating transducer/pole and the dispay head unit, a 2.4ghx wireless connection would be required between xducer and head unit.

That leaves the rotating mechanism, speed of rotation governed by the head unit, or lower cost, speed of rotation by an easily understood rule of thumb for display with "no information gaps", and powered from the 12V main battery.
A simple motor driving a pinion, mounted on the side of the transom bracket, winding a xducer pole with a crown gear fixed on it would work fine.

The final part of the system would be a software option to do the pie-chart display type, with some means of feedback so the head unit "knows" which way the xducer is facing at any moment. And an NMEA compass on the xducer pole would provide that information nicely, in a form the unit already is fitted to cope with. NMEA by wireless to head unit, along with ping returns.

_________________
Float Tube Fishing in Ireland

Dear Santa,

This is what I want this christmas, a software update with a display mode showing the pie-shaped sector that is swept when I scan with my portable system. It doesn't need to show a 360 degree image or the image's orientation. But it would be nice if the angle is adjustable. I have made a sketch for Santa to explain what I mean.

A very kind boy

Hi,

I stumbled over an image editing freeware which can do the transformations I need for the distorted rotation images. The programme is the commando operated ImageMagick. No GUI for the advanced options and it took some time to figure out how the syntax worked. Transformed images have better proportions but some close detail is lost and the image becomes a bit blurred farther away. (I don't think it's possible to fill a 360 degree image in a snapshot with acceptable quality. A full circle image would require use of the recording facility, I think. But when stationary there is no direction info from the GPS. Is it possible add a fake 360 degree turn to the recorded file and thereby fool a viewer show a circle although the position is constant?) It would certainly be much more convenient to have the polar view in the instrument from the beginning! I transformed a few of my snapshots:

Rickard

Small wrecks before:

and after:

Pipeline before:

and after:

Dredge barge before:

and after:

River overview before:

and after:

Excellent work Rickard.

I assume your examples were done with vertical axis reduction at low range and axis expansion at max range displayed .
Or did you work with arcs to transform the images?

The distortion is very acceptable to me ..... I expected it would be more, but those results are very good!

Now if that transformation can be done on a PDA palmtop computer, we could view results like those while on the water, at the waterside, right now!
I will see if I can find a PDA image manipulation software that will do it.

A new software option on the SI menu for this would be really neat.

Norm

_________________
Float Tube Fishing in Ireland

Thanks Norm,

I'm not sure exactly what the transformation command does but if I got it right it works with arces. The horizontal distance from left border to a pixel is the length of a vector and the direction of that vector (angle) depends on vertical distance from center of the image. The vectors, with their respective angles, are displayed with reference to a common origin, the center of the circle (or circle sector). Thus, a vertical, straight line should appear as a circle sector after transformation.

I agree the distorsion looks less severe than expected before transformation. In these examples the images are perfectly interpretable from about a 5th of max range to max range in the original images. But I still wish a polar display mode for Christmas!

Here in Luleå we have fresh, but thin, ice now and in a week or two I will try scanning wrecks through holes in the ice. Ice is a very stable platform so I hope to get clear, but distorted, results.

Rickard

I'm developing a method for scanning under ice by rotating the transducer when it hangs in the transducer cable. Until now I have tested with 10 meters of cable and the results are promising. The rotation movement gets very smooth because of the flexibility in the cable and the dampening effect of water. I hope I will be able to scan near the seafloor at large depths with this method. I have solved the problem with strong, interfering echoes from the ice but still have the problem with interference from the not used (I don't use the left) SI channel. This interference creates ghost images for two reasons: reflections from structures to the left are received by the right channel (the one I use) and crosstalk in long cables. A brutal way to solve this is covering the left side with cork - which works very well, but I think there must be a better solution. A simple solution would be to disconnect the left SI channel altogether, something that cannot be done in the menues as far as I can see. But could physical disconnection of one channel harm the instrument?

(Since my 981 is on repair I presently use a Matrix 37 with a SI transducer and the transducer selection set to Wide Side. This setting gives full screen SI images with a max range of 50 meters.)

Thanks,
Rickard

	Rickard wrote:
	... A simple solution would be to disconnect the left SI channel altogether, something that cannot be done in the menues as far as I can see. But could physical disconnection of one channel harm the instrument?

There is a remote possibility that over extended time one side will get "burned in" creating a "preferred route of lower resistance". This would cause a brighter side when the equipment was reverted to the proper setup.
I would be careful of overvoltage spikes too. With less than the designed current drain, voltage spikes could be hjigher. Maybe it's not a good idea to turn sensitivity or range to maximum while in this mode?

	Rickard wrote:
	... (Since my 981 is on repair I presently use a Matrix 37 with a SI transducer and the transducer selection set to Wide Side. This setting gives full screen SI images with a max range of 50 meters.)Rickard

A photo of the screen display would be interesting next time you have an opportunity.

Norm

_________________
Float Tube Fishing in Ireland

Hi,

As long as there is no guarantee from the "peanut gallery" that disconnection won't harm, I will continue as before. Cork is a wonderful material for controlling echo sounders!

I have a photo from my last test with the Matrix 37 that shows the same dredge bargue as the one shown before. I used the SI transducer (the one that came with the 981) and set transducer to Wide Side. In the M37 you can chose between Wide Side and QuadraBeam. However, the Quadra Beam mode has no display options as large as those the WideSide option has. All new sideimaging transducers can be used with the M37 (and also with most other older units with some sort of sideviewing at 455 kHz, thus, there are many HB owners who have a high resolution sideimaging unit and doesn't know about it.....) The photo might not look too impressive but it shows how one side can be displayed in full screen. It was dark - it usually is this time of the year close the arctic circle - and it was snowing. Scrolling can't be halted so I had to hurry and take (stiff fingers..) a snapshot with flash before the target disappeared from the screen. Screenshots cannot be saved to a memory card in the M37. Honestly, the photo does not give a fair impression of what the M37 can do. The M37 is fully capable to replace the 981 when testing the rotating imaging methods.

Note, the absence of any trace of surface/ice interference. The HB sideimaging transducers, as other transducers, have no good shield against transmitting/receiving in the direction of the surface. This is easily seen when scanning from a towfish at moderate depth or when scanning under ice. In the towfish case reflections from the surface appear as rather thin lines in the image. When "ice-scanning" these lines are replaced by broad bands which really can destroy the image if one is unlucky. This can be fixed in the usual way - with cork. I attached a "roof" of 2 mm cork (gasket material) over the whole transducer. This roof is sticking out about an inch from the transducer side. The cork sheet also protects, to some extent, from secondary returns which can cause ghost images. Note, this is an issue only when the transducer is operated at some distance under the surface/ice. When used close the surface, which is the normal situation, surface returns travel a very short distance and appear in the surface clutter zone in the display.

Rickard

I played with the photo and got this polar image:

Friends,

May I (proudly) present my first full, true 360 degree images. This involved rotating the transducer when hanging in the cable under ice, recording to memory card, conversion with Son2xtf, conversion with SIView, saving to bitmap, editing with Picture Manager and conversion to polar form with ImageMagick. Many, but rather easy steps (except perhaps for the first and last steps). Very little info is lost during the process. I still have trouble rotating the transducer at constant speed and will try solve this by adding more mass in the transducer end to get more inertia.

Depth is about 7 meters but the transducer was positioned at different depths from 1.5 meters over the sea floor to just under the ice. Range was set from 15 meters up to 45 meters. I kept the water column because removing it causes distortion as long as there is no slant range correction. Thus, the images have a black circle in the middle.

Rickard

A bargue.

A scan close the sea floor. Note the tyres.

Car, debris and the corner of a quay.

Very cool Rickard!

Excellent images, I'm looking to do the same thing, through ice.

For getting a constant rotational speed, I'm thinking using a 12V windshield wiper motor or a power window motor out of a newer-model old junk automobile. Maybe have a piece of round plywood, 24" diameter to set over the ice hole, with the transducer attached to a telescoping pole in the center, and the battery and head unit sitting on something that rotates. I may not be describing it well, I'll try to build one then post pictures.

Hi,

Thanks, and good luck with your project! There are some pictures from other home-built ice-scanning systems at Yahoo: http://groups.yahoo.com/group/sideimaging/photos/album/930006087/pic/list
Perhaps you have to be signed in as member to see them.

Waiting for your reports
Rickard

A 1 kg weight solved the problem with shaky rotation speed. Here are my first scans from inside a wreck. I have scanned this wreck many times with hull mounted transducers and from a towfish but these images told me much more. The transducer was located about 4' and 2' from the bottom of the wreck. Radius about 16 m. Depth 10 m. Length of wreck 18 m. Age...?

Rickard

Scanning in 360 degrees always, provided rotation speed is constant, produces images with correct proportions. No need for speed and position information to adjust proportions, the geometrical logic does the job for you. Slant range distortion is the same as in standard scanning, though. (I will try to solve that too to get rid of the "black holes"). But true proportions also means true influence of disturbance. Many of my images show substantial contamination from sidelobes. I haven't been aware of them before. I added two images to show this. The image was generated with a HDSI transducer (the one that comes with the 997) at 455 kHz. The first image is as it came out with no adjustments. One detail is encircled together with the sidelobe returns from the same detail. The mainlobe and the two first sidelobes are indicated with lines. The second image is adjusted to eliminate the sidelobes. Brightness is decreased and sharpness is increased. This also removes some mainlobe information but the image is more "true" than the original. I added a figure showing the beamform of a 100 mm rectangular array operated at 455 kHz.

From this one would wish better sidelobe suppression in the transducer. However, as far as I can see in simulations the potential for improvement, given a specific length, is small.

Rickard

Original with fuzzy sidelobes. Sidelobes create double images and overall impression gets shaky. A single object can sometimes appear as three objects, a well-defined object between two fuzzy copies.

Cleaned image. More "true" but dull.

Theoretical beamform generated with the standard formula for rectangular arrays. The beamform doesn't change much even if the array differs a lot from optimal configuration. Thus, there is not much hope for any significant improvement.

Added an image where even the second sidelobes to right and left can be seen. The target is a fish (not moving). The image is cut out from a larger image.

Last edited by Rickard on Sun Feb 08, 2009 6:51 pm; edited 1 time in total

Great work very impressive! Surprised