evaluation start service reverse

[re_solidworks]

Bye to all,

a few days ago I saw a demo of a 3d laptop scanner from the Creaform. I was particularly amazed at the simplicity of use (I tried it myself) and its cost could fall into an acceptable budget, since in my area there is no one who deals with this.
Now I'm evaluating what to do and the questions are 2: the first is if the equipment that they proposed me is valid or a bowl and the second is if I can take enough "manic" for surface management. If with the cad in the mechanical field, I can't say the same with the surfaces with which I do really little.

to me it seemed a valid solution and then with the scanner (optical and not laser) also give a license of rapidform for the conversion of stl.
The scanner has a precision of 0.1 mm, more than enough for what I should do.
I would mainly like to detect the carousel glassware, the required accuracy could be ±2 mm, so I'm fine in it with the scanner tolerances.

in a first period I would not like to buy anything else for which I will acquire stl, I will convert them into swx or proe with rapidform and then I will use them in cad.
I can evaluate in a second time and if I have the need to water the rex of proe, but I do not want to go to the ferrari and I do not think I will need surfaces in class a.

What do you say? Is it worth it? advice?

Hello and thank you all.

 

[stefanobruno]

and its cost could fall into an acceptable budget

What are the figures?

 

[re_solidworks]

What are the figures?

35k€ approx. I don't know how much space there is to treat.. . .

 

[maxopus]

Bye to all,

a few days ago I saw a demo of a 3d laptop scanner from the Creaform. I was particularly amazed at the simplicity of use (I tried it myself) and its cost could fall into an acceptable budget, since in my area there is no one who deals with this.
Now I'm evaluating what to do and the questions are 2: the first is if the equipment that they proposed me is valid or a bowl and the second is if I can take enough "manic" for surface management. If with the cad in the mechanical field, I can't say the same with the surfaces with which I do really little.

to me it seemed a valid solution and then with the scanner (optical and not laser) also give a license of rapidform for the conversion of stl.
The scanner has a precision of 0.1 mm, more than enough for what I should do.
I would mainly like to detect the carousel glassware, the required accuracy could be ±2 mm, so I'm fine in it with the scanner tolerances.

in a first period I would not like to buy anything else for which I will acquire stl, I will convert them into swx or proe with rapidform and then I will use them in cad.
I can evaluate in a second time and if I have the need to water the rex of proe, but I do not want to go to the ferrari and I do not think I will need surfaces in class a.

What do you say? Is it worth it? advice?

Hello and thank you all.

ahio, apart from the surfaces in class a, if you want info on the instrument I can collect a bit around.
because the problem is always the usual, to sell it they do very much, if by chance you have problems and in Italy they have 3 people on the cross ... to the first inconvenience are your own cabbages.

How much longer are the scans you have to acquire?
How long do you have to use it during the year?
How long do you plan to train for surface modeling?
Keep in mind that a lot of work is needed to cushion the fine.
If you want, we can hear it on the phone. :smile:

 

[re_solidworks]

ahio, apart from the surfaces in class a, if you want info on the instrument I can collect a bit around.
because the problem is always the usual, to sell it they do very much, if by chance you have problems and in Italy they have 3 people on the cross ... to the first inconvenience are your own cabbages.

How much longer are the scans you have to acquire?
How long do you have to use it during the year?
How long do you plan to train for surface modeling?
Keep in mind that a lot of work is needed to cushion the fine.
If you want, we can hear it on the phone. :smile:

Hello and thank you for your intervention, maybe then we feel so free to say goodbye.
If you can get some info on your device, I'd be grateful.

to answer your questions:

maximum dimensions of a car, I would say 2000x1000x1000. in this there should be no trouble, we made it live and the times are + that acceptable.
on quanot I have to use it... I don't know, depending on what they call me and this will depend on how much it costs. certainly all manufacturers around have hand-built details of which do not have accurate models, this tool would be really useful. However if the service costs too much is already out.
for training I don't know, this is the braid + big. taking the tool is easy, you look if there is the budget and the answer comes quickly, but to decide if and how much time I will have to devote to this... depends on many factors and now I don't know how to evaluate them. It is a completely new activity and not having well understood how it will be difficult to understand. I'd throw myself up,
I know that for amortization, and I'm indecisive about what to do.

 

[heterodyne]

Hi.

If I can give you my opinion, to do that kind of activity I suggest you take a look at this possible other alternative: www.3d3solutions.comcompared to what you saw you would have some advantage:

1 - greater dimensional accuracy because you would have the possibility to use photogrammetric markers and, in case the surface is set, you can align all sizes without using anything

2 - very high resolution: up to 3 megapixels

3 - measuring fields to acquire models in order of 3 meters but also very small objects or however of medium size

4 - spending abundantly under 20 k

5 - possibility, possibly, to use also top quality c-mount lenses such as zeiss and schneider

6 - speed on the single scan of 0.85 seconds

7 - white/structured light technology

8 - black/dark and colour surface acquisition without opacizing

 

[re_solidworks]

Hi.

If I can give you my opinion, to do that kind of activity I suggest you take a look at this possible other alternative: www.3d3solutions.comcompared to what you saw you would have some advantage:

1 - greater dimensional accuracy because you would have the possibility to use photogrammetric markers and, in case the surface is set, you can align all sizes without using anything

2 - very high resolution: up to 3 megapixels

3 - measuring fields to acquire models in order of 3 meters but also very small objects or however of medium size

4 - spending abundantly under 20 k

5 - possibility, possibly, to use also top quality c-mount lenses such as zeiss and schneider

6 - speed on the single scan of 0.85 seconds

7 - white/structured light technology

8 - black/dark and colour surface acquisition without opacizing

Thanks for the advice. much of what you say I have seen in this instrument, I speak of this:
http://www.creaform3d.com/en/metrology-solutions/handheld-portable-3d-scanner-goscan-3dI have seen using markers and the kind of light seems to me the same. the scanning speed is very high and using it in hand you can make huge scans, just have the time and power of cpu.
the test scan was also done on a glossy white piece, without opacizing. Keep in mind that in the figure there is rapidform understood and only that cota quite a lot.

 

[heterodyne]

are two different instruments; what I told you is part of the white light scanner family; if you want to document you can refer to the "German" measuring systems; atos, breuckmann and steinbichler.

what has been proposed to you is a little different though you seem equal; In fact the systems of that type have good advantages but if the goal is to obtain quality, cleanliness of mesh and of course to look for dimensional accuracy are comparable with the fixed sensors.

Maybe check if the rapidform they proposed opens only and exclusively the files produced by the scanner. there are bundled rp versions that cannot import the sl.
of course you can export the maths in iges, ste and also in stl...
this system was made with the aim of providing a reverse program combined with the measuring system... (at a reasonable price). .

Hi.

 

[re_solidworks]

are two different instruments; what I told you is part of the white light scanner family; if you want to document you can refer to the "German" measuring systems; atos, breuckmann and steinbichler.

what has been proposed to you is a little different though you seem equal; In fact the systems of that type have good advantages but if the goal is to obtain quality, cleanliness of mesh and of course to look for dimensional accuracy are comparable with the fixed sensors.

Maybe check if the rapidform they proposed opens only and exclusively the files produced by the scanner. there are bundled rp versions that cannot import the sl.
of course you can export the maths in iges, ste and also in stl...
this system was made with the aim of providing a reverse program combined with the measuring system... (at a reasonable price). .

Hi.

Thanks again for your intervention. I want to reiterate that for me it is a new world and the product I have seen and tried surprised me for the speed and simplicity of use, but if there is a better or equivalent product but cheaper....well come. I am happy because I realized that prices start to become affordable, a few years ago they were much more expensive tools.

with a fixed like what you indicated me would put markers when you put the tool to advance with the scan, did I understand?

In the case of Creaform I have seen using the tool without marker, it is used as an aerograph and you see the model appear in rendering in real time. If you hold your laptop in your arm (they have a suit on purpose) you can continue scanning until you want.
for this it seems easy and "open" to a great variety of works. then as I said the 100 microns of this tool are exaggerated for what I have to do.

to me it seemed a valid product, but I am a frog and in this matter I am easily radiant, so I ask lumi to the most experienced.

returning to the instrument you propose... Does 20k include software? What precision does the tool have?
Are there any particular limits in its use? What kind of marker do you use?

hello and thanks again to all.

 

[maxopus]

Thanks again for your intervention. I want to reiterate that for me it is a new world and the product I have seen and tried surprised me for the speed and simplicity of use, but if there is a better or equivalent product but cheaper....well come. I am happy because I realized that prices start to become affordable, a few years ago they were much more expensive tools.

with a fixed like what you indicated me would put markers when you put the tool to advance with the scan, did I understand?

In the case of Creaform I have seen using the tool without marker, it is used as an aerograph and you see the model appear in rendering in real time. If you hold your laptop in your arm (they have a suit on purpose) you can continue scanning until you want.
for this it seems easy and "open" to a great variety of works. then as I said the 100 microns of this tool are exaggerated for what I have to do.

to me it seemed a valid product, but I am a frog and in this matter I am easily radiant, so I ask lumi to the most experienced.

returning to the instrument you propose... Does 20k include software? What precision does the tool have?
Are there any particular limits in its use? What kind of marker do you use?

hello and thanks again to all.

will make you a goat so much

 

[heterodyne]

Thanks again for your intervention. I want to reiterate that for me it is a new world and the product I have seen and tried surprised me for the speed and simplicity of use, but if there is a better or equivalent product but cheaper....well come. I am happy because I realized that prices start to become affordable, a few years ago they were much more expensive tools.

with a fixed like what you indicated me would put markers when you put the tool to advance with the scan, did I understand?

In the case of Creaform I have seen using the tool without marker, it is used as an aerograph and you see the model appear in rendering in real time. If you hold your laptop in your arm (they have a suit on purpose) you can continue scanning until you want.
for this it seems easy and "open" to a great variety of works. then as I said the 100 microns of this tool are exaggerated for what I have to do.

to me it seemed a valid product, but I am a frog and in this matter I am easily radiant, so I ask lumi to the most experienced.

returning to the instrument you propose... Does 20k include software? What precision does the tool have?
Are there any particular limits in its use? What kind of marker do you use?

hello and thanks again to all.

Hi.

I answer your questions

As for markers (some also call them targets) the question is debated.

If you have particularly large surfaces and with large curvature (e.g. the hood of a car or a simple table) markers are the only alternative because, without them, no software could be able to determine the correct alignment of the acquisitions. I made you the example of a table not by chance. without the help of markers could be infinite alignment solutions due to the flatness of the surface. calculating the interdevices between markers the software is able to align by hand the various scans. therefore, without marker some surfaces cannot be pre-aligned well.

If the surface to be scanned is rich in details (or areas of interest) it is possible to exploit the fully automatic alignment algorithms that use the shape of the object to achieve the alignments.

Finally it is still possible to align the scans in a semi-automatic way: once the scan is done, it is sufficient to rotate the scan to be aligned in order to place it to large lines in its correct form; the software automatically calculates the alignment.

It should be said that alignment algorithms have taken giant steps in recent times so markers are used exclusively in limit situations

the fourth and last alignment is the most interesting. If a sufficiently high dimensional accuracy is required, it is the only alternative.
I make a summary on this point because it becomes too complicated to get into details.
Basically, before scanning, you use coded markers and support calibrated bars near the object. a reflex camera is used to capture from various views the object after which the result is exported in .txt format (in practice points have been obtained in space).
done this you carry out a normal scan but instead of using traditional we say markers you exploit the points previously detected in order to hook the various scans exactly on these points. It follows a very precise alignment. This is an application to use for very large objects if you want to minimize the progression of the error.

The scanner I was talking about runs all the alignment systems explained above.

everything depends on a tide of factors:

- environmental conditions
- surface to scan
- environmental temperature
- operator skill
- quality of the lens and intrinsic limits of the measuring instrument
- quality of calibration carried out

to great lines I can tell you this because to talk about precision without wondering about what is like talking about fried air.

on a piece of about 200 mm x 200 mm you can achieve accuracy in the order of +/- 0.03 mm (although you work well)

on about 1000 x 1000 mm you can stay in order of 5/6 tenths of millimeter

then there are techniques that allow to manage and minimize volumetric uncertainties. . .

I can tell you this. the greater the quality of the camera and the better lens will be the resolution.

by resolution I mean the possibility to grab the maximum surface detail.
In your case, I think an accurate scanner is more useful dimensionally while for those who need to scan the 50 cent aurel mark of Euros, it naturally serves more resolution :smile:

the scanner I have indicated is equipped with a software that produces a stl. then, if needed, you can complement it quickly (even here depends what you want to do). .

Excuse the Filipino... I hope you're useful

 

[re_solidworks]

Hi.

I answer your questions

As for markers (some also call them targets) the question is debated.

If you have particularly large surfaces and with large curvature (e.g. the hood of a car or a simple table) markers are the only alternative because, without them, no software could be able to determine the correct alignment of the acquisitions. I made you the example of a table not by chance. without the help of markers could be infinite alignment solutions due to the flatness of the surface. calculating the interdevices between markers the software is able to align by hand the various scans. therefore, without marker some surfaces cannot be pre-aligned well.

If the surface to be scanned is rich in details (or areas of interest) it is possible to exploit the fully automatic alignment algorithms that use the shape of the object to achieve the alignments.

Finally it is still possible to align the scans in a semi-automatic way: once the scan is done, it is sufficient to rotate the scan to be aligned in order to place it to large lines in its correct form; the software automatically calculates the alignment.

It should be said that alignment algorithms have taken giant steps in recent times so markers are used exclusively in limit situations

the fourth and last alignment is the most interesting. If a sufficiently high dimensional accuracy is required, it is the only alternative.
I make a summary on this point because it becomes too complicated to get into details.
Basically, before scanning, you use coded markers and support calibrated bars near the object. a reflex camera is used to capture from various views the object after which the result is exported in .txt format (in practice points have been obtained in space).
done this you carry out a normal scan but instead of using traditional we say markers you exploit the points previously detected in order to hook the various scans exactly on these points. It follows a very precise alignment. This is an application to use for very large objects if you want to minimize the progression of the error.

The scanner I was talking about runs all the alignment systems explained above.

everything depends on a tide of factors:

- environmental conditions
- surface to scan
- environmental temperature
- operator skill
- quality of the lens and intrinsic limits of the measuring instrument
- quality of calibration carried out

to great lines I can tell you this because to talk about precision without wondering about what is like talking about fried air.

on a piece of about 200 mm x 200 mm you can achieve accuracy in the order of +/- 0.03 mm (although you work well)

on about 1000 x 1000 mm you can stay in order of 5/6 tenths of millimeter

then there are techniques that allow to manage and minimize volumetric uncertainties. . .

I can tell you this. the greater the quality of the camera and the better lens will be the resolution.

by resolution I mean the possibility to grab the maximum surface detail.
In your case, I think an accurate scanner is more useful dimensionally while for those who need to scan the 50 cent aurel mark of Euros, it naturally serves more resolution :smile:

the scanner I have indicated is equipped with a software that produces a stl. then, if needed, you can complement it quickly (even here depends what you want to do). .

Excuse the Filipino... I hope you're useful

thanks again, all clear and useful for a newbie of this field.

 

[marcof]

the fourth and last alignment is the most interesting.
(cut)
Basically, before scanning, you use coded markers and support calibrated bars near the object. a reflex camera is used to capture from various views the object after which the result is exported in .txt format (in practice points have been obtained in space).
done this you carry out a normal scan but instead of using traditional we say markers you exploit the points previously detected in order to hook the various scans exactly on these points. It follows a very precise alignment. This is an application to use for very large objects if you want to minimize the progression of the error.

I'm interested in the subject, but I don't understand how you use the images from different views obtained with a reflex, in the sense that to have a base of support to reconnect the various scans you should get the 3 coordinates of the individual markers in the same reference system and, I think, you usually have to use topographic tools (total station etc). At least this is what I saw doing both in the field of traditional photogrammetry and when they sell made high resolution scans of large objects, then with many separate surfaces to reconnect.
In your case, even assuming that it is a reflex with a calibrated optics, I always miss the passage to get the coordinates of the individual markers from only images.
apart from the approach like that of photomodeler, whose accuracy and precision does not seem to me to fit in the orders of magnitude you referred to (+/- 0.03 mm), I miss something in the procedure you explained?

 

[Exatem]

a recommendation to re_
I can assure you that you have had a negative experience in reverse engineering so look for people who are extremely reliable.
my experience is witnessed by maxopus who, fortunately, intervened with valuable advice (for the future).

 

[heterodyne]

alignment through photogrammetric markers is suitable for medium/large size objects. as I explained you use a digital high resolution reflex; the goal is to photograph the markers from various views so as to fix their unique position in space.
Points detected, through appropriate software, are generally saved in text format and subsequently imported into 3d scanning software. done this by scanning the object using the same marker; alignment happens automatically; the various 3d measures are rotated from time to time using the grid of points obtained through the reflex.
generally every scanner manufacturer offers this possibility; If you are interested in the topic you can find some interesting point here: http://www.aicon3d.com - oppure anche which: http://www.gom.com/it/sistemi-di-metrologia/riepilogo-dei-sistemi/tritop.html -
however it is not the kind of accessory that serves re_solidworks

 

[marcof]

alignment through photogrammetric markers is suitable for medium/large size objects. as I explained you use a digital high resolution reflex; the goal is to photograph the markers from various views so as to fix their unique position in space.

Yes, this was clear to me, but I did not understand as from the photos, which do not have absolute dimensional references, you can get to the coordinates of the markers in space as you would with a traditional topograph system from which the exact coordinates of the individual markers are derived.

If you are interested in the topic you can find some interesting point here: http://www.gom.com/it/sistemi-di-metrologia/riepilogo-dei-sistemi/tritop.html -

and, I add, especially here:http://www.gom.com/fileadmin/user_upload/industries/shipbuilding_en.pdfI see that it makes explicit reference to a metric chamber, so I assume that the reflex-object set should somehow be calibrated as for any metric chamber. for dimensional references says that they are available "two certified scale bars guarantee the accuracy and process security of the measurement result" which I presume should be visible in individual shootings.
Now it's definitely clearer. thanks to the info.

Can you give a cost-size order between software (tritop), calibrated rows and calibrated reflex? (unless calibration can not be done independently on any reflex with a quality objective)

 

[marcof]

the goal is to photograph the markers from various views so as to fix their unique position in space.

I forgot something. Can you tell me what resolution you can get on the marker coordinates, of course according to the shooting distance or field framed, apart from the number of pixels occupying the reference calibrated line in the image?
Thank you:

 

[heterodyne]

Yes, this was clear to me, but I did not understand as from the photos, which do not have absolute dimensional references, you can get to the coordinates of the markers in space as you would with a traditional topograph system from which the exact coordinates of the individual markers are derived.



and, I add, especially here:http://www.gom.com/fileadmin/user_upload/industries/shipbuilding_en.pdfI see that it makes explicit reference to a metric chamber, so I assume that the reflex-object set should somehow be calibrated as for any metric chamber. for dimensional references says that they are available "two certified scale bars guarantee the accuracy and process security of the measurement result" which I presume should be visible in individual shootings.
Now it's definitely clearer. thanks to the info.

Can you give a cost-size order between software (tritop), calibrated rows and calibrated reflex? (unless calibration can not be done independently on any reflex with a quality objective)

through the reflex you don't take photos as "image files" but extract a dot matrix in space. only then, at the scan, you will make sure to hook the various clouds of points on the matrix of points. In fact the markers are also detected a second time (from the 3d scanner) and referential on the points previously obtained with photogrammetry system.

this technique is twofold because:

1 - sometimes only the coordinates of the points can be sufficient
2 - in other circumstances it is necessary (see above) to use these points as a mere reference, coupled with points clouds and/or mesh

Generally, point 1 is used in activities like gis, landscape, carpentry and naval- point two is definitely useful for reverse engineering, metrology, deformations, crash tests and things like that.

Yes, exactly, to carry out calibration are generally provided with calibrated bars, certified ball bearings, and encoded markers. each manufacturer has its own reference systems. . .

 

[heterodyne]

I forgot something. Can you tell me what resolution you can get on the marker coordinates, of course according to the shooting distance or field framed, apart from the number of pixels occupying the reference calibrated line in the image?
Thank you:

if you refer to the accuracy of the shooting you can consider about 0.006 mm per meter; This depends on the system you use. the tritop, for example, is very expensive. same as regards aicons.

photomodeler is much less expensive but judging by the specifications is much lower than the above systems.. .

in the lower band there would also be rhinophoto: http://www.rhinophoto3d.comThis is not bad for certain applications in the shipbuilding industry. . .

 

[marcof]

if you refer to the accuracy of the shooting you can consider about 0.006 mm per meter

Kill! maybe I didn't understand well... Starting from a 22 mpx reflex file, which then contains about 5800 pixels on the long side, can fix 0.006 mm between two markers distant one meter? If those two markers occupy the long side of the frame means that 1 pixel corresponds to about 0.17 mm. I know that there are algorithms working on the subpixels but I don't understand how to get to that 0.006 mm accuracy on the final processing of the coordinates, since it is almost three times lower than the correspondence between that pixel and real size

photomodeler is much less expensive but judging by the specifications is much lower than the above systems.. .

Yes, definitely, but at this point, using the two systems the same reflex/lents therefore the same shoots and only different software processing (also photomodeler previews the calibration of the machine and the insertion of measurement references known in the images) is incredibible that only via software one solves 0.006 mm on the meter and the other I do not remember how much, but much less of sure, maybe two orders of magnitude.. .