The biggest difference between HP and FlexScan3D?

Write here about construction, ideas, equipment, tips n tricks etc. related to structured light scanning
User avatar
greenskynet
Posts: 171
Joined: 12 Nov 2021, 19:02

The biggest difference between HP and FlexScan3D?

Post by greenskynet »

🔴🔴🔴 What is one of the biggest difference between HP Scanner and FlexScan3D? 👇👇👇

Answer:
As I know, there are several methods to calibrate and capture 3D data from Cameras+Projector

1- In some methods cameras are calibrated independently of Projector (two cameras are the two corners of the triangle in triangulation method. and object is the third one). So if during the scanning process the projector position or parameters get changed, it makes no problem for the scanning process and there is no need to re-calibrate the system.
This is very Important, because all projectors have some thermal drifts and due to other conditions, their parameters will change time by time (like clocking of the projector which is depend on Graphics Driver of PC. Specially horizontal clocking).
✅most professional scanner (like FlexScan3D and RV) are in this category.

2-in some other methods cameras (or even one camera) are calibrated dependently on Projector. Actually the projector act as one of the corner of triangle (in triangulation method for creating 3D data. . And object is the third one) so projector parameter must be completely fixed. Any small changes in projector can lead to system calibration failure.
🔴most scanner which use only one camera for data capturing (like HP 3D Scanner) are in this category.

in HP v5, the software use two camera for data capturing. but it is different from method 1. Actually it gets two separate 3D data from Cam1+Projector and Cam2+Projector and then create an average data between them. The averaged data is more smoothed and can have a more accuracy than each Cam+Projector data, but never can get as accurate as professional software like Flex and RV.

for more info: viewtopic.php?f=16&t=2305
User avatar
OBNRacerMan
Posts: 372
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: The biggest difference between HP and FlexScan3D?

Post by OBNRacerMan »

HP3DScan has one advantage (precisely because of its peculiarities of working as if two separate scanners, but at the same time "all in one") - this is the moment when it is necessary to scan relatively deep holes or slots. There you can switch to the extended mode "on the fly" and save the "double frame", which is not available for systems in which the projector is not involved in calibration.
This, as well as HP's omnivorous use of almost any camera and the speed of single frame calibration are the only real advantages.
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: LG PF1500G, Cam: Daheng Mercury MER2-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC FM12036MP5 (F2.8/5Mp/12-36mm) x2. And a handheld scanner Creality Ferret
User avatar
greenskynet
Posts: 171
Joined: 12 Nov 2021, 19:02

Re: The biggest difference between HP and FlexScan3D?

Post by greenskynet »

OBNRacerMan wrote: 25 Dec 2021, 10:19 HP3DScan has one advantage (precisely because of its peculiarities of working as if two separate scanners, but at the same time "all in one") - this is the moment when it is necessary to scan relatively deep holes or slots. There you can switch to the extended mode "on the fly" and save the "double frame", which is not available for systems in which the projector is not involved in calibration.
This, as well as HP's omnivorous use of almost any camera and the speed of single frame calibration are the only real advantages.
🟡 About deep area scanning you are right. extend mode is very useful for deep scanning, but the problem is that when we switch to extend mode, the accuracy will drop heavily (the most weakness point of HP is the accuracy in depth). but when the accuracy is not very important and also there is no other way, we can use this method.
but also I should mention that we use some other protocols for deep scanning in dual camera scanners without any need to switch to single camera mode. I will create some posts about this subject in future.

🟢 About using any cameras with the software, as you tested my modified version of Flexscan3D user can use almost any cameras with my modified version of Flex too, including very cheap webcam cameras up to very expensive industrial camera and also security cameras. (viewtopic.php?f=16&t=2305).

🔴 And about calibration, Hp calibration is designed at the most wrong method. I used HP for several years and I have lots of experiences about it.
Calibration in a single frame is a big disadvantage, because the software only will has a single mathematical model of calibration panel and any small deviation on the calibration panel can destroy all of the calibration process. And as you know in the real world it is impossible to produce any calibration panel with zero tolerance (it is an old topic about making a good calibration panel for HP and lots of users never got a good result from it, specially about scanning larger object, mots people experiences miss-alignment problem(viewtopic.php?f=16&t=115&start=10)).

🟢 But in multi frame calibration (like that used in FlexScan3D and other professional software), the software capture several photos of calibration panel and finally create an average mathematical model of the calibration panel from those several captured images. This will reduce the extreme sensitivities to small deviation on the physical calibration panel and allows for some tolerances in the manufacturing process.

So calibration in a single shot (like that used in HP) almost never be accurate enough, and if you don't need to an accurate calibration, and the speed of calibration process is the only parameter for you, it is possible to calibrate Flex only with 10 captured images in less than 40 seconds. (But I always capture more than 30~40 photos for calibration and it takes about 2~3 minutes and yield a perfect accuracy).
User avatar
OBNRacerMan
Posts: 372
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: The biggest difference between HP and FlexScan3D?

Post by OBNRacerMan »

I will not argue about the calibration by one frame - it, of course, has its drawbacks. But it adds another problem - the accuracy of the angle between the panels used in the calibration at HP.
Unfortunately, HP found some limitations on parameter settings, and the HP flat panel calibration was too high in terms of values.

viewtopic.php?f=16&t=2125
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: LG PF1500G, Cam: Daheng Mercury MER2-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC FM12036MP5 (F2.8/5Mp/12-36mm) x2. And a handheld scanner Creality Ferret
User avatar
greenskynet
Posts: 171
Joined: 12 Nov 2021, 19:02

Re: The biggest difference between HP and FlexScan3D?

Post by greenskynet »

OBNRacerMan wrote: 25 Dec 2021, 15:15 I will not argue about the calibration by one frame - it, of course, has its drawbacks. But it adds another problem - the accuracy of the angle between the panels used in the calibration at HP.
Unfortunately, HP found some limitations on parameter settings, and the HP flat panel calibration was too high in terms of values.

viewtopic.php?f=16&t=2125

Actually 90 degree angles between two sides of panel isn't very important. software can compensate any deviation in this angle. so you can change the angle to 80 degree too without any problem.

But you can't change this angle to 0 (flat panel), because at this situation software can't get any information about depth of field and as the result the calibration will done only for a flat volume (flat volume = zero volume = two dimension area)

So as much as you decrease the angles between panels sides, you are decreasing calibration depth info and in a zero angle (flat panel) you will loose all depth data. so at this situation the software can produce only flat scans (loosing all of third dimension data).

Meanwhile producing a flat panel with zero tolerance in impossible too. so there is no chance to get a perfect calibration with HP calibration method.
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: The biggest difference between HP and FlexScan3D?

Post by Micr0 »

greenskynet wrote: 25 Dec 2021, 17:19
OBNRacerMan wrote: 25 Dec 2021, 15:15 I will not argue about the calibration by one frame - it, of course, has its drawbacks. But it adds another problem - the accuracy of the angle between the panels used in the calibration at HP.
Unfortunately, HP found some limitations on parameter settings, and the HP flat panel calibration was too high in terms of values.

viewtopic.php?f=16&t=2125

Actually 90 degree angles between two sides of panel isn't very important. software can compensate any deviation in this angle. so you can change the angle to 80 degree too without any problem.

But you can't change this angle to 0 (flat panel), because at this situation software can't get any information about depth of field and as the result the calibration will done only for a flat volume (flat volume = zero volume = two dimension area)

So as much as you decrease the angles between panels sides, you are decreasing calibration depth info and in a zero angle (flat panel) you will loose all depth data. so at this situation the software can produce only flat scans (loosing all of third dimension data).

Meanwhile producing a flat panel with zero tolerance in impossible too. so there is no chance to get a perfect calibration with HP calibration method.
After a huge amount of work I was able to get good and consistent calibrations with David. However it was finicky. Any carelessness and I got bad calibrations, and had to redo. Calibration has always been a weak link for David. For that reason alone, I have been trying to decide whether to go with RV of Flex. I haven't tried RV but have seen OB get very goo results. The one thing that really bugs me about Flex is the color scheme. That's petty I know, but I really don't like it.
µ
User avatar
greenskynet
Posts: 171
Joined: 12 Nov 2021, 19:02

Re: The biggest difference between HP and FlexScan3D?

Post by greenskynet »

Micr0 wrote: 25 Dec 2021, 19:25 After a huge amount of work I was able to get good and consistent calibrations with David. However it was finicky. Any carelessness and I got bad calibrations, and had to redo. Calibration has always been a weak link for David. For that reason alone, I have been trying to decide whether to go with RV of Flex. I haven't tried RV but have seen OB get very goo results. The one thing that really bugs me about Flex is the color scheme. That's petty I know, but I really don't like it.
Actually RV has some big limitations:
1- very limited camera choice (only accept some Daheng cameras)
2- Russian language
3- the type of marker which it use (only a circle dot, so sometimes it detect wrong markers)
4- very special calibration panel
5- and I think not good support for out of Russia
6- and now regards to existence of my modified version of Flexscan3D, RV price is not competitive (viewtopic.php?f=16&t=2305).
User avatar
OBNRacerMan
Posts: 372
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: The biggest difference between HP and FlexScan3D?

Post by OBNRacerMan »

greenskynet wrote: 25 Dec 2021, 19:38 2- Russian language
RV has at least two interface languages - Russian and English
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: LG PF1500G, Cam: Daheng Mercury MER2-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC FM12036MP5 (F2.8/5Mp/12-36mm) x2. And a handheld scanner Creality Ferret
mading
Posts: 307
Joined: 31 Jan 2017, 13:09

Re: The biggest difference between HP and FlexScan3D?

Post by mading »

greenskynet wrote: 25 Dec 2021, 19:38 Actually RV has some big limitations:
1- very limited camera choice (only accept some Daheng cameras)
2- Russian language
3- the type of marker which it use (only a circle dot, so sometimes it detect wrong markers)
4- very special calibration panel
5- and I think not good support for out of Russia
6- and now regards to existence of my modified version of Flexscan3D, RV price is not competitive (viewtopic.php?f=16&t=2305).
The price is good: official software, good calibration panels, sw support.
And the english version is perfect, no mispelling or odd language.

I had no problem with markers.
No problem with support in Italy.
I will suggest it to anybody willing to use official software.
Excellent precision and accuracy.
Reasonable scanning time also for dark surfaces.
LG PF50, LG PF1500, RangeVision DIY: 2x DahengMer630, 2X12 and 2X16 mm 5Mp ZK lenses, RV turntable
User avatar
greenskynet
Posts: 171
Joined: 12 Nov 2021, 19:02

Re: The biggest difference between HP and FlexScan3D?

Post by greenskynet »

mading wrote: 26 Dec 2021, 16:57 The price is good: official software, good calibration panels, sw support.
And the english version is perfect, no mispelling or odd language.

I had no problem with markers.
No problem with support in Italy.
I will suggest it to anybody willing to use official software.
Excellent precision and accuracy.
Reasonable scanning time also for dark surfaces.
Hi
how are you?
glad to see your new posts in the forum

As I exactly follow all thread of the forum, I know you had lots of trouble to get a working version of RV as you are not a Russian.
viewtopic.php?p=3962#p3962

000.jpg

Also as I know, scanning dark surfaces is not a big issue. Both of Hp and Flexscan3D can also scan a fully black surface.
If you want I can share some results with Flexscan3D

and about markers, there is a big difference between a "ring" (like what Flex is using) and a "dot" (Like that RV is using)
I used many other 3d scanning and photogrametery softwares. when using only a "dot" as a marker, software can detect any black and even semi-round area on the object as a marker, so it can yield some errors. But when using "Ring" as a marker, both internal and external edge of the ring must be recognized to identify the ring as a marker, so there is a lower chance for any error.

Anyway, RV also can be a good software if you can deal with some of its major limitations (specially camera choice limitation). but I always prefer FlexScan3D as a very engineered and professional software.
Post Reply