Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Steve003

Pages: 1 [2] 3 4 ... 12

great tutorial on using multiple photo sessions of different sides of an object and combining into one chunk for better results, but how is this done if the camera was moved round the object then the object turned upside down and another photo session done, then another of a related part that had broken off.


The tutorial on the basic steps of metashape at:-
(see attached)
Has the user doing:-
1. Align Photos
2. Make a Dense Point Cloud, remove points unwanted with delete tools.

3. Build mesh
AFTER THE DENSE POINT CLOUD IS MADE, it says a choice then exists, use either Dense Point Cloud or Depth maps.
Depth maps supports GPU acceleration (GPU being the video processor) and provides better results for minor details.
It then shows the panel and settings for the depth map approach.


Click OK to start the build.

I bring our attention to:-
AFTER THE DENSE POINT CLOUD IS MADE, it says a choice then exists, use either Dense Point Cloud or Depth maps.

As I wish to get better detail, exactly what are the steps I should be following ? It seems one has to make dense point cloud then save depth maps then make mesh using depth maps and choose reuse them !  Where is this option to save depth maps in the Dense Point Cloud build ?

conflicts with advice :- after align photos skip DensePoint Cloud, as its slow and if using depth maps you don’t need the Dense Point Cloud !
I like speed and detail, so why are we being told to go via Dense Point Cloud here ?

I am now puzzled, I have been aligning photos then choosing build mesh use depth maps.

Am I doing wrong ?

it even goes on to say:-
If after the previous stage – Depth maps were built and save in project, we recommend using the parameter – Reuse Depth maps. It shows the Build mesh panel with depth maps selected and reuse depth maps ticked.

so again its saying make the dense point cloud then save the depth maps, I wasn’t aware one had an option to save or not to save the depth maps.

my method by not making Dense point cloud wont be making depth maps and then seeing an option to save them or not it appears.

I AM TOTALLY CONFUSED NOW. just done 9 models my way.

As I wish to get better detail, exactly what are the steps I should be following ?


General / Re: when using mesh from depth maps deleted areas reappear
« on: August 13, 2022, 01:22:27 AM »

In your case you could either adjust the reconstruction region (bounding box) to exclude certain areas from reconstruction.

newspaper has appeared, the region box wont allow non box shapes to be drawn, as the newspaper sees a W shaped item standing on it I cant get that box to delete the paper the item stands on, as it would need to be drawn onto thhe newspaper in the V area.

so I need to edit it out manually

Alternatively, you could delete unwanted points from the sparse cloud, mesh it and use the "import masks from model" to mask the images. Masked areas will not be reconstructed.

the sparse cloud is I presume the somewhat non dense dotty mess after align photos, impossible to see what you are doing trimming up close to the model.
the next option exists with the violet coloured mesh, or with the not so bad image mapped to it.

I do so.

then I follow 'import mesh from model, right clicking on an image, choosing that from the list, apply to all cameras, it does its green bar progress thing, then I click the mask icon above the photos, AND SHOCK HORROR THE W SHAPE ITEM BUT ALSO THE NEWSPAPER IS IN WHITE, its ignored the fact that just the W is the model and showing as such in that mesh violet stage, with all newspaper removed.



General / Variable density of polygons to suit surface ?
« on: August 12, 2022, 10:10:50 PM »
It strikes me that its wrong to choose ultra high to get the detail of items that sit on an otherwise smooth or slightly undulating surfaces. Take a vintage tractor, many smooth areas such as bonnet, sides, wheel hubs and rims, tyre sides (except for the lettering) cab structure, etc, with fasteners, wiring, and more detailed components showing lower on the sides, tread, axle joints etc.

How can one choose a setting to see a low poly count get used on smooth and a high poly count on detail.

I choose ultra high and get far more than I need, decimate attacks ALL polys and I lose the detail.

This is a fundamental need of the entire process and world of meshes AND EXTREMELY IMPORTANT.

How is this achieved, as I take the mesh into Rhino3D and its like wading through molasses, hellishly slow, when most of it doesnt need that density at all.

At times unworkable due to the speed hit, all for the detail parts I need on that tractor for example and I am looking at areas with far to many polys for what they need.

I dont see vids on this, everyone is doing rocks and aerial photos and fossils it seems.


General / Re: when using mesh from depth maps deleted areas reappear
« on: August 12, 2022, 09:31:12 PM »
Arie...It seems you might want to optimize your data acquisition methods. I've been using Metashape for a long time and can't even remember the last time I had alignment issues.

All my photos are correctly exposed evenly lit, 12 out of 22 failed the other day, ran them in photoscan and they all were aligned.
I had 1 fail in each of 2 sets, subject very matt, rusty covered in the most ideal surface pattern sat on newspaper, looking at the failure it had good progression from the previous one and into the next one that got aligned.

the photos couldnt be better, bounce flash off the ceiling in all cases or direct flash, only suitable of item has very matt surface, all are adjusted in camera raw in photoshop, detailed revealed in shadows, lighting variances evened out, honesyly those that fail have no apparent reason to at all.

Now if I feel the background might be causing issues then I want to mask it out of the process, and thats why I want to know the easy way of doing so on 144 photos for example.

But if one has to make the model before using mask from model, thats a catch 22, how does one overide that one ?

I am still baffled by that.

if only someone would even reply to my asking how to use Photoshop to make a mask.

I will be sending Agisoft the photo sets of those that fail align in metashape and get aligned by photoscan.

usually metashape lets me down on alignment and photoscan never ever does.

polarnick, unfortunately I dont use metashape pro.  I do wonder if pro also fails in same way as Standard.


General / Re: Where are the video tutorials on metashape ?
« on: August 12, 2022, 09:19:38 PM »
No I wasnt, thats great news , if they cover the things I am forever needing to know about.
I will study them and see.

Many Thanks


Thanks, I will follow those through.


General / Re: Ortho where is 2D grid and a 2D rotate tool ?
« on: August 12, 2022, 04:37:45 PM »
Please anyone, I am forever spoiling my aligned model or region as there is no way it seems to reset the rotation sphere to have its coloured red blue green lines to suit ortho, i.e round the edge of the sphere and exact down middle, where I wish to rotate in 2D mode, as ortho is 2D.

and we need a grid in front side as well as top views, how ?


General / Re: when using mesh from depth maps deleted areas reappear
« on: August 12, 2022, 04:25:07 PM »
No alignment errors in the project I am running (though in 2 of the sets of photos one wouldnt align, no apparent reason,  solved by using photoscan. as always.)

but in an 'its possible' scenario, for us all,  still hoping someone somewhere can explain the role 'masks' play or could play if a background was causing problems, lets say a human was in a few shots moving about., when photographing a tractor at a show for example. the workflow steps to solve such.

good for us all to know.


General / Re: when using mesh from depth maps deleted areas reappear
« on: August 12, 2022, 03:33:15 PM »
Hi, I will try both suggestions.
As the region is a box and as this is an arched item standing on the newspaper, its a sort of W shape, the region cannot be made to exclude the newspaper under the arched item.

If the background is such that its confusing the build, and as 'mask from model' apply to all cameras,  clearly requires the build to succeed to be used, it strikes me that mask from model has a use after the model is made, and cannot help make the model in the first place.

Could someone explain the steps involved from importing photos,  and how useful mask from model is and its use and purpose and how it avoids an apparent catch22 situation ?


General / unable to see the thread I just replied to
« on: August 12, 2022, 03:22:44 PM »

I replied to this yet its not visible under general>masking which is the address is shows whilst doing so.

I expect to see a post called masking in the general section but nothing shows.

so as such no one especially Alexey will see it.

having thousands of views it was proving useful especially awaiting the answers.


Such wouldnt work.
The item was sat on a picnic table covered in flat newspaper which was taped down , each time the item was moved to photograph its underside or topside etc (called a set for each photo set) the newspaper was then different to the item.
I happened to make a mistake and fed two sets to metashape at the same time, and when align photos was done I had a double image going on.
So I had in fact done what you suggest by accident.

If I were to spend many hours masking out the background and newspaper from a few hundred photos then yes that would work. I dont have time to do such, as this is one of many projects, and I am years behind as it is. That metashape is often failing to align a GOOD set of images and photoscan manages it perfectly everytime is also a real pain as it messes me about. Wish Agisoft would solve that failure with metashape, they are aware and have photosets that prove it.

As such the newspaper helps as it provides a nice detailed part of the subject to ensure good alignment, bit like targets but far better.

That it keeps reappearing despite trimming it off in initial cloud after align, as I am using mesh from depth maps, IS DOING MY HEAD IN.

another issue Agisoft should really fix as its a fundamental flaw.

Anyhow it sounds as if I should trim away the model (I have each set as textured now) leaving only quality areas to merge.

Should I take the shaded stage and trim and merge or take the final textures stage and trim and merge ?

I wondered if 'merge chunks' was 'intelligent ' enough to merge the best parts and discard the not so good parts, lessening our workload.


we all know those areas that look like the base board has a snow drift up to the model, covered in texture thats gone blurred.

I have a metal object photographed on a table. laid on its side then standing upright, so  have two chunks 1. top and sides, 2. underside and sides,

can I just go ahead and merge chunks, after first bringing them into the relationship they are, which in effect will see them intertwined with each other, and expect merge chunks to remove the not so good bits, or do I have to trim off the parts the other chunk will provide and repeat such trimming on the second chunk ?

What is the correct way ?


General / when using mesh from depth maps deleted areas reappear
« on: August 11, 2022, 10:26:24 PM »
Metashape 1.8.2
I like to tidy up the model, when after align photos I delete the base its sitting on for example, currently a newspaper is my victim !
I am using mesh from Depth maps to see if its better, HOWEVER the base (newspaper) keeps reappearing.

It is NOT POSSIBLE to delete an unwanted base a model sits on, when using workflow ' mesh from depth maps' is my conclusion.

Am I right ?

surely we dont want that tedious time spent erasing up to the model to be wasted,  what is the solution to this ?


General / Re: Numpad navigation removed from 1.8.2 why ?
« on: August 11, 2022, 12:29:08 AM »
Hi, Now you might be onto something there, I have never considered the numlock button , never in life ever used it, maybe I hit it off by accident.
I have remapped the numpad keys to 5 toggle ortho/persp, 4 left side, 6 right side,  2 front 8 back 9 top and 3 bottom,
I would need to run agisoft as it comes out of the box to see if numlock makes a difference.
on that note, i launch photoscan which I havent altered the key allocations and yes numlock sees 5 work as toggle, as well as 4 2 8 6, cant fathom out what top view is as the key allocs in preferences are not there. where are they ?

I hit 9 and got a view suiting 3d glasses..   :) must explore further !


Pages: 1 [2] 3 4 ... 12