chupacabra20 you will find that you spend a lot of time processing your photos , hours And sometimes a whole day. This type of operation puts your PC to the test. I had no luck trying to get this software to work. A lot of wasted time. I was looking for a cheaper alternative to Pix4d. Pix4d is very forgiving and the whole process makes sense. Very pricey but it works and you don’t have to be a GIS Tech to get it to work. This video however is very informative and I’m hat goes off to the woman who made it.
You are done all perfect. The problem are the photopoints. For second time I got the same mistake. The photo marks must be at the corners of the area you want to scan. If you put the photo marks straight line like these you have you will have error at the datums when you export surface
Hello, I have a rather urgent question. I performed a "low" quality treatment (cloud, alignment...etc). If I want to switch to "high" quality, is there a way to reuse the work done in low quality to reduce the working time or do I have to redo everything from the beginning in high quality? Thanks.
Hi! If you are using GCP, and you are optimizing sparse point cloud after alignment, you don't realign the photos but you have to build the dense point cloud again
@@justynajeziorska6205 Thanks for answer. For example, if i add photos in order tjo optimize the result in some places where there are a lack of details, so i should start again or there is a solution to maintain the current result and just modify that place. Thank you so much.
@@plentyfacts8301 There is a way of aligning just a subset of photos. Just select the group you want to deal in the reference pane and right click. There is an option: Align selected cameras. Other than that you have to go through the building of the dense cloud anyway - each realignment, even of a small part, resets the dense point cloud
@@plentyfacts8301 If you realign just a selected group of photos, the rest of them remain aligned. but if you have built the dense point cloud and then try to realign, the dense point cloud is reset. The Alignment stays
@@justynajeziorska6205 Thank you so much. I got a last question, i have a very large surface to modelize, so i divided it to parts. So i have to fusion them, i tried with metashape, it doesnt give good results, i would like to manipulate manually the two parts and regroup them, is there any easy way to do it with another software? Thank you again !
Hi, thank you for the video, it help me to place my markers. My points are in the correct Lat Long position, but they appear in the real elevations, I mean, I want to correct the elevations of my dense cloud using the elevations of my markers, but I can, the markers appear be floating in the dense clouds. Any advise? thanks you so much
First, let me clarify - by "no data" you mean that you have only pictures and no flight log and no GCPs set before the flight? I also assume that you have already flown and looking for the way to georeference the model without going back, setting the targets (GCPs) and retaking the photos with a new flight mission. 1. Some of the cameras have GPS and save the position of the photo in the EXIF file. This is rare though and the positions are very inaccurate. Additionally - you have a position of the CAMERA, so the altitude registered is higher that the ground. There are additional problems with varying ceiling of the flight path. I don't recommend this way 2. Look at your pictures and find so-called "photo-identifiable points". Corners of the buildings, permanent objects that you can precisely locate on multiple photos. It is important that they will be on the ground and they will be well distributed (you need at least 3, preferably 5 or more). Then go to the field and measure them with the GPS (the more accurate your GPS is the more accurately positioned will be your model). Come back to your Agisoft project, find previously measured objects on the photos and place the markers there inputting the measured coordinates and altitude. Optimize alignment based on the newly created markers. 3. Process the data and create not georeferenced outputs. Then in the GIS software find the measured photo-identifiable points and georeference the orthophoto. I don't recommend this since you can accurately georeference only the orthophoto. Not DSM and not the point cloud. My advice is: always set the targets before the flight.
hi, very good explanation. but i still have some questions. i hope you can help me. i have old remote sensing data, but there are no GCP's on it. So i have to search/set them in ArcGIS first and convert them to a .txt-file. but i dont know which coordinate system i have to use. i think if i look for GCP's in ArcGIS i have to look them up for the coordinate system i want to set in Photoscan. Im not sure how to make it.
does the lat/long has to be from the same coordinate system like i set it in Agisoft? for example: i want to georeference my data for WGS84/Zone 3N, so i set the coordinate system in agisoft to this. then i need do look manually for some gcp's (like you said in the comment below). i search for them in arcgis. do i need to set the coordinate system for the WMS i uploaded to the same coordinate system (WGS83/Zone3N) or can i use the lat/long/easting from the geographic WGS 1984?
Hi Hoang, Unfortunately, ArcGIS is not really good in working with projections. Reprojecting "on the fly" and adjusting view to the arbirtarly set projection of the layers creates a lot of confusion. I highly recommend using open source GRASS GIS, where the projections are set permanently and can't be changed for the project (therefore you are always sure what are you working with). What you can do is to check the coordinate system of the environment you were setting the GCPs in ArcGIS (right click on the LAYERS in the layers tree, 'Properties' in the drop down menu, and in the 'Coordinate System' tab, chek the 'Current coordinate system') then set the same system in Agisoft ('Settings' icon in the reference Pane and look up the coordinate system, the easiest way is through EPSG code).
If the Agisoft will be set for WGS84/Zone 3N you will see the EASTING and NORTHING column for the GCP coordinates, therefore you can't use lat/long there.
hi Justyna and thank you for your quick reply, what i just did (before reading your reply): i opened a WMS-Layer of the region where i should search my GCP points. The WMS Layer was set to geographic WGS 1984. Then i created a point.shp which i will edit my GCP points. I set the coordinate system of the shp-file to WGS84/Zone 3N. Unfortunatelly i could not edit the WMS-Layer, so i added a worldbase map. Then i started editing. (Should be alright so far?) In the attribut table i added some fields where i calculated the X,Y (The question is, is it the X,Y from the geographic or the Zone3N coord. system?) then i had a problem, because i needed some Z-Values. so a DEM was required. i downloaded some hgt-files for the region. After extracting the values to an .txt-file i tried to import it to agisoft. it was hard work to readjust the txt-files so agisoft can read it properly. i hope that is the right workflow. looking for your reply, you are a lifesaver! :)
Hoang, did you "reproject" the shapefile in ArcGIS or just set the coordinate system of the layer? The reprojection issues are one of the reasons why I abandoned ArcGIS. There are so many problems when it comes to correct reprojection and you can't really look into what the software is doing because it's not opensource. You definitely need to reproject the vector with the GCP points.. What is the resolution of the data that you are working with? Usually, the UAS data aims to produce high-resolution outputs and the precision here plays a crucial role. And if you are taking the z values from the bare earth DEM, remember that RGB data always shows the surface, therefore you are creating DSM, not DEM. If your DEM has like 1m vertical resolution, you can't precisely reference your model with it.
Bardzo fajny filmik, dziękuję i proszę o więcej, może jakiś film wyjaśniający w poszczególnych etapach generowania modelów możliwe opcje do wyboru np. pair selection: generic, reference, disable itp.. Pozdrawiam serdecznie :)
Dziękuję i cieszę się, że instruktaż okazał się przydatny. Na jesieni ruszamy z warsztatami przetwarzania danych z UAS i wtedy powinno się pojawić więcej filmików. Pozdrawiam również!
This was my old computer and I don't remember the specs (I think it was just 8 GB RAM and 2 or 4 cores). In this video, there is no processing shown, so the speed depends on the graphics card. I have a new computer now that works really great and is much faster with the processing. Processor Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz, 3696 Mhz, 6 Core(s), 12 Logical Processor(s)Installed Physical Memory (RAM) 32.0 GB Installed Physical Memory (RAM) 32.0 GB NVIDIA GeForce GTX 1080 Graphics Cards
Thanks. Excellent. I've seen a lot of videos explaining some procedures and this one is just great.
This software gives me a great appreciation for Pix4D.
can you explain why?i am still deciding between pix4d and metashape.this software seems to be better but i am a beginner.
chupacabra20 you will find that you spend a lot of time processing your photos , hours And sometimes a whole day. This type of operation puts your PC to the test. I had no luck trying to get this software to work. A lot of wasted time. I was looking for a cheaper alternative to Pix4d. Pix4d is very forgiving and the whole process makes sense. Very pricey but it works and you don’t have to be a GIS Tech to get it to work. This video however is very informative and I’m hat goes off to the woman who made it.
Thanks for the video. It helped me a lot for exactly what I was looking for. :)
You are done all perfect. The problem are the photopoints. For second time I got the same mistake. The photo marks must be at the corners of the area you want to scan. If you put the photo marks straight line like these you have you will have error at the datums when you export surface
Una volta inserite le coordinate dei target, come si fa a georeferenziare?
Hello, I have a rather urgent question. I performed a "low" quality treatment (cloud, alignment...etc). If I want to switch to "high" quality, is there a way to reuse the work done in low quality to reduce the working time or do I have to redo everything from the beginning in high quality?
Thanks.
Hi! If you are using GCP, and you are optimizing sparse point cloud after alignment, you don't realign the photos but you have to build the dense point cloud again
@@justynajeziorska6205 Thanks for answer.
For example, if i add photos in order tjo optimize the result in some places where there are a lack of details, so i should start again or there is a solution to maintain the current result and just modify that place. Thank you so much.
@@plentyfacts8301 There is a way of aligning just a subset of photos. Just select the group you want to deal in the reference pane and right click. There is an option: Align selected cameras. Other than that you have to go through the building of the dense cloud anyway - each realignment, even of a small part, resets the dense point cloud
@@plentyfacts8301 If you realign just a selected group of photos, the rest of them remain aligned. but if you have built the dense point cloud and then try to realign, the dense point cloud is reset. The Alignment stays
@@justynajeziorska6205 Thank you so much. I got a last question, i have a very large surface to modelize, so i divided it to parts. So i have to fusion them, i tried with metashape, it doesnt give good results, i would like to manipulate manually the two parts and regroup them, is there any easy way to do it with another software? Thank you again !
what is the next process????
Hi. I can not see X, Y and Z values in the software. Please advise me how to setup in the software to see the values. Thanks.
where did you get the GCP?
Thank you for the video, please can i know how the file .txt is made?
I;ve no GCP Point, so i collect it from google earth, but after final processing my image is distorted. How can i solve it?
Hi, thank you for the video, it help me to place my markers. My points are in the correct Lat Long position, but they appear in the real elevations, I mean, I want to correct the elevations of my dense cloud using the elevations of my markers, but I can, the markers appear be floating in the dense clouds. Any advise? thanks you so much
You are awesome! Thank you so much.
Do you need to realign after place markers?
what are the steps after optimizing?
after you put markers what step you do? Is align photo? is this markers consider in coordinates estimation? Thanks
The next step is optimizing the alignment. The preliminary photo adjustment will be corrected based on the placed markers and their coordinates.
is there a way how to geotagging pictures if we have only pictures no data ?
First, let me clarify - by "no data" you mean that you have only pictures and no flight log and no GCPs set before the flight? I also assume that you have already flown and looking for the way to georeference the model without going back, setting the targets (GCPs) and retaking the photos with a new flight mission.
1. Some of the cameras have GPS and save the position of the photo in the EXIF file. This is rare though and the positions are very inaccurate. Additionally - you have a position of the CAMERA, so the altitude registered is higher that the ground. There are additional problems with varying ceiling of the flight path. I don't recommend this way
2. Look at your pictures and find so-called "photo-identifiable points". Corners of the buildings, permanent objects that you can precisely locate on multiple photos. It is important that they will be on the ground and they will be well distributed (you need at least 3, preferably 5 or more). Then go to the field and measure them with the GPS (the more accurate your GPS is the more accurately positioned will be your model). Come back to your Agisoft project, find previously measured objects on the photos and place the markers there inputting the measured coordinates and altitude. Optimize alignment based on the newly created markers.
3. Process the data and create not georeferenced outputs. Then in the GIS software find the measured photo-identifiable points and georeference the orthophoto. I don't recommend this since you can accurately georeference only the orthophoto. Not DSM and not the point cloud.
My advice is: always set the targets before the flight.
Merci infiniment. Super interessant
수동 마커 기능은 standard에는 없습니다. 프로페셔널에만 있다네요.
hi, very good explanation. but i still have some questions. i hope you can help me.
i have old remote sensing data, but there are no GCP's on it. So i have to search/set them in ArcGIS first and convert them to a .txt-file. but i dont know which coordinate system i have to use. i think if i look for GCP's in ArcGIS i have to look them up for the coordinate system i want to set in Photoscan. Im not sure how to make it.
does the lat/long has to be from the same coordinate system like i set it in Agisoft?
for example: i want to georeference my data for WGS84/Zone 3N, so i set the coordinate system in agisoft to this. then i need do look manually for some gcp's (like you said in the comment below). i search for them in arcgis. do i need to set the coordinate system for the WMS i uploaded to the same coordinate system (WGS83/Zone3N) or can i use the lat/long/easting from the geographic WGS 1984?
Hi Hoang,
Unfortunately, ArcGIS is not really good in working with projections. Reprojecting "on the fly" and adjusting view to the arbirtarly set projection of the layers creates a lot of confusion. I highly recommend using open source GRASS GIS, where the projections are set permanently and can't be changed for the project (therefore you are always sure what are you working with).
What you can do is to check the coordinate system of the environment you were setting the GCPs in ArcGIS (right click on the LAYERS in the layers tree, 'Properties' in the drop down menu, and in the 'Coordinate System' tab, chek the 'Current coordinate system') then set the same system in Agisoft ('Settings' icon in the reference Pane and look up the coordinate system, the easiest way is through EPSG code).
If the Agisoft will be set for WGS84/Zone 3N you will see the EASTING and NORTHING column for the GCP coordinates, therefore you can't use lat/long there.
hi Justyna and thank you for your quick reply,
what i just did (before reading your reply):
i opened a WMS-Layer of the region where i should search my GCP points. The WMS Layer was set to geographic WGS 1984. Then i created a point.shp which i will edit my GCP points. I set the coordinate system of the shp-file to WGS84/Zone 3N. Unfortunatelly i could not edit the WMS-Layer, so i added a worldbase map. Then i started editing. (Should be alright so far?)
In the attribut table i added some fields where i calculated the X,Y (The question is, is it the X,Y from the geographic or the Zone3N coord. system?)
then i had a problem, because i needed some Z-Values. so a DEM was required. i downloaded some hgt-files for the region.
After extracting the values to an .txt-file i tried to import it to agisoft. it was hard work to readjust the txt-files so agisoft can read it properly.
i hope that is the right workflow. looking for your reply, you are a lifesaver! :)
Hoang, did you "reproject" the shapefile in ArcGIS or just set the coordinate system of the layer? The reprojection issues are one of the reasons why I abandoned ArcGIS. There are so many problems when it comes to correct reprojection and you can't really look into what the software is doing because it's not opensource. You definitely need to reproject the vector with the GCP points.. What is the resolution of the data that you are working with? Usually, the UAS data aims to produce high-resolution outputs and the precision here plays a crucial role. And if you are taking the z values from the bare earth DEM, remember that RGB data always shows the surface, therefore you are creating DSM, not DEM. If your DEM has like 1m vertical resolution, you can't precisely reference your model with it.
Bardzo fajny filmik, dziękuję i proszę o więcej, może jakiś film wyjaśniający w poszczególnych etapach generowania modelów możliwe opcje do wyboru np. pair selection: generic, reference, disable itp.. Pozdrawiam serdecznie :)
Dziękuję i cieszę się, że instruktaż okazał się przydatny. Na jesieni ruszamy z warsztatami przetwarzania danych z UAS i wtedy powinno się pojawić więcej filmików. Pozdrawiam również!
Great explanation. I see your pc is extremely fast. Could you please tell me which processor, ram and video card it has?
This was my old computer and I don't remember the specs (I think it was just 8 GB RAM and 2 or 4 cores). In this video, there is no processing shown, so the speed depends on the graphics card.
I have a new computer now that works really great and is much faster with the processing.
Processor Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz, 3696 Mhz, 6 Core(s), 12 Logical Processor(s)Installed Physical Memory (RAM) 32.0 GB
Installed Physical Memory (RAM) 32.0 GB
NVIDIA GeForce GTX 1080 Graphics Cards
Thanks
Гуд епта!
THEN what???
You would need to explain me what do you mean by "then what" :-)