Mosaic Images

[Bethann Gendernalik]

Thankyou Gabriel. That is closer to what I am going for. I should have been clearer with what I am working with. I have a series of images taken with a microscope of a slide and I need to convert it into a single images of the entire slide, as though I had taken one image instead of dozens. The stacking worked to get all the images together, but it is more of a collage.

I am attempting to mosaic 12 rasters. All rasters have been exported to the geodatabase as .tif . After creating the mosaic dataset, I attempt to add all 12 rasters. The process is completed quickly, but only 6 of the rasters are displaying in the mosaic dataset.

The other images do not render at higher or lower scales. All 12 images appear in the "footprint" attribute table. Overviews, stats and pyramids have also been built duing the "Add Raster to Mosaic Dataset" process.

I was able to have all raster's displayed after calculating the correct MaxPS (in attribute table) for the rasters that were not displaying. Once calculated, I overwrote the original MaxPS values that were in the attribute table.

It is possible that the missing rasters have invalid or incomplete metadata, or that they are not properly aligned with the mosaic dataset. You can try to check the metadata of the missing rasters and ensure they are in the same coordinate system and have the same cell size as the mosaic dataset. Additionally, you can try to adjust the mosaic method and blending options in the mosaic dataset properties to improve the display of the rasters.

Does your mosaic have overviews? Zoom in and see if the rasters appear at a very small scale. If they do, then you probably just have overviews that were generated prior to the new rasters being added.

Building overviews didn't work for me, but if you're interested in how I was able to solve this issue check the accepted solution. There was something funky going on with the MaxPS in the images that weren't displaying correctly.

I have 380 DEMs. All the rasters are in .tif format. I am trying to mosaic them together. But the issue is that every raster image is in a separate folder. There are 380 folders for the 380 different rasters. I have tried creating a file geodatabase, adding a raster catalog and then adding the DEMs to the catalog. But it did not work.

All three of these had the same result, so I think if all your "Main Folder X" folders are in a separate "top folder" you should be able to run this command once. I've done a similar process with about 56k raster,

Both images correspond to the same path/taken the same day so they should have the same number and configuration of bands. I have tried also to mosaic only two of contiguous granules (lower right corner granule for the northern image and upper right corner for the southern image) but I get the same error.

Probably you have opened the Sentinel 2 products in their native resolution. It is better to chose one of the three offered resolutions when you try to open an S2 product. Either 10, 20 or 60 meter resolution. Probably you want the 10m resolution.

If you open the 10 m bands from the .xml file the other bands remain. However, I get still the same error when I attempt to mosaic.

The image shows the word view for when the native resolution is open (left) and the 10 m resol is open (right), just in cas it is of any help.

Additionally, what I exaclty would need is to mosaic the granules 1 and 2 (see right side of the image attached), since it covers my study area. Problem is that one is in zone 36 south and the other one in zone 37 south. If I open both in the same zone (either 36 or 37) I end up losing part of my study area.

I have the same problem by making mosaics . SNAP gives me an error message : the bands do not have the same size .

I tried semi auto- classification with Qgis but it really takes a lot of time.

Is there a solution to have Bands of the same size?

Thank you

Hi Javier,

Do you still have the same problem? Was it fixed since then?

Is it also the case with version 4.0?

Please let us know, I inform Julien, Nicolas and Omar about this behaviour, but they surely are already.

Thanks a lot for these elements of observation.

Cheers

Olivier

It went well untill I clicked run and got an error message about my cloud and cloud shadow condition, when I tried to remove the condition the whole mosaic app froze and I could not select anything. I restarted SNAP and started new mosaic, but the tabs Map Project Definition and Variables & Conditions remained frozen. Maybe I need to restart unity? Or reboot?

It's free, fast, and easy. You just pick the big picture you want your mosaic to look like, and then add all the small pictures that make up the mosaic. Then your photos will automatically be arranged to look like your big picture. After that, you can fine-tune your mosaic to your liking. You have full control over the mosaic dimension, mosaic colorization, and which photos go where.

DS9 is supposed to do this, and I know it has in the past. If the '-mosaic' and/or '-smosaic' flags aren't working, then try the options under the File load menu. If there are problems or bug reports you should contact SAO.

Research included checking out a range of open-source and low-cost tooling in the drone mapping space. The minimum three pieces of software needed for drone mapping include mission planning software (how you instruct the drone what needs to be collected), mission control software (how you execute the mission and control the drone in the field) and something to perform the image processing and mosaicing. I did check out a range of options, each of which fulfilled one or more of those roles:

Elevation change over the collection area meant I needed to have terrain-aware mission planning. Otherwise, if all the images were taken from the same flight elevation (height above sea level), the relative distance to the ground would be varied across all the images. This could lead to wonky stitching of the mosaic. This helped me pick Drone Harmony over other options; they had an obvious terrain-aware mission planning option.

I got some advice from the Open Drone Map community and settled on my missions: two separate missions, one running north-south, one 20 skewed. Both higher than normal, both with over 70% front/back and sides overlap. And with the sensor panned 5 up from nadir (nadir => looking straight down). With all that, I took to Drone Harmony to build out these missions.

With the DEM ready, I just needed to outline my mission area, set the parameters for each of my missions, and then sync those missions back to the cloud. Once the plans were synced to the cloud, they were available for me to download on the Drone Harmony mission controller application, which in my case I was running on an Android phone. The regular DJI Mavic Pro setup allows you to control the drone via a combination of hooking your smart phone into the Mavic Pro remote controller. The Drone Harmony mission controller would take over the software role that was usually provided by the DJI flight controller software. After downloading the plans to my Android device, I was ready to fly!

Processing with OpenDroneMap was alarming easy. I already had Docker downloaded, and after bumping up its memory allocation (to help with all the image processing it was about to be running), I just had to drop the images on to my laptop and pop a simple command line instruction into a terminal prompt: docker run -ti --rm -v /Users/twatson/datasets:/datasets opendronemap/odm --project-path /datasets round1 --dsm The result was pretty legit at a first glance.

One thing to note is that the collection area I started with was a large rectangle and this resulting image is an unusual smaller shape. The difference is a result of images that were thick with many homogenous trees. These images could not be meaningfully stitched into the larger mosaic with the other images. Try stitching a bunch of these together!

As for the image quality while largely decent, there were a bunch of image glitches (often referred to as processing artifacts). Here is a sample, including some wonky tennis court lines and wonky asphalt, and a pool slide that certainly does not curve like that.

I did mess about with the processing a little and tried to get images from both missions to process (again with some advice from the OpenDroneMap community). You can see how ugly that got in this next image

These missions were achievable with under one battery of flight each, so much more manageable. I was lucky with the light on one morning; the sun was veiled behind long thin clouds. This took the sharp edges off the shadows, while there still being enough light to pick out features. Collection was seemless through the Drone Harmony mission controller, and once I got home, processing ran smoothly (although I should note that OpenDroneMap also got updates in the gap between the different processing runs, so that could have improved things as well). Overview below:

There are still some interesting artifacts, mainly focused on the challenges provided by the tall trees all over the property. Perhaps these could be worked out with some tweaks to the processing. I have included some choice cuts of these oddities below.

I really enjoyed this process. I got to step back into the nitty-gritty of actual collection ops, and create my bespoke aerial photo mosaic. Now I get to use it to help plan out the garden at this spot!

EDIT: i came across, Isotope, which is almost perfect and i will probably end up using it, unless there's something that fixes my needs exactly. Therefore my question still remains the same...

I suggest Freewall. It is a cross-browser and responsive jQuery plugin to help you create many types of grid layouts: flexible layouts, images layouts, nested grid layouts, metro style layouts, pinterest like layouts ... with nice CSS3 animation effects and call back events. Freewall is all-in-one solution for creating dynamic grid layouts for desktop, mobile, and tablet.

3a8082e126