RTK with Smart Controller is working! And Base Station!

[Dave Pitman]

se il drone tramite il suo software memorizza i dati gnss sul file jpeg che viene generato insieme al file dng (es. scatto in condizioni dng + jpeg) non capisco che, avendo tutte le informazioni, non sia in grado, il software, per generare un file xmp con le stesse informazioni, già presenti, memorizzate su un file che deve essere semplicemente trasferito. Trovo questo assurdo, di facile soluzione e che ha un impatto notevole dal punto di vista operativo quando lavori in zone dove puoi andare una sola volta (ad esempio per problemi di autorizzazione) e se hai foto con ombre e luci da sistemare sono limitati nel farlo. Per quanto riguarda la qualità delle foto jpeg, attenzione a non confondere le cose. il P4rtk corre che è una bellezza, magari anche l'evo 2 rtk sarà così ma dal punto di vista del software deve mangiare ancora tanta pasta al sugo! Non voglio essere polemico ma darmi una giustificazione plausibile o dirmi che dal punto di vista informatico non è possibile, perché se lo è, come credo e anche facilmente, mi sembra assurdo non farlo. un saluto da un utente al momento, visti gli innumerevoli guai dell'evo 2 rtk, è davvero deluso. ciao

Hi Paolo, I do not have an EVO yet. But, if you are intent on shooting .dng, and your workflow allows for the additional time needed for .dng image storage on the drone, you should be able to configure Exiftool to write all of the exif data from the original .jpg copies to new .jpg copies derived from edited .dngs.

One thing I'm not certain about, and maybe Ken can comment, is if there is any actual time differential between the "capture" of the .jpg vs the .dng. I suspect that any differential is in the processing and caching and not the capture. But I am not possitive.

 

[Blade Strike]

Hi Paolo, I do not have an EVO yet. But, if you are intent on shooting .dng, and your workflow allows for the additional time needed for .dng image storage on the drone, you should be able to configure Exiftool to write all of the exif data from the original .jpg copies to new .jpg copies derived from edited .dngs.

One thing I'm not certain about, and maybe Ken can comment, is if there is any actual time differential between the "capture" of the .jpg vs the .dng. I suspect that any differential is in the processing and caching and not the capture. But I am not possitive.

Hi Paolo, I do not have an EVO yet. But, if you are intent on shooting .dng, and your workflow allows for the additional time needed for .dng image storage on the drone, you should be able to configure Exiftool to write all of the exif data from the original .jpg copies to new .jpg copies derived from edited .dngs.

One thing I'm not certain about, and maybe Ken can comment, is if there is any actual time differential between the "capture" of the .jpg vs the .dng. I suspect that any differential is in the processing and caching and not the capture. But I am not possitive.

That’s actually something I didn’t think of. I’m just not sure if the subsystem can handle jpg+DNG and keep the trigger times accurate.

should be easy to test if he wants to spend the time.

 

[jmason702]

se il drone tramite il suo software memorizza i dati gnss sul file jpeg che viene generato insieme al file dng (es. scatto in condizioni dng + jpeg) non capisco che, avendo tutte le informazioni, non sia in grado, il software, per generare un file xmp con le stesse informazioni, già presenti, memorizzate su un file che deve essere semplicemente trasferito. Trovo questo assurdo, di facile soluzione e che ha un impatto notevole dal punto di vista operativo quando lavori in zone dove puoi andare una sola volta (ad esempio per problemi di autorizzazione) e se hai foto con ombre e luci da sistemare sono limitati nel farlo. Per quanto riguarda la qualità delle foto jpeg, attenzione a non confondere le cose. il P4rtk corre che è una bellezza, magari anche l'evo 2 rtk sarà così ma dal punto di vista del software deve mangiare ancora tanta pasta al sugo! Non voglio essere polemico ma darmi una giustificazione plausibile o dirmi che dal punto di vista informatico non è possibile, perché se lo è, come credo e anche facilmente, mi sembra assurdo non farlo. un saluto da un utente al momento, visti gli innumerevoli guai dell'evo 2 rtk, è davvero deluso. ciao

English, please...

 

[Dave Pitman]

That’s actually something I didn’t think of. I’m just not sure if the subsystem can handle jpg+DNG and keep the trigger times accurate.

should be easy to test if he wants to spend the time.

A quick test should tell the tale. And, for the timing question, process the originals in one project and the copies in another and see what the metrics are. If they are the same, then all is good.

 

[Blade Strike]

A quick test should tell the tale. And, for the timing question, process the originals in one project and the copies in another and see what the metrics are. If they are the same, then all is good.

Yep, exactly. I’m just worried about it hammering jpg + DNG that fast. If intervals are 3-4 sec I can see. But if asking 1.5-2sec, I’m not sure.

I would go out and test today but still getting the trailer ready for winter. I’m ready for a hotel vacation, a lot less work.

 

[paolo]

English, please...

if the drone through its software stores the gnss data on the jpeg file that is generated together with the dng file (e.g. shooting in dng + jpeg conditions) I do not understand that, having all the information, the software is not able to generate an xmp file with the same information, already present, stored on a file to be transferred. I find this absurd, easy to solve and that has a significant impact from an operational point of view when you work in areas where you can only go once and if you have photos with shadows and lights to fix they are limited in doing so. As for the quality of the jpeg photos, be careful not to confuse things. the P4rtk runs which is a beauty, maybe even the evo 2 rtk will be like this but from the software point of view it still has to eat a lot of pasta with sauce! I don't want to be controversial but give me a plausible justification or tell me that from an IT point of view it is not possible, because if it is, as I believe and also easily, it seems to me assured not to do so. a greeting from a user at the moment, given the countless troubles of the evo 2 rtk, he is really disappointed. Hello

 

[paolo]

Hi Paolo, I do not have an EVO yet. But, if you are intent on shooting .dng, and your workflow allows for the additional time needed for .dng image storage on the drone, you should be able to configure Exiftool to write all of the exif data from the original .jpg copies to new .jpg copies derived from edited .dngs.

One thing I'm not certain about, and maybe Ken can comment, is if there is any actual time differential between the "capture" of the .jpg vs the .dng. I suspect that any differential is in the processing and caching and not the capture. But I am not possitive.

thanks for the suggestion but I have already tried exif tool and unfortunately works a couple of files at a time .. I at each survey (buildings or urban complexes) minimum shot 1800 photos. Exif tool helped me to understand that the drone has two versions of "software" with which it stores data: one for gnss and one for gps. And in fact my amazement is given by the fact that the dng file has coordinates like the jpeg with the difference that it does not have the info on gimbal, pitch, roll yaw. Now if I work in rtk and I make the system memorize only the gnss data on both files the difference is second mè minimal and a 3 second interval as storage limit is sufficient. Also because with the electronic shutter it is not that you can go beyond certain speeds if you aim for precision. Now it is true that not all drone systems have this possibility and a fortiori it makes sense to do so. Why castrate a product when it has room for improvement ....?

 

[paolo]

Yep, exactly. I’m just worried about it hammering jpg + DNG that fast. If intervals are 3-4 sec I can see. But if asking 1.5-2sec, I’m not sure.

I would go out and test today but still getting the trailer ready for winter. I’m ready for a hotel vacation, a lot less work.

in attachment I put you the screen shot (on two monitors) of two point clouds made with the same set of photos by taking dng + jpeg. it is already evident from the screen that the difference between the two point clouds is immense, I would say incredibly too much, too much. both clouds were made with high alignment and the cloud generated with high quality and light filter. that's all .... maybe this way you understand why you can't leave your jpeg-generating software the freedom to modify and simplify the digital negative (raw). this happens in all cameras. You do it but it would be better to say we can't do it, even if I don't believe it!

 

[Dave Pitman]

thanks for the suggestion but I have already tried exif tool and unfortunately works a couple of files at a time ..

Hmmm.. With the help of the Exiftool developer, Phil Harvey, I have used a command that will convert batches of files (in folders) at a time.

Copying tags from original image to edited one.

Copying tags from original image to edited one.

exiftool.org

 

[paolo]

Hmmm.. With the help of the Exiftool developer, Phil Harvey, I have used a command that will convert batches of files (in folders) at a time.

Copying tags from original image to edited one.

Copying tags from original image to edited one.

exiftool.org

I tried but it has some problems in transferring the data between jpeg and dng. I will do other attempts and I will try other programs because from what I understand autel is not able to solve this problem!

 

[Dave Pitman]

I tried but it has some problems in transferring the data between jpeg and dng. I will do other attempts and I will try other programs because from what I understand autel is not able to solve this problem!

You won't be able to transfer the exif data from the jpgs to the dngs. What you want to do is make any exposure/contrast, etc. edits to your dngs that you need, then export as a jpg. Then you apply the exif from the original jpgs to your new jpgs. Is that what you have tried?

 

[paolo]

You won't be able to transfer the exif data from the jpgs to the dngs. What you want to do is make any exposure/contrast, etc. edits to your dngs that you need, then export as a jpg. Then you apply the exif from the original jpgs to your new jpgs. Is that what you have tried?

succeeded, with a different procedure, but I did

 

[paolo]

Yep, exactly. I’m just worried about it hammering jpg + DNG that fast. If intervals are 3-4 sec I can see. But if asking 1.5-2sec, I’m not sure.

I would go out and test today but still getting the trailer ready for winter. I’m ready for a hotel vacation, a lot less work.

by way of demonstration, after having copied the exif data of the jpeg files of the same set of 441 photos into the dng files (expeditious survey of an urban complex), I will give you the screen sot of the scattered cloud processing and, of course, the dng file containing more information in the alignment process ensures greater accuracy. I repeat the same files of the same photos with the same coordinates (if you want I can send you the data set). For absurdity, the dng is better, which can be easily converted into jpeg without special procedures and software, than to go the other way. In the end I solved in an artisan way with a more or less simple procedure to do. good job

 

[Dave Pitman]

succeeded, with a different procedure, but I did

In Metashape, you can export the camera locations (xyz) of the .jpgs to a file and then import the file to use with the .dng copies. I'm not sure if more than xyz metrics, like yaw, pitch, roll, are needed, or if they are included in the .dng exif.

 

[paolo]

thanks davide del'interessamento, I have seen that by processing the dng file the information is more and the better the product and specifically the point cloud is well structured without too much noise. Working in jpeg is a disgrace for me, then there are types of reliefs for which jpeg is sufficient. The dng requires a more powerful pc but in the end the product is of another level and the customer is happy! At the moment the autel product is the same as others but less mature, and the customer is angry! when I'm good to do the migration of exif on dng I'll let you know.

 

[Blade Strike]

by way of demonstration, after having copied the exif data of the jpeg files of the same set of 441 photos into the dng files (expeditious survey of an urban complex), I will give you the screen sot of the scattered cloud processing and, of course, the dng file containing more information in the alignment process ensures greater accuracy. I repeat the same files of the same photos with the same coordinates (if you want I can send you the data set). For absurdity, the dng is better, which can be easily converted into jpeg without special procedures and software, than to go the other way. In the end I solved in an artisan way with a more or less simple procedure to do. good job

That's actually pretty good but can you do like 3 checkpoints and no GCP? It looks like you only have one check point, could be wrong because I can read everything.

 

[jmason702]

What I would like to see is an accuracy report. Against some checkpoints, only then will we know that you actually have an absolute model. Now if you're only concerned about a pretty stitched photo then there's no need.

I have not been able to fly or do any testing in weeks, do to work. @Blade Strike what is your testing showing when using checkpoints, please?

 

[Blade Strike]

I have not been able to fly or do any testing in weeks, do to work. @Blade Strike what is your testing showing when using checkpoints, please?

All within spec of craft.

 

[jmason702]

How's this for pretty?

 

[jmason702]

All within spec of craft.

Thank you!