Announcing Mavis

We are happy to announce Mavis - a new UAV photogrammetry and mapping software.

Mavis is specifically designed for agricultural, archeological and mining surveys.

Mavis covers the entire workflow from single UAV images to GIS ready maps. It comprises precise georeferencing, on-site image analysis and stitching, image preprocessing, the generation of digital surface models and orthophotos, color post-processing, the computation of vegetation indices, as well as the generation HTML reports including WebGL 3D models.

Mavis is designed with an intuitive workflow in mind - from georeferencing to 3D mapping and the generation reports - while providing state-of-the-art algorithms.

Detailed information is available at mavis.bitmapping.de

Preprocessing

Precise georeferencing is realized by a combination of Pixhawk/APM dataflash log file analysis and a specific CHDK script for Canon cameras. This combination allows accounting for shutter delays with high precision.
For other cameras the camera trigger information recorded in the log file can be used for georeferencing. Images which have already been geotagged can be used as well.

3D Mapping

Orthomosaics and Digital Surface Models form UAV imagery are generated using MicMac. MicMac uses a multi-scale, multi-resolution approach for dense matching and sophisticated camera calibration models.

Color Manipulation

Orthomosaics often require further processing for scientific analysis or interpretation purposes. Therefore, Mavis provides various tools to enhance and manipulate colors such as white balance, gamma correction and decorrelation stretching amongst others. Spatial filters to remove noise are implemented as well.

Vegetation Indices

Various vegetation indices can be derived using Mavis.
For single camera setups the filter type can be selected. Blue and red filters are supported.

Reports

HTML reports can be generated following 3D Mapping. A 3D WebGL model is generated and linked to the report.

Availability

Mavis is distributed free of charge with our UAVs.
It is currently in the final beta testing stage for regular distribution.
Please contact us for availability.

Mavis is made to evolve! Please share your thoughts, requirements and suggestions regarding Mavis and UAV mapping workflows in general.

Best regards,

Thorsten

Hi Chris, Mario, and RPM,

In general the resulting positional accuracy is mainly based on three components:

GPS accuracy, geo-tagging, image processing. (Some others are environmental conditions and camera mount)

With Mavis, we provide most probably the best geo-tagging possible for Canon CHDK cameras. Regarding the other two components we will soon run a comparison against an L1 RTK setup and survey grade GCPs. We’ll see if we can differentiate between the influence of GPS accuracy and image processing. I’ll post it.

As @RPM mentioned be happy to run some of your datasets. In individual cases I'd also be happy to provide a full version of Mavis for any (third party) test against L1 RTK setups and survey grade GCPs. Send me a PM so we can discuss things.

I agree with Mario that some public dataset would be prefect!

Best regards,
Thorsten

You need to be a member of diydrones to add comments!

Join diydrones

Comments

T3

Thorsten May 19, 2016 at 1:52pm

Hi all,

we now have a trial version available. You can request it at: http://mavis.bitmapping.de/request-trial/

Best regards,

Thorsten
T3

Thorsten February 20, 2016 at 9:46am

Hi Antonio,

we are always interested in research cooperations! Can you send me a PM?

Best regards,

Thorsten
Antonio Vigoni February 20, 2016 at 9:40am

I work with the university of Milano we use Photoscan and pix4d and I would like try your software. Thanks
T3

Thorsten February 20, 2016 at 9:36am

Hi Michael,

on the way to put my kids to bed :-)

A trial version is definitely on our list.

Until then, I would be happy to help and run your data through Mavis.

I'll send you a PM tomorrow!

Cheers,

Thorsten
Michael Papenhagen February 20, 2016 at 9:24am

Thorsten......where are you??????
Michael Papenhagen February 20, 2016 at 7:06am

Hi Thorsten

I know it is a weekend, but any progress on the possibility of having a Trail version? We are flying 2 Quarries and one farm next week and would really like to have a copy of your Software to evaluate.

Regards

Michael
Michael Papenhagen February 18, 2016 at 9:49am

Hi Thorsten

We would really like to have a trail version of your program to evaluate it against our completed Projects where we used GCP and also to compare it to Photoscan and Pix4D. At that price we will buy a couple of programs!

Regards

Michael
Pascal P. February 18, 2016 at 2:28am

Hi Thorsten,

Best way to promote your software is to quickly give access to a trial version, I know you are considering it, just give people the opportunity to try with their own data.

Testing CGP vs RTK accuracy is another story.Even if I think we all look to this dream : no more CGPs.
T3

Thorsten February 18, 2016 at 12:21am

Hi Chris, Mario, and RPM,

In general the resulting positional accuracy is mainly based on three components:

GPS accuracy, geo-tagging, image processing. (Some others are environmental conditions and camera mount)

With Mavis, we provide most probably the best geo-tagging possible for Canon CHDK cameras. Regarding the other two components we will soon run a comparison against an L1 RTK setup and survey grade GCPs. We’ll see if we can differentiate between the influence of GPS accuracy and image processing. I’ll post it.

As @RPM mentioned be happy to run some of your datasets. In individual cases I'd also be happy to provide a full version of Mavis for any (third party) test against L1 RTK setups and survey grade GCPs. Send me a PM so we can discuss things.

I agree with Mario that some public dataset would be prefect!

Best regards,
Thorsten
RPM February 17, 2016 at 8:21pm

Thanks @Chris for elaborating on what I was getting at. What someone like myself would like to see is several data sets showing a test site using survey grade GCPs, the usual APM Ublox GPS set up and another using RTK. Half the GCPs would be used for the creation of the DEM, the other half used as accuracy checks (for example). Of course the results would be better with the more advanced systems. The performance of other software is tested and known. @Thorsten has offered to process my data through email but I think these things should be done third party. Perhaps since it uses Micmac this data is already known? The price is certainly right but we need to know for sure it is a competitor with industry standard software. Happy to help!

of 3

This reply was deleted.