Every image taken with a P1 camera contains the usual EXIF information relating to the exposure, GPS coordinates, etc. These values can be extracted and read by most traditional EXIF programs, including Windows if you right click on the image and select "Details".

However, with every image file stored by the P1 camera there is an "XMP" segment internal to the EXIF section that records much more of the details about not only the image itself but also the gimbal and the UAV at that inatant.

Here is an example of some of the information recorded in the XMP section of a single JPG image file...

   tiff:Make="DJI"
   tiff:Model="ZenmuseP1"
   dc:format="image/jpg"
   drone-dji:Version="1.4"
   drone-dji:GpsStatus="Normal"
   drone-dji:AltitudeType="RtkAlt"
   drone-dji:GpsLatitude="+29.282190206"
   drone-dji:GpsLongitude="-94.922448090"
   drone-dji:AbsoluteAltitude="+14.687"
   drone-dji:RelativeAltitude="+39.839"
   drone-dji:GimbalRollDegree="+0.00"
   drone-dji:GimbalYawDegree="+57.20"
   drone-dji:GimbalPitchDegree="-89.90"
   drone-dji:FlightRollDegree="+5.00"
   drone-dji:FlightYawDegree="+55.90"
   drone-dji:FlightPitchDegree="-7.40"
   drone-dji:FlightXSpeed="5.1"
   drone-dji:FlightYSpeed="7.4"
   drone-dji:FlightZSpeed="0.0"
   drone-dji:CamReverse="0"
   drone-dji:GimbalReverse="0"
   drone-dji:SelfData=""
   drone-dji:RtkFlag="16"
   drone-dji:RtkStdLon="0.92389"
   drone-dji:RtkStdLat="1.08331"
   drone-dji:RtkStdHgt="2.54741"
   drone-dji:RtkDiffAge="0.00000"
   drone-dji:SurveyingMode="1"
   drone-dji:UTCAtExposure="2021-05-26T15:21:49.073960"
   drone-dji:ShutterType="Mechanical"
   drone-dji:ShutterCount="23122"
   drone-djironeModel="Matrice 300 RTK


I'm still learning about it all, but as far as I can tell, the "Relative Altitude" is GPS based, and doesn't take into account any info from the downward facing sensors to track trees or hills. The platform speed at the instant of the image can be calculated using the X and Y speeds being East/West and North/South respectively.

I have tried several commercial EXIF programs, but so far only one showed what was in the XMP section, named "EXIFTOOL by Phil Harvey". It actually works from the DOS command line, and as fas as I can tell, only on a single image file at a at time, but it was enough to show that there is WAY more information in each JPG than can be read by normal EXIF programs.

After figuring out how to open the XMP info I was able to write several Windows programs that read and export the information.

By reconstructing the EXIF/XMP data it is possible to recreate a log file at 1 second intervals, export to GIS software, plot mission paths and boundaries,, and add the direction the camera was facing to photogrammetry programs like Agisoft Metashape.

I've been trying to write something that correlates pixel locations in one image with the same location in a surrounding image by comverting to Lat/Lon. So far I've come upon some weird offset of around 1 meter always in the same direction. If it was related to the direction the UAV was facing I would think it had something to do with the offset of the camera sensor to the GPS location in the EXIF, which I can calculate. Once I figure that out I'm expecting to need a correction for pixel distance from the center of the image, becaue pixels at the corners should be farther away from the camera sensor than the center pixel by the viewing angle.

If anyone is interested in seeing what non-standard EXIF info is recorded for other cameras, send me a raw JPG to HMArnold (at) MSN.com

Every image taken with a P1 camera contains the usual EXIF information relating to the exposure, GPS coordinates, etc. These values can be extracted and read by most traditional EXIF programs, including Windows if you right click on the image and select "Details".

However, with every image file stored by the P1 camera there is an "XMP" segment internal to the EXIF section that records much more of the details about not only the image itself but also the gimbal and the UAV at that inatant.

Here is an example of some of the information recorded in the XMP section of a single JPG image file...

   tiff:Make="DJI"
   tiff:Model="ZenmuseP1"
   dc:format="image/jpg"
   drone-dji:Version="1.4"
   drone-dji:GpsStatus="Normal"
   drone-dji:AltitudeType="RtkAlt"
   drone-dji:GpsLatitude="+29.282190206"
   drone-dji:GpsLongitude="-94.922448090"
   drone-dji:AbsoluteAltitude="+14.687"
   drone-dji:RelativeAltitude="+39.839"
   drone-dji:GimbalRollDegree="+0.00"
   drone-dji:GimbalYawDegree="+57.20"
   drone-dji:GimbalPitchDegree="-89.90"
   drone-dji:FlightRollDegree="+5.00"
   drone-dji:FlightYawDegree="+55.90"
   drone-dji:FlightPitchDegree="-7.40"
   drone-dji:FlightXSpeed="5.1"
   drone-dji:FlightYSpeed="7.4"
   drone-dji:FlightZSpeed="0.0"
   drone-dji:CamReverse="0"
   drone-dji:GimbalReverse="0"
   drone-dji:SelfData=""
   drone-dji:RtkFlag="16"
   drone-dji:RtkStdLon="0.92389"
   drone-dji:RtkStdLat="1.08331"
   drone-dji:RtkStdHgt="2.54741"
   drone-dji:RtkDiffAge="0.00000"
   drone-dji:SurveyingMode="1"
   drone-dji:UTCAtExposure="2021-05-26T15:21:49.073960"
   drone-dji:ShutterType="Mechanical"
   drone-dji:ShutterCount="23122"
   drone-djironeModel="Matrice 300 RTK


I'm still learning about it all, but as far as I can tell, the "Relative Altitude" is GPS based, and doesn't take into account any info from the downward facing sensors to track trees or hills. The platform speed at the instant of the image can be calculated using the X and Y speeds being East/West and North/South respectively.

I have tried several commercial EXIF programs, but so far only one showed what was in the XMP section, named "EXIFTOOL by Phil Harvey". It actually works from the DOS command line, and as fas as I can tell, only on a single image file at a at time, but it was enough to show that there is WAY more information in each JPG than can be read by normal EXIF programs.

After figuring out how to open the XMP info I was able to write several Windows programs that read and export the information.

By reconstructing the EXIF/XMP data it is possible to recreate a log file at 1 second intervals, export to GIS software, plot mission paths and boundaries,, and add the direction the camera was facing to photogrammetry programs like Agisoft Metashape.

I've been trying to write something that correlates pixel locations in one image with the same location in a surrounding image by comverting to Lat/Lon. So far I've come upon some weird offset of around 1 meter always in the same direction. If it was related to the direction the UAV was facing I would think it had something to do with the offset of the camera sensor to the GPS location in the EXIF, which I can calculate. Once I figure that out I'm expecting to need a correction for pixel distance from the center of the image, becaue pixels at the corners should be farther away from the camera sensor than the center pixel by the viewing angle.

If anyone is interested in seeing what non-standard EXIF info is recorded for other cameras, send me a raw JPG to HMArnold (at) MSN.com