How to Process MicaSense Sensor Data in Pix4D

Introduction

This guide is intended to outline a standard image processing workflow for use with Pix4Dmapper. Pix4D is compatible with RedEdge (3, M, MX models), Altum, and the RedEdge-MX Dual Camera system. 

Video

Watch on YouTube

Please note that this video was published in 2017. There have been several software updates to both Pix4D and MicaSense cameras since then. The video remains quite accurate, but we recommend reading this guide in addition to the video for the most updated information.

Importing the Data

Imagery taken with MicaSense sensors will be automatically recognized since the camera is in the Pix4Dmapper camera database. 

NOTE: At this time, Altum data is only supported on the Windows version of Pix4D. RedEdge is supported by both Windows and Mac versions. 

  • Create a new project
  • Import all of the images for each band. We recommend clicking "Add Directories" instead of "Add Images" to ensure all files are added simultaneously. This is very important!
  • The order in which the bands should appear are as follows:
  • In the Edit Camera Model window, ensure the Global Shutter model is selected
  • Choose the Ag Multispectral processing template

Radiometric Processing and Calibration

On the Menu bar, click Process > Processing Options...

processing_options.png

the Processing Options pop-up appears. Press 3. DSM, Orthomosaic and Index

Select the Index Calculator tab:

calibration2.PNG

Under Radiometric Processing and Calibration, you can choose the following options, some of which may not be available for MicaSense Sensor data. 

  • No correction: no radiometric correction will be done
  • Camera only: corrections will be applied for the parameters that are written in the EXIF metadata and relate to the camera (vignetting, dark current, ISO, etc...). This will still use the calibration panel images.
  • Camera and Sun Irradiance:  corrections will be applied for the camera parameters from the point above as well as for the sun irradiance information written in the XMP.Camera.Irradiance EXIF tag. Use this option to include data from your DLS.
  • Camera, Sun Irradiance and Sun angle: corrections will be applied to take into account the sun position, as well as the camera information and the irradiance data. This option should only be chosen for flights that were done in variable sky conditions. 

See this article for more details: Using Panels and/or DLS in Post-Processing

Important: Since the thermal imager in Altum is already radiometrically calibrated, you should choose the correction type "No Correction" for the LWIR band.

The camera used is displayed. Users can calibrate the sensor to perform an illumination adjustment in order to obtain more accurate reflectance values. If there are more than one camera models in the project, all the cameras will be listed in the Radiometric Processing and Calibration section.

We recommend to also navigate to DSM and Orthomosaic, and check "Raster DSM" (merged tiles) and "Orthomosaic" (merged tiles) so that you will have additional outputs to use. These are optional, but can be very useful for additional analysis.

 

Using the Calibration Panel

The Calibrate button allows users to take into account the information from a picture with radiometric calibration target if such a target was used during the project.

Obtaining your panel's reflectance values

If you have a MicaSense Reflectance Panel, you can get the reflectance values for your panel by filling out this form.

This procedure will need to be done for each of the five bands, with the corresponding "albedo" value provided for each band. Please note the band order (Blue, Green, Red, NIR, Red Edge) and ensure you are entering the appropriate value for each band.

After clicking Calibrate for a camera model, the Radiometric Calibration pop-up appears.


drawpolygon.png

 

If a radiometric calibration target is used, make sure that the radiometric target data is taken into account by ensuring the symbol in the Processing Options is a green check mark.

In File Name, the Browse button, opens the Select a radiometric calibration image pop-up. This pop-up allows the user to select the image in which the radiometric calibration target appears.

When an image is browsed, the user can draw a region on the image that will define the radiometric calibration. The Reset button, resets the area drawn by the user. 

Use the left mouse button to draw the first three points, then draw the final point with the right mouse button.

 

Important:  The albedo values range between 0.0 and 1.0.

Processing

After configuring Pix4D to use the calibration data, you can continue to process your data set. You will need to generate a reflectance map. For up-to-date details on processing, please see the Pix4D knowledge base.

Once you have generated a reflectance map, see How to export to Atlas from Pix4D* to export the final output to Atlas for analytics.

NOTE: Altum data is not currently compatible with Atlas. 

Merging the Reflectance Map

The reflectance map generates a GeoTIFF for each band. To combine these into one multi-band image, you can either upload the data to Atlas and download the GeoTIFF, or use external software:

  • GDAL's gdalbuildvrt via the command line
  • QGIS Virtual Raster Builder

Using gdalbuildvrt

This GDAL program builds a VRT (Virtual Dataset) that is a mosaic of the list of input GDAL datasets. Using a command line, navigate to your Pix4D output folder containing the reflectance map tiles. (...\project_name\4_index\reflectance\tiles)

 

For RedEdge-M/MX (5 bands): you will build 5 .vrt files before merging them.

gdalbuildvrt blue.vrt *blue_*.tif
gdalbuildvrt green.vrt *green_*.tif
gdalbuildvrt red.vrt *_red_*.tif
gdalbuildvrt rededge.vrt *"red edge"_*.tif
gdalbuildvrt nir.vrt *nir_*.tif

Then merge the 5 .vrt files with -separate

gdalbuildvrt -separate reflectance.vrt blue.vrt green.vrt red.vrt rededge.vrt nir.vrt

 

For Altum (6 bands): you will have to build an additional .vrt file for the LWIR (thermal) band before merging. Follow the instructions for the 5 bands above, then make a final .vrt file by using:

gdalbuildvrt lwir.vrt *_lwir_*.tif

 Then merge the 6 .vrt files with -separate

gdalbuildvrt -separate reflectance.vrt blue.vrt green.vrt red.vrt rededge.vrt nir.vrt lwir.vrt

 

For RedEdge-MX Dual Cam system (10 bands), you will build a total of 10 .vrt files before merging them. 

gdalbuildvrt blue444.vrt *blue-444*.tif
gdalbuildvrt blue.vrt *blue_*.tif
gdalbuildvrt green531.vrt *green-531*.tif
gdalbuildvrt green.vrt *green_*.tif
gdalbuildvrt red650.vrt *red-650*.tif
gdalbuildvrt red.vrt *_red_*.tif
gdalbuildvrt rededge705.vrt *"red edge-705"*.tif
gdalbuildvrt rededge.vrt *"red edge"_*.tif
gdalbuildvrt rededge740.vrt *"red edge-740"*.tif
gdalbuildvrt nir.vrt *nir_*.tif

 Then merge the 10 .vrt files with -separate 

gdalbuildvrt -separate reflectance.vrt blue444.vrt blue.vrt green531.vrt green.vrt red650.vrt red.vrt rededge705.vrt rededge.vrt rededge740.vrt nir.vrt  

 

Once you've merged all .vrt files to create a multi-layered .vrt (if you used our code it will be reflectance.vrt), you can now open this in any GIS software (such as QGIS) for analysis.

Using QGIS Virtual Raster Builder

Note: This method does not create a new raster. It is a way to organize your existing rasters into one catalog for easy access and visualization. 

Load the GeoTIFFs for each band into the Layers Panel (drag and drop or Layer > Add layer > Add raster layer...)

  • Ensure that the individual band geoTIFFs are in the correct order by dragging and dropping
    • Correct order: Blue, Green, Red, Red edge, NIR
    • For 10-band data sets, the correct order is Blue-444, Blue, Green-531, Green, Red-650, Red, Red edge-705, Red edge, Red edge-740, NIR
  • Select (turn on) all geoTIFFs

BGREN.JPG

Open Virtual Raster Builder from the QGIS menu bar: Raster > Miscellaneous > Build Virtual Raster (Catalog)...

BVR_tutorial_1.jpg

 

The Virtual Raster Builder dialog opens (see below).

  • Since all geoTIFFs are activated in QGIS, select the option to use visible raster layers for input.
  • Select an output file location for the virtual raster file.
  • In order to create a virtual raster with all discrete bands in one file, select the "Separate" option.
  • The output file will automatically load into QGIS unless this box is deselected.
  • Press OK to build the virtual raster.

VRB_tutorial.JPG

The new virtual raster is added to the Layers Panel in QGIS. Now we can move on to editing the layer properties to display properly.

From the Layers Panel, right click on the new virtual raster layer and select Properties. Within the Layer Properties, navigate to the “Style” tab. Ensure you are in the “Multiband color” option. Using the drop down menu for each band, you will see that this layer now contains 5 bands. Change the Red band to “Band 3”, Green band to “Band 2”, and Blue band to “Band 1”. 

For 10-band data (from the Dual Cam system), you can instead change the Red band to "Band 6", Green band to "Band 4", and Blue band to "Band 1". Follow the rest of the steps as you normally would.

Bands_style__2_.jpg

Now, copy the smallest "Min" value and paste it into the other "Min" value boxes. Then copy the largest "Max" value and paste it into the other "Max" value boxes. The above example will then look like the example below. This step is not necessary but stretching these values generally helps to create a more visually pleasing image for the human eye. 

Layer_props2.JPG

 

 

Have more questions? Submit a request