Capturing multispectral data in ideal and adverse lighting conditions

While consistent lighting conditions are the best for ensuring radiometric accuracy, there will be times when data collection is necessary under sub-optimal lighting conditions. In this article, we will begin by defining what ideal and adverse lighting conditions are, then provide details on how the DLS and Calibrated Reflectance Panel (CRP) are used to mitigate the impacts of poor lighting conditions.

Optimal lighting conditions for data capture

The ideal conditions for multispectral data capture are a clear sunny day or an overcast day with uniform cloud coverage. In both cases, the lighting is consistent.

The most difficult lighting conditions to work with are partially cloudy days with sharp transitions from shadow to sun. These kinds of conditions will most likely produce anomalies in the multispectral data. These anomalies negatively affect the composites and vegetation indices generated from the dataset, rendering the data sub-optimal in many cases.

When ideal lighting conditions are not available

When it comes to radiometric calibration, a strong value point for MicaSense series sensors is that they offer options for adaptability in a variety of light conditions. Included in each camera kit are two calibration tools: the Calibrated Reflectance Panel (CRP) and the DLS.

An image of the CRP is taken before and after each flight to capture a baseline reflectance value, whereas the DLS collects data continuously during the flight to capture variations in light intensity. Specifically, the DLS captures downwelling light through 10 different light sensors positioned on different planes to best determine the position of the sun and the effects of light on the images collected. The DLS gathers this information from the 10 sensors and then applies an appropriate value to the metadata of each image collected by the MicaSense series sensor throughout the mission. This gives the DLS a distinct advantage over other light sensors on the market which use only one light sensor on one plane.

Using both the CRP and DLS

By using both the CRP and the DLS when collecting data, you have the ability during processing to experiment and use data from either tool or both tools to create the best possible orthomosaic. 

For example, when processing in Agisoft Metashape, Pix4D Mapper, or Simactive Correlator3d, you have the freedom to decide whether to calibrate using the CRP, the DLS, or both.

See an example of choosing the calibration method in the Metashape workflow below:

A good starting point on whether to use CRP and DLS data is to consider the lighting conditions in which the data was collected. We generally follow this guideline when processing data in-house:

• Completely clear sunny day: use CRP
• Overcast & other cloudy conditions : use CRP + DLS 

To illustrate the differences in using the CRP, DLS, or a combination of the two tools, we have provided the examples below, each captured under different lighting conditions.

Cumulus clouds // Both DLS and CRP used for calibration

Below, in the image on the left is an example showing the data impacts of cloudy conditions where illumination is highly variable.  Fortunately, in this instance data from the DLS could be used to correct for shadow-related effects during processing, correction shown on the right.

As evidenced in the images above, the cloud effects aren’t entirely removed but the map quality is greatly improved and perhaps the dataset has been converted from unusable to usable. This is a best-case outcome. It should be noted that sometimes variable illumination is so severe that even the use of DLS data does not remove the anomalies from the clouds, however, having the DLS data offers the best possibility for improvement.

No clouds, sunny // CRP used for calibration

When skies are clear, using only the CRP for calibration is the best decision in most instances.  Using the DLS on clear sunny days can sometimes lead to striping in the data on long straight surveys, as shown below:

Cirrocumulus clouds // Both DLS and CRP used for calibration

Another instance where the DLS can improve data quality is in variable overcast conditions, similar to what is shown in the image below.  In the image, it is apparent that there are bands of heavier clouds mixed in among clouds, thin enough to allow for sunbreaks.

The DLS’s ability to provide accurate radiometric corrections in variable lighting conditions can be seen in the example below. The DLS is proven to be effective in removing the impact of the localized sunbreak, allowing for a high quality and uniform orthomosaic to be generated.

Conclusion

Although you may be faced with challenging lighting conditions while collecting multispectral data, if you are using a MicaSense series camera paired with the CRP and DLS you will have the ability to generate high-quality, radiometrically calibrated outputs across a range of conditions. This flexibility in calibration options ensures that you can have consistent and dependable data for analysis and temporal studies.