Adobe Lightoom Photo Merge

Overview
Adobe Lightroom is an image organization software that allows you to stitch, edit, and manipulate photos. In this post, you will learn how I use Lightroom to merge photos of large structures like bridges.

The Need for Merged Photos 
Due to restrictions on altitude when capturing images with aircraft like the DJI M210, I often need to fly far away and capture oblique photos rather than nadir photos. Oblique photos can be stitched together to create a more detailed image, hence the need for merged photos.

Alternatives to Merging Photos 
While photogrammetry platforms like Pix4D can be used to create an Orthomosaic, I found Adobe Lightroom to be a more straightforward and efficient alternative. Lightroom is particularly useful when dealing with projects that include water.

Why Adobe Lightroom?
I found adobe lightroom to be fairly easy to merge photos compared to Pix4D. Many projects I fly include water and as of this post, Lightroom has done well with fewer distractions in my experience. 

A Case Study with the DJI Zenmuse XT2
The XT2 camera captures both visual and thermal images and is compatible with the DJI Matrice 200 and 300 series. To scan infrastructure like bridge decks, I take a series of thermal photos and visual photos. The visual photos are used to map the structure and serve as comparison photos to the thermal images.

After importing the 47 visual photos captured by the XT2 into Lightroom (see Figure 1), I select all the photos and click the Perspective Merge function (See Figure 2). The photos take about a minute to process. Refereencing Figure's 2 and 4, the Boundary Wrap function helps correct the fisheye distortion common with the XT2 camera, and any remaining lens distortion can be corrected using open-source software or Adobe Photoshop.

Figure 1: 47 Photos Loaded to Adobe Lightroom

Figure 2: Perspective Merge Function

Figure 3: Merge Before Applying Boundary Wrap

Figure 4: Merge After Applying Boundary Weap

Results of Stitching Images From the XT2 

Referencing Figure 5,the results of stitching images from the XT2 show a bridge with one truck photo bombing it. Although the image quality decreases when zoomed in, this is due to the XT2's 12-megapixel camera, which is not designed for cinematography. The barrier wall on the right side is not perfect and could be improved in Photoshop, while the left barrier wall appears even.

Figure 5: Distortion Correction Settings in Adobe Photoshop

Figure 6: Merged Imagery From Zenmuse XT2

Conclusion 

Considering that we capture numerous thermal images at low elevations, the visual reference photos can further enhance the overall assessment by facilitating structural mapping. For sizable structures, this approach proves particularly advantageous, as it allows for preliminary reference mapping using the visual imagery without exceeding the 400 ft limit or needing to unlock the M210.