AT 209 Lab 8: Multispectral Indices

Introduction

Building off of Lab 7, this lab continued to increase knowledge of ArcGIS Pro. Looking more into what multispectral is, two different images were used to look at the different color bands used in creating the image. Multispectral's color bands contain useful information, and Lab 8 explores the different band combinations and they type of data that can be pulled off them.

Methods

Lab 8 consisted of using four different images of an area that was used for controlled burnings. The first two images were the original photos before and after burning, and the second pair of images were the same before and after but with thermal energy. For the lab, students were tasked with changing the different bands of color to create different false color imaging.

Discussion

While working on the lab, the images were taken from their original state into a true color pattern. Figure 1 compares the pre and post burn images in their original state. For the color bands, the formatting will follow the same pattern: Red band value, Green band value, Blue band value (Example Red 1, Green 2, Blue 3: 1,2,3).

Figure 1: The image shows the area in what was captured originally with the camera, which the band combination is 1,2,3.

For the sake of length, the remaining image color band changes use only the post burn image. Although the image is pretty realistic, the image's bands can be manipulated to create a true color filter. Figure 2 will show the post burn image in a true color way.

Figure 2: This image shows what the land actually looked like at the time, and the band configuration was 3,2,1.

Comparing Figure 1 to 2, it is pretty obvious that the original image compensated for some excess sunlight. Figure 1 would be used more for a more clear image. Figure 2 would be for more realism.

Transitioning to false coloring, the burnings would cause a high amount of IR radiation to come off of the areas. Figure 3 will show the image in IR.

Figure 3: The area is now presented in IR false coloring, which the band sequence is 5,3,2.

After being introduced with the different types of sequences that are commonly used, different types of sequences were tried. Figure 4 will present three different sequences that I used.

Figure 4: From Left to Right - The band sequences are 4,3,2; 5,3,5; 5,5,1.

After using the normal images, we switched to the data set that are based on thermal energy. While using those images, they were manipulated to a symbology image. Figure 5 will show the first image that was constructed. 

Figure 5: Above, the image is using a discrete color scheme, which means blocks of color rather than a spectrum/gradient.

For the next symbology, a more gradient style was used for the same image. Figure 6 pictures the spectrum used.

Figure 6: The spectrum that was used was a light blue to a purple.

Referring to all of the figures above, these images gave a big insight into multispectral imaging and what kind of data can be pulled off the many bands that are present.

Conclusion

Lab 8 was very insightful of multispectral imaging. For doing such task, satellites are used to gather the images. In terms of UAS, it has high applicability to be used. The systems themselves are already automated, but the use of drones for this purpose could pose for better quality data. Since the drones do not fly as high, they can be used to generate more accurate and higher resolution data. UAS can also be adopted to be used as satellites further down as the technology develops. In terms of the data, multispectral makes it easy to pull the most data out of one small set.