Comparing Spectral Bands for LANDSAT 8 LANDSAT 7
Comparing Spectral Bands for LANDSAT 8, LANDSAT 7, LANDSAT 5, and LANDSAT 4 Using Google Earth Engine CEE 6003 Remote Sensing of Land Surfaces Abigail Repko, M. E. Utah State University I. Project Introduction LANDSAT satellites collect multispectral data at a temporal resolution of 16 days. The collected data can be used to create false-color images, which can quickly identify areas of plant growth or urbanization over a period of time. The same data is collected by every LANDSAT satellite, but the spectral bands vary. This analysis uses near-infrared (NIR), red, and green bands to generate false-color images from LANDSAT 8, LANDSAT 7, LANDSAT 5, and LANDSAT 4 data. Data was available for Santo Tomas, Baja California, for all four LANDSAT satellites of interest. The site location is pictured in Figure 1. Google Earth Engine was used to process LANDSAT data and generate false-color images at the analysis location for all four LANDSAT satellites. IV. Discussion II. Data Processing Google Earth Engine was used to process data for four LANDSAT satellites. To prepare the false-color images, the following steps were performed: i. Import LANDSAT product (i. e. LANDSAT 8) into Google Earth Engine and input correct bands for NIR, red, and green ii. Create an image using the selected location as a filter point, cloud cover as a metadata filter, and range for each LANDSAT product. Thedate different LANDSAT products used and iii. Stretch the image to a visible range spectral bands for NIR, red, and green for each product can be seen in Table 1. Comparison of different LANDSAT products used. LANDSAT product Spectral Bands LANDSAT 8 B 5: NIR B 4: Red B 3: Greed LANDSAT 7 B 4: NIR B 3: Red B 2: Green LANDSAT 5 B 4: NIR B 3: Red B 2: Green LANDSAT 4 B 4: NIR B 3: Red B 2: Green III. Results False-color images for each LANDSAT product can be seen in Figure 2 through Figure 5. V. Conclusion False-color images produced by different LANDSAT products were similar. The top right corner of each image shows the expansion and urbanization of Mexicali. LANDSAT 8 has the most bands, and the image produced is more detailed than the false-color images produced by other LANDSAT products. The false-color images produced by LANDSAT 4 and LANDSAT 5 were very similar, . Urbanization can be seen in the top right, while coastal vegetation is observed on the left side of the images. LANDSAT 7 produced a similar image, however there is noticeably less vegetation in this image. LANDSAT 7 collected data from January 1999 to present. The false-color image was generated using data from the entire lifespan of LANDSAT 7. Climate change and drought could be responsible for the decrease in vegetation observed in Figure 4. Figure 5 shows the highest growth of urban areas and an increase in coastal vegetation. LANDSAT 8 was launched in February 2018. The data shows a more accurate, fully integrated image of Santo Tomas, which illustrates how much internal instrumentation has improved between the different LANDSAT satellites. Overall, the spectral resolution of LANDSAT 8 is much more specific and results in a clearer, more distinct image that can be used to identify sensitive parameters of a specific area. LANDSAT satellites provide multispectral data across the globe. LANDSAT 8 has the most distinct spectral resolution and, as a result, can produce the most accurate false-color images out of the four LANDSAT products analyzed. The next step in this analysis would be to use top of atmosphere (TOA) data to create a more standardized image, which could further identify differences between LANDSAT products. After quantifying the differences between LANDSAT 8, LANDSAT 7, LANDSAT 5, and LANDSAT 4, a linear regression could be created. The resulting regression would allow for earlier LANDSAT products to be adjusted to “match” LANDSAT 8 values. LANDSAT 9 will launch in 2020. Data from LANDSAT 9 can be incorporated into this study. Ultimately, previous LANDSAT products can be adjusted to match current and future LANDSAT Torres-Rua, A. (2019). CEE 6003 class lectures. products. VI. References Torres-Rua, A. (2019). CEE 6003 lab exercises. Gorelick, N. , Hancher, M. , Dixon, M. , Ilyushchenko, S. , Thau, D. , & Moore, R. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment. VII. Earth Engine Code https: //code. earthengine. google. com/875 aab 18077761 d 71 c 6 c 43 0 ee 5847 af 5 Figure 1. Analysis site location at Santo Tomas, Baja California. Images generated for Santo Tomas show the difference in instrument technology between the different LANDSAT satellites. The 16 -day temporal resolution is the same for all four satellites. Spectral resolution is also consistent, with all four satellites having a resolution of 30 meters. Figure 2: LANDSAT 4 false-color image of Santo Tomas. Figure 3: LANDSAT 5 false-color image of Santo Tomas. Figure 4: LANDSAT 7 false-color image of Santo Tomas. Figure 5: LANDSAT 8 false-color image of Santo Tomas.
- Slides: 1