Landsat+Lab+9+Principal+Component+Analysis


 * Landsat Lab 9 Principal Component Analysis**

Quicklook of True color original subset of Landsat Path 197 Row 29:



Quicklook of all six principal components: The first image contains the majority of the variation. Bands 2 and 3 continue to show some variation, while bands 4, 5, and 6 increase in noise respectively.

Band 1.

Band 2.

Band 3.

Band 4.

Band 5.

Band 6.


 * Band Statistics:**


 * Basic Stats || Min || Max || Mean || Stdev || Num || Eigenvalue ||
 * Band 1 || -189 || 4264 || 260.453213 || 175.180168 || 1 || 1199624.672763 ||
 * Band 2 || -166 || 5421 || 598.763249 || 245.587324 || 2 || 662824.406542 ||
 * Band 3 || -128 || 6478 || 529.752533 || 309.282256 || 3 || 35747.352501 ||
 * Band 4 || -508 || 8367 || 4194.978886 || 1082.157193 || 4 || 7021.723855 ||
 * Band 5 || -96 || 6744 || 2194.029888 || 583.970789 || 5 || 2531.560541 ||
 * Band 6 || -47 || 8235 || 1129.607900 || 458.963283 || 6 || 1640.391715 ||


 * Covariance Matrix:**


 * Covariance || Band 1 || Band 2 || Band 3 || Band 4 || Band 5 || Band 6 ||
 * Band 1 || 30688.091117 || 40459.147284 || 51084.924465 || -9109.449615 || 77089.969982 || 68716.166043 ||
 * Band 2 || 40459.147284 || 60313.133529 || 71785.170177 || 30315.117396 || 119139.727611 || 97879.114618 ||
 * Band 3 || 51084.92.4465 || 71785.170177 || 95655.514079 || -37365.780970 || 143374.727520 || 128711.846593 ||
 * Band 4 || -9109.449615 || 30315.117396 || -37365.780970 || 1171064.190932 || 148742.537620 || -29623.989727 ||
 * Band 5 || 77089.969982 || 119139.727611 || 143374.727520 || 148742.537620 || 341021.882875 || 246229.925267 ||
 * Band 6 || 68716.166043 || 97879.114618 || 128711.846593 || -29623.989727 || 246229.925267 || 210647.295384 ||


 * Correlation Matrix:**

The correlation matrix indicates if there is a high degree of correlation between two components. Highly correlated bands yield image components similar in appearance. Bands 1, 2, 3, 5, and 6 show the highest correlation, and band 4 shows the least correlation.
 * Correlation || Band 1 || Band 2 || Band 3 || Band 4 || Band 5 || Band 6 ||
 * Band 1 || 1.000000 || 0.940429 || 0.942872 || -0.048053 || 0.753567 || 0.854665 ||
 * Band 2 || 0.940429 || 1.000000 || 0.945091 || 0.114068 || 0.830729 || 0.868373 ||
 * Band 3 || 0.942872 || 0.945091 || 1.000000 || -0.111642 || 0.793828 || 0.906746 ||
 * Band 4 || -0.048053 || 0.114068 || -0.111642 || 1.000000 || 0.235371 || -0.059645 ||
 * Band 5 || 0.753567 || 0.830729 || 0.793828 || 0.235371 || 1.000000 || 0.918696 ||
 * Band 6 || 0.854665 || 0.868373 || 0.906746 || -0.059645 || 0.918696 || 1.000000 ||


 * Eigenvectors and Eigenvalues**
 * Eigenvector || Band 1 || Band 2 || Band 3 || Band 4 || Band 5 || Band 6 ||
 * Band 1 || 0.007070 || 0.047010 || -0.003724 || 0.981819 || 0.182561 || 0.020703 ||
 * Band 2 || -0.188886 || -0.268356 || -0.353838 || 0.147328 || -0.660019 || -0.556574 ||
 * Band 3 || 0.380777 || 0.481407 || 0.561469 || 0.079103 || -0.547176 || -0.048476 ||
 * Band 4 || -0.088376 || -0.287654 || -0.097821 || 0.085140 || -0.475003 || 0.816700 ||
 * Band 5 || 0.678998 || 0.260222 || -0.681321 || -0.020199 || -0.007769 || 0.081110 ||
 * Band 6 || 0.591973 || -0.737267 || 0.292864 || 0.020281 || 0.077181 || -0.117765 ||

The first three bands contain 99.41% of the information; whereas the remaining bands contain noise.
 * ENVI ASCII || Plot File ||  ||
 * Column 1: || Eigenvalue Number ||  ||
 * Column 2: || Eigenval:1984Stan || Variability ||
 * 1 || 1199624.672763 || 62.83% ||
 * 2 || 662824.406542 || 34.71% ||
 * 3 || 35747.352501 || 1.87% ||
 * 4 || 7021.723855 || 0.37% ||
 * 5 || 2531.560541 || 0.13% ||
 * 6 || 1640.391715 || 0.09% ||
 * || 1909390.1079170 ||  ||


 * Inverse PC Rotation**

The first image is the true color original and the third is a zoomed in area of the original. The second is the Inverse PC Rotation image and the fourth image is a zoomed in image with Inverse PC Rotation. The colors red and yellow seem to have been lost in the Inverse PC Rotation and the overall image appears dull; however, the areas near the river seem to have more clarity, whereas the original image seems muddled in that area. For parcels of land or farm tracts, the original image seems to show these better possibly because of the colors.