Does anyone know how to perform a orthogonal transformation with the planetscope images, to perform a Tesselles cap transformation?
Best answer by idilgumus
View originalDoes anyone know how to perform a orthogonal transformation with the planetscope images, to perform a Tesselles cap transformation?
Best answer by idilgumus
View originalHi Andres!
Based on our investigation, here is the answer: “
PC1 = coastal_blue_coeff * Coastal_blue + blue_coeff* Blue + green_i_coeff * Green_I + green_ii_coeff * Green_ii + yellow_coeff * Yellow + red_coeff * Red + sentinel_red_edge_i_coeff * RedEdge + nir_coeff * NIR
PC2 = ....
where *_coeff are the coefficients (Eigenvectors) of the PCA (e.g. in the csv file) and Coastal_blue, Blue, Green_i, ..., NIR are the pixels/bands of the image that is being transformed
Result is an orthogonal transformed image / an image that was transformed to the feature space of the fixed PCA.
The screenshots show how the pixels of images (here the images that were used to derive the PCA Eigenvectors) fall into the new, orthogonal, fixed PCA feature space.
The "cumulative_explained_variance" graph shows how much variance (information) was contained in the Principal Components (that were derived from the Palouse and Koethen areas).”
Let us know if you have further questions. Have a great day!
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.