cloudless
cloudless copied to clipboard
BradNeuberg
Application to multispectral imagery
@BradNeuberg I've been looking into your blog post regarding this project, and it seems really nice. I wonder if you guys have applied this to multispectral imagery, which could narrow down the classification errors.
We haven't applied this to multispectral imagery but I'm interested in that domain. Are you planning to try it?
On Thu, Jun 29, 2017 at 8:25 AM Ricardo Barros Lourenço < [email protected]> wrote:
@BradNeuberg https://github.com/bradneuberg I've been looking into your blog post regarding this project, and it seems really nice. I wonder if you guys have applied this to multispectral imagery, which could narrow down the classification errors.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/BradNeuberg/cloudless/issues/14, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJD0ZYPftN13OQuiRNF3gHa1GTm5m81ks5sI8HxgaJpZM4OJeEz .
I'm working on a cloud classification problem for my master's paper. Perhaps working on the top of your findings would be good. Have you published your findings in a journal?
I didn't get to publish my findings as I don't think they were quite strong enough.
On Thu, Jun 29, 2017 at 10:50 AM Ricardo Barros Lourenço < [email protected]> wrote:
I'm working on a cloud classification problem for my master's paper. Perhaps working on the top of your findings would be good. Did you publish these findings in a journal?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/BradNeuberg/cloudless/issues/14#issuecomment-312044542, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJD0Tb5yjTkwEP2PmR2AnZARK-8UIDrks5sI-PogaJpZM4OJeEz .
Ok. Anyway is interesting even on the subject of image segmentation. I'm willing to reproduce your experiment, with the models you've provided. My research is regarding cloud classification, and the segmentation would be an initial part. I assume that on your models, you are working with a png rib visible composite, right? Because my assets come from MODIS, and probably on GeoTIFF, so I need to map the scheme you've used.