danfenghong / IEEE_TGRS_MDL-RS

Danfeng Hong, Lianru Gao, Naoto Yokoya, Jing Yao, Jocelyn Chanussot, Qian Du, Bing Zhang. More Diverse Means Better: Multimodal Deep Learning Meets Remote Sensing Imagery Classification, IEEE TGRS, 2021, 59(5): 4340-4354.
101 stars 21 forks source link

Data preprocessing #8

Open G-z-w opened 1 year ago

G-z-w commented 1 year ago

您好,请问您是怎么将原SAR和MSI数据进行归一化操作的呢?原图数据范围层次不起,分布不均,但是我看您处理后的数据都在[0,1]范围内,可以告知具体的操作方式吗?谢谢 以下是某些SAR数据某些通道的数据分布: image image 以下是某些MSI数据某些通道的数据分布: image image 非常期待您的回复。感谢!

danfenghong commented 1 year ago

我们这个工作都是基于每个数据每个波段进行的0-1的归一化

G-z-w @.***> 于2023年6月3日周六 15:15写道:

您好,请问您是怎么将原SAR和MSI数据进行归一化操作的呢?原图数据范围层次不起,分布不均,但是我看您处理后的数据都在[0,1]范围内,可以告知具体的操作方式吗?谢谢 以下是某些SAR数据某些通道的数据分布: [image: image] https://user-images.githubusercontent.com/95175307/243064856-95ed2f1a-e7d3-4a15-af50-02d7cbcaaef9.png [image: image] https://user-images.githubusercontent.com/95175307/243064868-74d0709b-7bb8-4951-b6b8-5b91f1ee0874.png 以下是某些MSI数据某些通道的数据分布: [image: image] https://user-images.githubusercontent.com/95175307/243064907-ef76226d-f455-4c36-9bd8-dc040a8b0944.png [image: image] https://user-images.githubusercontent.com/95175307/243064916-f9b3bc4d-f10a-44a2-9084-9fbaa01c4bf0.png 非常期待您的回复。感谢!

— Reply to this email directly, view it on GitHub https://github.com/danfenghong/IEEE_TGRS_MDL-RS/issues/8, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFL2GZRCG5E2IKLS4OZ35FTXJLQBLANCNFSM6AAAAAAYZD3MR4 . You are receiving this because you are subscribed to this thread.Message ID: @.***>

G-z-w commented 1 year ago

这样会不会损失掉每个通道的偏置信息,只留下每个通道的相对值。

danfenghong commented 1 year ago

任何归一化都会损失一定的偏置信息,这是个trade-off的问题。看你主要是什么应用或是目的了,或是数据的质量如何,遥感图像质量良莠不齐,不像自然图像。

G-z-w @.***> 于2023年6月4日周日 00:32写道:

这样会不会损失掉每个通道的偏置信息,只留下每个通道的相对值。

— Reply to this email directly, view it on GitHub https://github.com/danfenghong/IEEE_TGRS_MDL-RS/issues/8#issuecomment-1575056790, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFL2GZRYD763DMOBG5TJVZTXJNRKLANCNFSM6AAAAAAYZD3MR4 . You are receiving this because you commented.Message ID: @.***>

G-z-w commented 1 year ago

任何归一化都会损失一定的偏置信息,这是个trade-off的问题。看你主要是什么应用或是目的了,或是数据的质量如何,遥感图像质量良莠不齐,不像自然图像。 G-z-w @.> 于2023年6月4日周日 00:32写道: 这样会不会损失掉每个通道的偏置信息,只留下每个通道的相对值。 — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFL2GZRYD763DMOBG5TJVZTXJNRKLANCNFSM6AAAAAAYZD3MR4 . You are receiving this because you commented.Message ID: @.>

感谢感谢!!