Hello Prof Makki Kaze your question is not clear.

If you are looking for a novel strategy it could be that combines the DL with OBIA. For remote sensing image preprocessing, you may look at SAEs and CNNs .

You may want to look at filling the gap of DL in image registration.

Good luck

Makki Kaze , It's depend on your Applications & Research Areas for Remote Sensing Data. But here is the sample methods for remote sensing Application :

For this special issue, we welcome the most recent advancements related, but not limited to:

* Deep learning architecture for remote sensing

* Machine learning for remote sensing

* Computer vision method for remote sensing

* Classification / Detection / Regression

* Unsupervised feature learning for remote sensing

* Domain adaptation and transfer learning with computer vision and deep learning for remote sensing

* Anomaly/novelty detection for remote sensing

* New dataset and task for remote sensing

* Remote sensing data analysis

* New remote sensing application

* Synthetic remote sensing data generation

* Real-time remote sensing

* Deep learning-based image registration

-------------------

Sample Articles for Remote sensing with DL, pls visit the link

(*)Article Deep Learning in Remote Sensing: A Review

(*)Article A survey of remote sensing image classification based on CNNs

"... As a field of research, AI traces its history back to the dawn of computer science, in the 1950s. The term “AI” itself was coined in 1956. In the 1960s, “perceptrons” (see Section 3.1for technical details) were taken to exemplify the possibility that a machine could “learn” from data." .... "The basic constructs of deep neural networks , convolutional neural networks , and back-propagation were all in place by the 1980s. However, DNNs did not become the technology of choice for many applications until after 2010. The delay of more than 20 years in the application of DNNs to key problems was because two key ingredients were missing: labeled data sets for training, and sufficiently powerful hardware for training." from 'Perspectives on Research in Artificial Intelligence and Artificial General Intelligence Relevant to DoD' at https://fas.org/irp/agency/dod/jason/ai-dod.pdf

In 1961, the Samos E-1 satellite had a film readout system: the camera recorded an image on film, and then this film was developed while in orbit, scanned by an electronic system and the data downlinked back to the ground. ( https://satelliteobservation.net/2016/07/30/history-of-the-us-reconnaissance-system-i-imagery/ ). Even before this, U-2 aerial imagery was being digitized. So DoD was confronted with continent sized volumes of vast amounts of imagery, that no amount on human photo interpreters could process, so they massively funded research into various algorithms, such as automated object detection and what would later be known as 'machine learning'. ( https://www.hsdl.org/?view&did=441606 )

Of course, this eventually spilled over into the commercial world, but bottle-necked for decades because of the expensive high performance hardware needed. CNNs? How about 1975? "... We implement an improviser for Scenic scenarios and apply it in a case study generating synthetic data sets for a convolutional neural network designed to detect cars in road images. ..."

Most of the common 'A.I.' algorithms have a deep history in Remote Sensing, albeit sometimes with different nomenclature. It has been only recently when economical computing made them available to the mainstream.

And Everything Old is New Again.

Attached is some 1965 era defense satellite imagery - imagine trying to look over the entire area of Russia at .5 meter resolution.

I am not sure what you are asking, yes, there are a lot of algorithms but the use of one depends on the problem at hand, familiarity, etc.

If you are looking for a novel strategy it could be that combines the DL with OBIA. For remote sensing image preprocessing, you may look at SAEs and CNNs .

Peace, mercy and blessings of God be upon you. Thank you very much. Happy New Year

happy holiday