I need to know about the deep learning algorithms used in land cover classification and which one is best suited. Planning to use Sentinel-2 satellite images.
I also want to know about GAN in Landcover classification.
Have a look this may help you out
Article Land cover and land use classification performance of machin...
https://towardsdatascience.com/land-use-land-cover-classification-with-deep-learning-9a5041095ddb
Preprint A Cycle GAN Approach for Heterogeneous Domain Adaptation in ...
This is not peer-reviewed but interesting. This article uses cycle gan for land classification.
Ansith Sivadasan I do not know why do you want to use deep learning for land cover classification! There are many algorithms which are sometime comparable in accuracy to deep learning and less expensive computation wise.
Read the following:
https://www.kdnuggets.com/2016/04/deep-learning-vs-svm-random-forest.html
if you are not convinced than you can use the following deep learning land classification:
Deep convolutional neural networks for land-cover classification with Sentinel-2 images
Article Deep convolutional neural networks for land-cover classifica...
I suggest you follow a similar query as demonstrated here : Chapter Land Cover Classification Based on Sentinel-2 Satellite Imag...
Thank you so much for your replies to Muhammad Ali , Mohamad M. Awad , Byunghyun Ban and Mokhaled N. A. Al-Hamadani .
using deep learning for land cover classification may require more costly computing resources
Agree with Mohamad M. Awad. However there is not a one line answer to your question. the only way you will know which algorithm is suited for your application is to implement it and tune it. with High end APIs of python its a matter of few days. get your hands dirty.
Dear Ansith,
One of the most widely used machine learning algorithms are random forests (RF). The popularity of this algorithm is due to the fact it can be used for both classification and regression purposes. Second, you can use Earth-observing satellites, such as Sentinel-2 and Landsat-8. There are different ML classifications: with ensemble decision tree classifiers (e.g., Random Forest (RF), Bagging Trees (BT), Boosted Trees, etc.) and Support Vector Machine (SVM) being among the most commonly utilized for LULC classification. More recently, algorithms employing deep learning (DL) (i.e., Deep Neural Networks (DNNs)) have also become very popular for LULC classification.
Important articles you can get help:
https://arxiv.org/pdf/1710.03959.pdf
https://link.springer.com/article/10.1023/A:1010933404324
I hope it helps you.
Cheers
Ankur
Thank you sir Ankur Srivastava for your valuable comments. I will go through which you are recommending.
Dear Ansith,
There are several deep learning methods for suitable for land use classification, such as the Genetic algorithm, Convolutional Neural Networks (CNNs), Random forest classification etc.
The following latest articles may be helpful:
Article Hydrological Response to Agricultural Land Use Heterogeneity...
Cheers
- Badges
- Science topic