Dear Mr. Pimiskern

Thank you for your kind answer. but actually I didn't get your idea. can you please make it more clear or if you have some websites or article about that.

Thank you in advanced

Yes it is, but...

The most important questions are: what accuracy do you want to achieve? what spatial resolution do you need? Does the purpose of your work agree with the copyright of google earth?

You can of course work with screenshots and apply supervised classification techniques. Preferably OBIA if you work with very high reoslution images. The georeferencing of the tiles will be very time consuming. You will also suffer from a low spectral resolution and radiometric corrections from GoogleEarth you don't know about. So the accuracy of your result wont be as high as with original data.

Depending on the purpose of your work I suggest you have a look on the datasets available from NASA or ESA. Especially from ESA you can get access to high quality images with a small project proposal.

In google earth images there is no NIR band you only get RGB bands. I also tried to to do similar study for my work, but the reflectance value from the rooftops, roads, barren land was very similar to each other and it became very difficult to distinguish between these features.

I also put the same question and got some good replies from the fellow researchers.

You can go through these replies here -->

https://www.researchgate.net/post/How_to_achieve_higher_accuracy_of_land_use_cover_classification_using_Google_Earth_Imagery

I hope it may be useful for you.

Regards

DS Bisht

I must say before using Google Earth Imagery, please read permission guidelines. For your ready reference, please have a look attachment.

The whole concept is annotation....one concept is annotation or manual labeling like creating hotspot over images and another by automatic classification for annotated images.Thanks for the questions and more ideas and researches over the concept.

Thank you all for the valuable answers

The best you could do is create a shapefile or an ROI that could be used in a supervised classification within one of these three software: ArcGIS, ERDAS or ENVI. There are most likely other software that could help with the supervised classification,. but Google Earth is not one of them. Plus, the imagery does not hold enough information, as far as bandwidths, to be useful.

hy all, I have a question for you.

images from Google Earth are ortophotoplans!

I could do an analysis of land use using images from Google Earth?

I do not think so.You can not get requested spatial and spectral information from google earth images..

Object based classifier works better for high resolution images rather than parametric classifiers .

@Ana-Maria

You could do analysis however the values can not be validated values because the data is not original. My question is how you could get the orthoplans from google earth ?  

Hello! I am just reading the comments on this discussion topic. Since 2014, have the permission guidelines of Google Earth evolved? I can not anymore find the faq/permission guidelines illustrated above (today, in french: https://www.google.com/permissions/geoguidelines.html). In other words, is that possible that now it is allowed to use Google Earth as a basis?