The straight forward choice will be to use Landsat images. But all remote sensing projects are objective driven. Depending on the level of precision you require, it will determine the data you go in for. So my advice will be that you have to be very clear in your mind what you want to achieve in the end. Landsat images have spatial resolution of 30m and so if your work is not for very detailed analysis then you can use it.
Landsat 8 and Sentinel 2a are good for land use pattern in urban areas. They are free. I think Sentinel 2a is better than Landsat 8 because it has spectral and spatial resolution batter than landsat.
You can download Sentinel 2a with this link: https://scihub.copernicus.eu/
Dear friend,i would like to inform you that choosing satellite image depends on the purpose of the study, if you're studying urban areas (areas - distance.....etc) so it should be used satellite images with high spatial resolution suh as Landsat 8 and Sentinel 2a they are good for land use pattern in urban areas and vice versa.
But if the purpose of your study, the study of climate elements in urban areas such as clouds or the Earth's temperature, in this case you must use satellite images with high spectral resolution (36 band) such as MODIS, and time can be overlooked spatial accuracy.
I agree with statement that Landsat is the best one due to its availability and time coverage! But the choice of image highly depends on the purpose or task! The Landsat images spatial resolution (pixel size 15-30 meter) sometimes is not enough, when you study small cities, or plan any other detailed research. Sentinel-2 has better spatial resolution (10-20 m), but it is "very young" - not much Earth area was covered by this satellite. There are many images with very high resolution (very detailed) which are represented at WWW-map services: eg. https://www.google.com/maps (Earth view), yandex.ru/maps (Satellite)... etc. I like to use Google Earth for my students-geographers (https://www.google.com/earth/ - need to be installed), as it contains new high resolution images for many places on the Earth, although they are mostly "pictures" - not suitable for computer analysis
Combination of optical data along with High incidence angle like polarized RADAR data acquired during summer season is probably the best data for Urban studies. Please go through one of my research publications using the link provided with the answer:
Comparative evaluation of potential of optical and SAR data for the detection of human settlements using digital classification (2006),
HS Srivastava, P Patel, Y Sharma, RR Navalgund - International Journal of Geoinformatics, Vol. 02, No. 03, pp. 21-28.
Right with regards to the type image data you need is based on the region you are located is that Africa or any part of the world is depend on the acquisition period a good example of sub-Sahara Africa a dry season is best time for identifications and other image interpretations of land use and land cover change and the good data for that is QUICK BIRD very high resolution image other 30 meter resolutions are land sat is free data from GLOVIS with ERDAS Imagine you can be able to classified land use and land cover in supervise classifications analysis of images of two different date data for change detection's with an accuracy assessments.
For urban studies you can use: Global Human Settlement Layer (derived from Landsat multitemporal), European Settlement Map (derived from Spot 5/6) or Global Urban Footprint (TerraSAR-X) datasets.
See its depends on your objective of study. If you want to study building level study then you need high resolution data but if you want to study on urban expatiation or regional level study then course resolutions data is needed. If you mention more clearly about your study then I can help you more.
I've seen Landsat and Sentinel both mentioned, and both are good options, but it also depends on where you will be conducting the analysis.
In many higher-income countries, such as the US, there is freely available high-resolution aerial photography, which can often include a near-infrared band, which is useful for vegetation classification.
If you're interested in doing LCLUC research in US urban areas, check out :