Howdy Peter,

Mind that CORONA imagery is NOT multispectral. so forget about classifications based on multispectral CORONA data. On my ResearchGate site at  : 

https://www.researchgate.net/profile/Frank_Veroustraete

you can find a paper at:

https://www.researchgate.net/publication/235959317_Diachronic_Mapping_of_LUCC_in_the_Northwest_of_China_over_the_last_half_Century_Conversion_of_CORONA_Panchromatic_reflectance_into_LANDSAT_NDVI_Imagery

in which I used CORONA imagery for the Norhtwest of China (Xinjiang Province) in which I converted CORONA monochromatic data into NDVI imagery. The methodology is described. You can perform LUCC studies covering a period of 50 years with this methodology.

I do not know of multispectral missions with a spatial rsesolution of less than 20 meters from before 1986. Maybe some archived airborne campaigns? If you are happy with less than 30 meters multipsectral, LANDSAT offers possibilities.

See the LANDSAT mission timeline at:

http://landsat.usgs.gov/about_mission_history.php

The LANDSAT timeline starts in 1970. If that 20 meters is critical, then there is nothing left before 1970, just monochromatic data!

Cheers,

Frank

Article Diachronic mapping of LUCC in the northwest of China over th...

Dear Peter,

if b/w imagery is ok for you, I suggest you take a look at the declassified Corona images, going back to the 1960s:

https://lta.cr.usgs.gov/declass_1

The data can be accessed via EarthExplorer (Declassified Data):

http://earthexplorer.usgs.gov/

Many scence are available for free, others at very low prices (~30 USD, if I remember right).

Best regards,

David

Dear,

You can explore the USGS earth explorer web for Landsat data as well as CORONA http://earthexplorer.usgs.gov/

I am very interested to see if there are any answers to this - SPOT1 (20m resolution) went up in 1986 but I always assumed if you were looking for higher MS resolution than Landsat back then you should look for airborne imagery rather than satellite. 

Howdy Peter,

Mind that CORONA imagery is NOT multispectral. so forget about classifications based on multispectral CORONA data. On my ResearchGate site at  : 

https://www.researchgate.net/profile/Frank_Veroustraete

you can find a paper at:

https://www.researchgate.net/publication/235959317_Diachronic_Mapping_of_LUCC_in_the_Northwest_of_China_over_the_last_half_Century_Conversion_of_CORONA_Panchromatic_reflectance_into_LANDSAT_NDVI_Imagery

in which I used CORONA imagery for the Norhtwest of China (Xinjiang Province) in which I converted CORONA monochromatic data into NDVI imagery. The methodology is described. You can perform LUCC studies covering a period of 50 years with this methodology.

I do not know of multispectral missions with a spatial rsesolution of less than 20 meters from before 1986. Maybe some archived airborne campaigns? If you are happy with less than 30 meters multipsectral, LANDSAT offers possibilities.

See the LANDSAT mission timeline at:

http://landsat.usgs.gov/about_mission_history.php

The LANDSAT timeline starts in 1970. If that 20 meters is critical, then there is nothing left before 1970, just monochromatic data!

Cheers,

Frank

Article Diachronic mapping of LUCC in the northwest of China over th...

Dear Peter,

CORONA is a good solution if there is sufficient high quality coverage of your area. One of my colleagues, Niko Galiatsatos has worked a lot with CORONA - here is his researchgate profile pages with links to some relevant publications.

https://www.researchgate.net/profile/Nikolaos_Galiatsatos/publications

As earlier mentioned by David Loibl and Muhammad Ateeq Qureshiby earthexpolrer is the best one and i hope you will find there.

In addition to the earlier answers, I suggest you  to look at Landsat data via USGS

SPOT 1 launched February 22, 1986 with 10 panchromatic and 20 meter multispectral picture resolution capability. Withdrawn December 31, 1990 I suggest pansharpen image for your study.

As Frank has mentioned  CORONA imagery is not multispectral data, so aerial imagery is the way to go. 

If you can not find fine resolution multispectral data, you can try to combine lower res multispectral and high res panchromatic data from different sensors, like Landsat + CORONA or Landsat + airborne data. One way to do it is pansharpening. If your task is focused on image classification/segmentation/object recognition, there are also methods which can use multi-resolution data directly (but this is not my specialization, I just have seen some articles).

I do recommend again as others to look at the USGS website where you can find the Landsat imageries, by registering on the https://earthexplorer.usgs.gov/  link. However, the resolution is not less than 30 m. 

The less than 20m resolution criteria places a huge limit on possibilities before 1986. LANDSAT is definitely out of the question for less than 20m resolution.

Cheers

You can try employing SPOT imagery that has a spatial resolution of 10m and has been in use for a long time and certainly before 1986.

Resurs-F also may be realy good decision. But I don't know anything about availability of its images. You may try to contact SRC Planeta (see a link below) to get detailed information on Resurs imagery.

http://planet.rssi.ru/english/index_eng.htm

SPOT 1 was launched in early 1986 it had a spatial resolution of 10 m in panchromatic and 20 m in each of the three multispectral bands. However, the spectral resolution was rather low (only in green, red and near infrared). Before 1986 Landsat TM had a spatial resolution of 30 m, but it had a better spectral resolution than SPOT, with seven bands. It is to be noted that Russian space photographs of before 1986 achieved spatial resolutions of 5 m and less, again not many spectral bands though, and they were analogue photos which you may well digitize.

 I recommend you to use Landsat 4 or 5 TM for the time before 1986 although the images have 30 m resolution and possibly you can do re-sampling the pixels to 15 m which may help.