Hi

Please see http://www.cropsreview.com/crop-classifications.html

The principle of all remote sensing software is the same, and accuracy of classification mainly depends on data type and quality. Landsat MSS has a coarse spatial resolution (80m). also it uses only 4 spectral wavebands, so it is good for large field size and level one or two classification. It is difficult with such resolution to get detailed crop classification.

Hello. I am agree with Ahmed, Landsat has too high spatial resolution to classify sub crops. Maybe, you can use Landsat if the croplands have big size.

Thank you Ahmed and Ana, I will try with more high resolution satellite data because I want to classify sub crops because in my study area its all mixed crop and definitely it will be difficult....

Your question maybe which best satellite data can used for sub agriculture crops, anyway this depend on the spatial and spectral resolution of the image used and so SPOT 2.5 is good because it high spatial resolution, or hyperion as it has a very good spectral resolution, the any remote sensing software could do the classification

I could strongly recommend you RapidEye or Worldview-2 satellite images which are commonly used for crop type mapping. For the case of software, I could suggest you ENVI or Idrisi Taiga which are commercial softwares. If you prefer open source software, Orfeo toolbox or Qgis could be proper for you.There is not any best software.Each one has pros and cons therefore there is not best software...

Thank you Mustafa, as I am already using SPOT satellite images and for classification I am using ENVI 5.0 software and in one of my case study Maximum Likelihood method and Majority and minority 3*3 and 5*5 filters are giving good results. I highly appreciate for your answer.

SVM works quite better than MLC if optimum parameters were selected. For image classification, it would be sometimes better to choose machine learning algorithms such as SVM, RF etc.

Yes, I tried with SVM software to classify Landsat image but it didnt worked as I am also new one to use SVM. I searched for reasons why it not worked, so I thought better to not waste time with stucking with SVM and get work done by using ENVI. Because I have very short time to complete my project.

First, try to find optimum parameters for each type of kernel .Then try again.

Check the articles below:

Comparison of artificial neural networks and support vector machine classifiers for land cover classification in Northern China using a SPOT-5 HRG image

Xianfeng Song, Zheng Duan, Xiaoguang Jiang

International Journal of Remote Sensing

Vol. 33, Iss. 10, 2012

Support vector machines for classification in remote sensing

M. Pal, P. M. Mather

International Journal of Remote Sensing

Vol. 26, Iss. 5, 2005

Dear, Mustafa, can you send me this above mentioned papers on my email.

my email id is :[email protected].

I will be very thankful to you.

Sure.I hope it could help you.

All The Best!

Thank you very much Mustafa Ustuner, I will study this papers in detail and try to use SVM for classification.

Hi

Please see http://www.cropsreview.com/crop-classifications.html

Dear, Dr. Mostafa Eidiani thank you for sharing important info related to all the crop types and all basic required  info related to them....