Oh this is a really broad question. That will depend of the purposes of your work!! Lately I am using numerical python and GDAL, because I am involved in process automation. Erdas is really nice, but you can also try free software like Spring (Brasil) or Ilwis (ITC). The smaller and easier Bilko.

This really depends on what you want to classify, the spectral quality of your images (number of bands, noise, etc...), and what you mean by classification: Clustering (SAM, KNN...)? Supervised classification/class-modelling (PLS-DA, SIMCA,...)? Non-linear classification (SVM, ANN)?

As far as I know, ENVI software is one of the standard ones. The problem is that it only has SAM (Spectral Angle Mapping) implemented for classification. SAM is a "kind" of clustering technique where you put some boundaries to define the classes by using the angle between the spectra (a kind of correlation coefficient). It works fine. But there are many other techniques that might be used.

Here, in our research group we have a PhD course entitled: "Hyperspectral and Multichannel Image Analysis". This is the webpage:

http://phdcourses.life.ku.dk/course/lphd190

Here we talk about classification, segmentation, clustering, etc. If you want, I can send you the material in pdf and you can have a look at it. Of course, we use MATLAB-based functions (HYPER-Tools - http://www.models.life.ku.dk/HYPERTools)

Another software that might be useful for you, if you want to skip MATLAB, is ImageJ. Easy to use. But the main problem is that you need to download (or create) macros. For sure there is a users community where you can download macros for whatever you want to do.

If you want more information or you have questions, just write me an e-mail ([email protected])!

Hope this is useful!

Jose

ENVI would be my choice as well because its functionality can be easily extended (if you want to) using IDL and/or ArcGIS.

ENVI is a good choice. Simple interface. The strength - a set of spectral tools.

Thanks for all helpful answers

There is no best software..Just use the software which find out your problem.Sometimes it could be open source software

I use IDRISI. I think it's easier to apply compared to ENVI but less flexible. ENVI and R are scripting languages which are very customizable. because you can write your own scripts. That's not possible with IDRISI. If your major goal is classification you should also take a look on eCognition.

From my experience during the last 13 years in remote sensing the best software for optical images analysis is ERDAS and the best software for microwave images analysis or (SAR) is GAMMA.

Depending on the product you use and processing you need, you may have a look at the Orfeo Toolbox : http://orfeo-toolbox.org/otb/

I used eCognition software. You can try eCognition. The multiresolution segmentation algorithm in eCognition is very good for remote sensing images.

We are offering an excellent forum for this type of discussions. Why not attending ForestSat 2014? Deadline for abstracts on April 30th, visit http://forestsat2014.com

I am voting ERDAS Imagine. It has a lot of tools for image processing/analysis.

ENVI definitely. In my opinion, it is the most friendly and powerful software for image analysis.

ERDAS and eCognition are best tools for classification. R can be used but a lot of study needs to be done for that. ENVI is good for hyperspectral dataset.

Oh this is a really broad question. That will depend of the purposes of your work!! Lately I am using numerical python and GDAL, because I am involved in process automation. Erdas is really nice, but you can also try free software like Spring (Brasil) or Ilwis (ITC). The smaller and easier Bilko.

I think that what will really solve this question is to determine the capability of these softwares commonly available with the vendors by caring a comprehensive comparative analysis of their ability in solving a particular real world problem, e.g Land use and Land cover Classification.

there are some programms using in remote sensing application , u must determine which the types of application u talk about. in my opinion and in my low experience in this field envi programm is better than other programms to deal with vegetations.

Hello Chemesse, Please go through herewith the attached code in matlab for ann based landslide mapping, i thought that this may helpful to you for further work.

ok thank you

In my opinion, Idrisi and ERDAS IMAGINE are appropriate and user-friendly softwares.

 IM Agree WITH Sumi Merlin Ecognation is the best for classification 

i suggest you to use both of ecognation & envi

I prefer Erdas Imgaine and ILWIS.

I will suggest you ERDAS IMAGINE. Its user interface is nice and easy to learn.

What is the best open sources software ?

The best software is the software that you feel comfortable with. This statement could mean a lot of things:

A software that has a friendly GUI (note that some people do not need friendly GUI)

A well documented software and with good support

A software that is used by your organization

A software that fulfill your processing requirenments.

A software with a complete set of processing algorithms

A software that support the kind of images you want to process

etc.,etc. etc.

The election of the "best software" will depends of the characteristics of your project, your needs, your level. But trying to do the things simple, people in general are using ERDAS or ENVI.

Cheers

Erdas Imagine is cool but not as affordable like ENVI. It should be made to have some free open source versions so that researchers in some economies can can come to terms with it.

there are some programs using in remote sensing application , you must first determine which the types of application you talk about.

Due to different applications, there are different software.