A Superpixel Spatial Intuitionistic Fuzzy C-Means Clustering Algorithm for Unsupervised Classification of High Spatial Resolution Remote Sensing Images

Round 1

Reviewer 1 Report

The paper « 

 A superpixel spatial intuitionistic fuzzy C-means clustering al-gorithm for unsupervised classification of high spatial resolution remote sensing images  » is well written and proposes 

a superpixel spatial intuitionistic fuzzy C-means (SSIFCM) clustering algorithm to address the problems of misclassification, salt-and-pepper noise and classification uncertainty arising in pixel-level unsupervised classification of high spatial resolution remote sensing (HSRRS) images. 

To reduce information redundancy and ensure antinoise and image detail preservation, the authors  use a superpixel segmentation to obtain the local spatial information on the HSRRS  image. 

On the detail preservation earlier works are not cited, I suggest to the authors to refer to the papers of Albert Cohen & al. (2001-2002) on the details reconstuction based of local spatial information.  

 

The paper show numerical experiments shows that the superpixel segmentation can reduce information redundancy and make comprehensive use of the spectral and spatial information of the image.  I propose to accept the paper with minor revisons under the condition that the authors add the references  on the details reconstuction based of local spatial information.  

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper proposes a method for unsupervised image segmentation.  The  novelty of the proposed algorithm (section 2) is unclear:

- "The SSIFCM algorithm *was* proposed to solve the problems of misclassification..." (section 2, 1st sentence ) : are you referring  to a previously published SSIFCM algorithm? if so what is the reference?

- SLIC segmentation algorithm is well known and section 2.1 is not new. Please separate more clearly what is state of the art work from claim of novelty.

- mathematics notations are inconsistent and difficult to follow e.g.  definition of S_j and S (equation (1)), what is G_g (Eq. (5)), etc.

- The proposed unsupervised approach is badly justified considering the success of the Convolution Neural Networks (supervised learning) for segmenting aerial imagery:  the sentence  *In the practical applications of supervised classification, there are some problems, such as cumbersome sample preparation and weak transferability* (in introduction)  is not accurate.

The writing and presentation needs to be improve  e.g. *antinoise* (abstract), what is *internal distribution law of target data* ? etc.

 

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

A superpixel spatial intuitionistic fuzzy C-means  clustering algorithm to classify high spatial resolution remote sensing images is proposed.

This paper is interesting and the reasons and objectives of the research are well described.

However, there are the following critical points that I recommend to the authors to consider:

1) A more structured description of the functional components of the proposed algorithm is required, reported in the flow diagram in fig. 1. In particular, it is necessary to highlight the characteristics of the functional components of SSIFCM by connecting them with the phases and formulas described above.

2) There is no presentation in Pseudocode of the SSIFCM algorithm; It is also necessary to analyze the computational complexity of the proposed algorithm.

3) The authors must discuss the computational complexity of SSIFCM and show how it varies in varying the size and resolution of the Remote Sensing Image.

4) It is necessary to add a comparison on the processing times. The authors must show that the elaboration times of SSIFCM are not higher than those of SFFCM, FRFCM and SPFCM, and therefore allow you to elaborate also Very High Resolution Remote Sensing Images.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Authors took into account all my suggestions. I consider this paper publishabe in the present form.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.