After calculating water indices NDWI , MNDWI ; how to check its accuracy ?

21 February 2024 1 339 Report

we have calculated water indices after that how can we check its accuracy .

Verifying the accuracy of water indices like NDWI and MNDWI requires ground truth data and some statistical analysis. Here are the steps involved:

1. Gather Ground Truth Data:

High-resolution imagery with accurate water body boundaries is crucial. This can be:Field surveys: Collect GPS points or delineate water bodies manually on high-resolution aerial photographs. Reference data: Look for publicly available water body maps or datasets aligned with your study area.

2. Compare Calculated Indices with Ground Truth:

Overlay your NDWI and MNDWI maps with the reference water body boundaries.
Use GIS tools to calculate various metrics:Overall Accuracy: Percentage of correctly classified pixels (water and non-water). Producer's Accuracy: Percentage of true water pixels classified as water by the indices. User's Accuracy: Percentage of pixels classified as water that are actually water. Kappa Coefficient: Measures agreement beyond chance, considering all four confusion matrix categories.

3. Analyze and Interpret Results:

Compare the accuracy metrics for NDWI and MNDWI to understand which performs better for your specific study area and conditions.
Consider factors like:Image resolution and quality. Presence of other water bodies or land cover types that might confuse the indices. Thresholding method used to separate water from non-water pixels.

Additional Tips:

You can also perform visual inspection to identify areas where the indices misclassify water or non-water.
Consider using advanced accuracy assessment techniques like stratified random sampling or object-based image analysis.
Consult relevant scientific literature for established methods and accuracy benchmarks for your specific application.

Remember, accuracy assessment is crucial for validating your water body mapping results and ensuring their reliability. By following these steps and carefully analyzing the results, you can gain valuable insights into the performance of NDWI and MNDWI for your specific context.

Badges
Science topic

More Shriram Kathar's questions See All

Can anyone tell me when to use Shannon and Phylogenetic diversity (PD_whole_tree) while calculating species richness in metegenomic data?

Currently I am analyzing amplicon sequencing data generated from diseased and control individuals. In that observed OTUs are more in control individuals but Shannon index is significantly higher...

16 March 2016 2,671 5 View

Publishing in two different journals with same genomic data set will counted as plagiarism?

Now a days journals like Data in Brief, Genome Announcements and Genomics Data publish short manuscripts or brief data about genomics or metagenomics work. Role of such journals is to inform...

05 August 2015 739 9 View

How effective is the RCA-1 and RCA-2 cleaning of ITO substrates followed by O2 plasma treatment?

12 March 2015 2,462 2 View

How should I analyze the multiple scans in Malvern Zetasizer software?

Hello, I am currently using the Malvern Zetasizer software for the purpose of obtaining the size distribution of nanoparticles produced in the lab. When I perform the runs in the software to get...

15 February 2021 4,251 1 View

Which optimization algorithm is most efficient for fitting a regression model using training dataset?

The goal of regression model is either to minimize the prediction error or to maximize a classification accuracy depending on the type of the model. Is there any optimization algorithm that is...

10 February 2021 4,128 3 View

Why does the accuracy of NN decreases when two datasets are combined?

I am working on gait analysis using a feed-forward NN. When I use data from the accelerometer (3 columns for each axis if stored in a data frame), I get an accuracy of 95%, and when I used data...

07 February 2021 9,654 6 View

Which techniques are used to improve the accuracy of deep learning models by compressing neural networks?

Hi all, I am looking for enhancing the accuracy of a deep learning model using some techniques to compress neural networks. Could someone suggest to me some of the tools/techniques,...

29 January 2021 1,781 4 View

When is the right time to stop training a classification algorithm?

In this example giving 400 samples looks like there is an accuracy of 91.5% while training beyond that point, accuracy is getting decreased. Are 400 samples enough to keep for this example? If...

27 January 2021 5,657 3 View

45% top1 Accuracy on TinyImgeNet?

I trained a model from scratch no transfer learning of any kind I got top1 validation accuracy of 45.79 and top5 validation accuracy of 72.39. I did a lot of hyperparameters tuning I have seen...

19 January 2021 6,883 2 View

How to express accuracy of the neural network where the dataset has not been split into two, training set and test set?

In use, the weights obtained are considered with the greatest accuracy when the data set is divided, because the weights are initialized at random and the training data is swapped at random...

19 January 2021 5,714 4 View

Multicollinearity between bioclimatic variables in spatial distribution modelling?

Is there any manual guide to perform multicollinearity test between environmental (19 Bioclimatic) variables for MaxEnt modelling? Using all variables without removing the correlated one affects...

14 January 2021 3,807 7 View

Why do I have good performances for the test set but bad performances in predicting new data ?

Hello :) I'am trying to predict churn for the year of 2020 based on historical data from 2014 to 2019 on R Studio. The aim is to predict churn probability for every person. At the beginning...

13 January 2021 5,565 7 View

How to change an internal parameter during MATLAB fminseach?

An optimization is performed using the MATLAB fminseach function. This optimization has a parameter that adjusts the precision, and the larger the value, the wider the search range of solution...

12 January 2021 3,832 1 View