How do I extract the specific values from NDVI time serie?

I have generated ndvi input file time series over crop season(1 year period) and want to extract the following; POS value;MAX sum;AVG sum;SOS;POS;EOS The discription of above mentioned parameters are as follows; (a) Phenological markers • Start of season (SOS): the date of growth start and crop emergence • Peak of season (POS): the date of the maximum NDVI value in the time series • End of season (EOS): the date of the harvesting and chlorophyll ab- sence • POS value: maximum NDVI value in the time series •

Average sum: sum of average NDVI values in the time series

• Maximum sum: sum of maximum NDVI values in the time series

• Base: the range between minimum and maximum NDVI

I am attaching input ndvi file as excel csv file. I am also attaching the output file as a sample it has not been derived from the above mentioned two input files. This is just sample output file. I am requesting to matlab signal processing comunity to please suggest me how to write this code. I would be highly obliged for any kind help.

Jyoti

Answer 1

[Suraj Kumar]

Hi Devendra,

From what I understand, you are trying to extract specific phenological values from an NDVI time series using MATLAB.

For this you can refer to the following steps and the attached code snippets:

You can start by importing your NDVI data from the CSV file into MATLAB and then extract the dates and the NDVI values from the dataset.

% Load NDVI data from CSV 
data = readtable('ndvi_time_series_input.csv'); 
% Extract dates and NDVI values 
dates = data.Properties.VariableNames(2:end); 
ndvi_values = data{:, 2:end}; 

Initialize the arrays using ‘NaT’ function in MATLAB, to store the desired results.

sos = zeros(size(ndvi_values, 1), 1); 
pos = zeros(size(ndvi_values, 1), 1); 
eos = zeros(size(ndvi_values, 1), 1); 
pos_value = zeros(size(ndvi_values, 1), 1); 
avg_sum = zeros(size(ndvi_values, 1), 1); 
max_sum = zeros(size(ndvi_values, 1), 1); 
base = zeros(size(ndvi_values, 1), 1); 

Calculate the phenological markers by looping through each row of the NDVI data.

% Loop through each row in the NDVI data 
for i = 1:size(ndvi_values, 1) 
    [pos_value(i), pos_idx] = max(ndvi_values(i, :)); 
    pos(i) = datenum(dates{pos_idx}, 'yyyymmdd'); 
    sos_idx = find(ndvi_values(i, :) > (min(ndvi_values(i, :)) + 0.1 * (pos_value(i) - min(ndvi_values(i, :)))), 1); 
    if ~isempty(sos_idx) 
        sos(i) = datenum(dates{sos_idx}, 'yyyymmdd'); 
    end 
    eos_idx = find(ndvi_values(i, pos_idx:end) < (min(ndvi_values(i, :)) + 0.1 * (pos_value(i) - min(ndvi_values(i, :)))), 1); 
    if ~isempty(eos_idx) 
        eos(i) = datenum(dates{pos_idx + eos_idx - 1}, 'yyyymmdd'); 
    end 
    avg_sum(i) = sum(ndvi_values(i, :)) / length(ndvi_values(i, :)); 
    max_sum(i) = sum(max(ndvi_values(i, :))); 
    base(i) = range(ndvi_values(i, :)); 
end 

Compile these calculations into a table for easy analysis and interpretation and display the results.

results = table(data{:, 1}, sos, pos, eos, pos_value, avg_sum, max_sum, base, ... 
    'VariableNames', {'Location', 'SOS', 'POS', 'EOS', 'POS_Value', 'Avg_Sum', 'Max_Sum', 'Base'}); 
disp(results); 

You can look at the output for a better understanding:

For more details on the ‘NaT’ or ‘datenum’ functions in MATLAB, you can refer to the following links:

• https://www.mathworks.com/help/matlab/ref/nat.html
• https://www.mathworks.com/help/matlab/ref/datenum.html

Happy coding!