How can I find nearest value in large grid data?

Question

Zubair Nizamani 2016-7-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/295448-how-can-i-find-nearest-value-in-large-grid-data

评论： Chad Greene 2016-7-14

I have a grid having cells 100x60, containing contours data. Some of the cells have filled with whole number from 1 to 15 showing the elevation of contours, now how can I interpolate points between the contours line to fill all empty cells?

Note: Points are not in regular way, these scattered points

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2016-7-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/295448-how-can-i-find-nearest-value-in-large-grid-data#answer_228545

See John D'Errico's File Exchange contribution "inpaint_nan"

2 个评论
显示无隐藏无

Zubair Nizamani 2016-7-13

Dear Roberson thank you for your response and guideline. The function you have mentioned above is right now harder for me to understand because I am getting started with MATlab, So can you direct me any simplest way possible?

Chad Greene 2016-7-14

在 MATLAB Online 中打开

If you download inpaint_nans, it makes the problem pretty easy. For example, say you have some gridded dataset like this:

Z = peaks(100);

but a few points are NaNs:

Z([423 898 1234] = NaN;

You can fill in the missing values like this:

Z_filled = inpaint_nans(Z);

请先登录，再进行评论。

Answer 2

Image Analyst 2016-7-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/295448-how-can-i-find-nearest-value-in-large-grid-data#answer_228557

How about scatteredInterpolant()?

5 个评论
显示 3更早的评论隐藏 3更早的评论

Zubair Nizamani 2016-7-13

编辑：Walter Roberson 2016-7-13

在 MATLAB Online 中打开

@Image Analyst, I have used the code mentioned below for interpolation, but I made an excel sheet in a way that, divided all the grid in 5x5 square boxes and then I assigned values on every corner of that boxes according to my best assumption, but when I plotted the contours, my contours lines were not smoothed and kinda uneven. So for now I want to make code that reads the excel as it and then interpolate.

If it is possible please do me favor here via helping me out.

[In excel sheet, only contour digitized values are present]

[Code]

clear all;
clc;
data = xlsread('Trial1.xlsx');
rlt = data ;
[height, width] = size(data);
[x,y] = meshgrid(0:5:width,0:5:height);
x(:,1) = 1; y(1,:) = 1;
z(1,1) = data(1,1);
for i = 1:height/5
      z(i+1, 1) = data(5*i, 1);
  end
  for j = 1:width/5
      z(1, j+1) = data(1, 5*j);
  end
  for i = 1:height/5
      for j = 1:width/5
          z(i+1,j+1) = data(5*i,5*j);
      end
  end
  [x1,y1] = meshgrid(1:width, 1:height);
  z1 = interp2(x, y, z, x1, y1, 'spline');
  for i = 1:height
      for j = 1:width
          if ~(data(height-i+1,j) > 0 )
              rlt(i,j) = z1(height-i+1,j);
     else
              rlt(i,j) = data(height-i+1,j);
          end
      end
  end
  % figure; mesh(x1,y1,z1); xlabel('longitude'); ylabel('latitude');
  % F = [.05 .1 .05; .1 .4 .1; .05 .1 .05];
  % ZC = conv2(rlt,F,'same');
  xlswrite('digital_value.xlsx', rlt);
  figure;
  %contour(rlt,20);
  [C,h] = contour(rlt);
  clabel(C,h);
  set(h, 'LevelList', [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19], ...
         'TextList', [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19])

Image Analyst 2016-7-13

I need to leave now, but my first attempt would be to use scatteredInterpolant, like I already suggested. But I'm sure you've tried that by now, so let's see your code - it may speed me up when I return later today. If you didn't try my suggestion, then why not?