inverse distance weighting on matrix

Question

jenifer Ask 2019-12-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/497965-inverse-distance-weighting-on-matrix

编辑： Image Analyst 2019-12-25

采纳的回答： Image Analyst

Dis_all.mat

hi

I have Below matrix that distanse from 17 point in image

How i can calculate inverse distance weighting on this matrix?

how weighting on distance that have minim distanse for each row?

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

Image Analyst 2019-12-25

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/497965-inverse-distance-weighting-on-matrix#answer_407657

在 MATLAB Online 中打开

This 17-by-17 matrix looks like it was made by pdist2(). And you say it is the distance of every point to every other point. If you want 1 over the distance, just do

inverseDistances = 1 ./ Dist;

To get the min value of inverseDistances, do

minsPerRow = min(inverseDistances, [], 2);

To get the min distance of the original Dist, do

dist2 = Dist; % Initialize
dist2(dist2 == 0) = inf; % Trick to get it to ignore zero distances (point to itself)
minsPerRow = min(dist2, [], 2);

3 个评论
显示 1更早的评论隐藏 1更早的评论

Image Analyst 2019-12-25

编辑：Image Analyst 2019-12-25

在 MATLAB Online 中打开

You can call sort on each row.

s = load('Dis_all.mat')
m = 1 ./ s.Dist
[rows, columns] = size(m)
[sortedDistances, sortedIndexes] = sort(m, 2)

sortedIndexes are the indexes in the original, unsorted array. The first 3 columns of that are the indexes of the three other points that are closest to that point. So

sortedIndexes =
  16     7    15    14    11    13    12     8    10     5     9     6     4     3     2     1
  16     7    15    14    11    13    12     8    10     5     9     6     4     3     1     2
  16     7    15    14    11     8    13    12    10     1     2     5     9     6     4     3
  16     7    15    14    11     8    13    12    10     1     2     5     9     6     3     4
  16     7    15    14     1    11    13     2    12     9     8     6    10     4     3     5
  16     7     8    11    15    14    13    12    10     1     5     2     3     4     9     6
  16     1     9     2     6     4     3    15    14     5    12    13    10    11     8     7
  16     1     2     9     6    15    14     7     4     3    13    12     5    11    10     8
  16     7     8    11    15    14    13    12    10     1     5     2     3     4     6     9
  16     7     1     2     9     6     3     4    15    14     8     5    11    13    12    10
  16     1     2     7     9     6     3     4     5     8    15    14    10    12    13    11
  16     7     1     2     6     9     3     4     8     5    15    14    10    11    13    12
  16     7     1     2     6     9     3     4     5     8    15    14    10    11    12    13
   7     1    16     2     6     3     4     9     8     5    10    11    12    13    15    14
   7     1    16     2     6     3     4     9     8     5    10    11    12    13    14    15
   2     7     3     4     6     5     9     8    10    11    12    13    14    15    17    16
   2     3     6     4     9     7     5     8    10    11    12    13    14    15    16    17
     

You can see that for point 1 (represented by row 1), point 17 is closest, followed by point 16, followed by point 7, with point 2 being farthest away (actually closest in actual distance but farthest in inverse distance).

As another example, for point 16 (represented by row 16) point 1 is closest, followed by point 2, followed by point 7, and point 17 is farthest away.

Why do you want inverse weighting rather than just on the actual distances directly?

jenifer Ask 2019-12-25