如何处理图像范围错误？

Question

以下 matlab 脚本将较大的蒙版图像剪辑到较小灰度图像的边界框，并屏蔽掉较小图像的区域。只要较小的图像在较大图像的范围内，这就会很好。但是，如果较小的图像超出了较大图像的范围（附截图），我会收到以下错误：

Index exceeds matrix dimensions.

Error in createMask (line 33) 
In = I(NI(1):NI(2),NI(3):NI(4));  % clipped binary

我希望将大图范围之外的小图的任何像素值转换为 0。我可以使用哪些方法来处理此图像范围错误？

---

---

%% this program clips a larger image using the extent of a smaller image..
% resample the images to match the pixel size and multiply with each other
%This program requires that the clip image is entirely within the larger image

[I B] = geotiffread('MASK.tif'); % larger image
[Ic R] = geotiffread('NDVI.tif'); % clip image
infoNDVI = geotiffinfo('NDVI.tif'); % geoinfo of clip layer
infoMASK = geotiffinfo('MASK.tif'); % geoinfo of larger layer

ncolsMASK = infoMASK.Width; 
nrowsMASK = infoMASK.Height;
nrowsNDVI = infoNDVI.Height;
ncolsNDVI = infoNDVI.Width;

B1 = infoMASK.BoundingBox;
Bm = infoNDVI.BoundingBox;

res = infoMASK.PixelScale(1); % res is the pixel dimension 

% Calculate the X and Y distances of the four corners of the clipping image from the origin of the larger image

d1 = fix([ (Bm(1) -B1(1)) (Bm(3) -B1(3))]/res);
d2 = fix([ (Bm(1) -B1(1)) (Bm(4) -B1(3))]/res);
d3 = fix([ (Bm(2) -B1(1)) (Bm(4) -B1(3))]/res);
d4 = fix([ (Bm(2) -B1(1)) (Bm(3) -B1(3))]/res);

% calculate row column indices of the clip layer and extract values for the clip extent
% Downscale resolution of the larger image to match the clip image
% Multiply the binary images and write that output as a geotiff file

NI = [ (nrowsMASK - d3(2)) (nrowsMASK - d1(2))  d1(1) d3(1) ];    
In = I(NI(1):NI(2),NI(3):NI(4));  % clipped binary

resampledIn = imresize(In, [nrowsNDVI, ncolsNDVI]);  % resample the binary layer to the size of clipped layer
intersected = immultiply(resampledIn,Ic);
outfilename = ['clipIntersect1'  '.tif'];
geotiffwrite('OUTPUT.tif', intersected, R, 'GeoKeyDirectoryTag', infoC.GeoTIFFTags.GeoKeyDirectoryTag);

Answer 1

我能想到的最简单的方法是使用逻辑索引来检查索引的有效性。下面是一个小例子来展示它是如何完成的：

I = ones(10);
NI = [-2 3 8 12];

% Get your rows and columns
iRows = NI(1):NI(2);
iCols = NI(3):NI(4);
desiredSize = [numel(iRows), numel(iCols)];

% Check validity
ilValidRows = iRows > 0 & iRows <= size(I, 1);
ilValidCols = iCols > 0 & iCols <= size(I, 2);

% set everything to zero
In = zeros(desiredSize);

% assign the values based on logical indexing
In(ilValidRows, ilValidCols) = I(iRows(ilValidRows), iCols(ilValidCols));

您最终得到In 的以下值

In =

     0     0     0     0     0
     0     0     0     0     0
     0     0     0     0     0
     1     1     1     0     0
     1     1     1     0     0
     1     1     1     0     0