MATLAB Picture Processing Code
On this instance, our code is performing characteristic detection on circuit boards to seek out factors of curiosity (corners) which can be utilized by the exterior software for some course of. In MATLAB, we have to load the picture, detect some options and extract them. The extracted options can be utilized in our software, so we’ll want entry to the ensuing array factors which point out the characteristic location.
The uncooked MATLAB code to carry out that is:
% Picture file title on the consumer aspect this can be a parameter
% High quality issue — on the consumer aspect this can be a parameter
% Learn the grayscale picture
% Discover nook candidates within the picture
corners = detectHarrisFeatures(I, ‘MinQuality’, qFac);
% Extract the legitimate corers
[~,valid_corners] = extractFeatures(I,corners);
To validate the outcomes on the MATLAB aspect, we’ll overlay the legitimate corners we extracted onto the picture. (Our consumer code will use these corners for different functions.)
J = insertMarker(I, valid_corners, ‘+’);
imshow(J);
operate [X, valid] = cornerFinder(imFileName, qFac)
% Use Harris nook detector on the enter file with a high quality issue given
% by qFac.
% Copyright 2022 The MathWorks(R), Inc.
% Carry out fundamental argument kind checking
arguments(Enter)
imFileName (1,:) char
qFac single
finish
% Allocate sorts for outputs (single, uint32, boolean)
X=zeros(0,2, ‘like’, qFac);
legitimate = false;
% Guarantee high quality issue is in vary.
if qFac < 0 || qFac > 1
fprintf(“Error: High quality issue have to be between 0 and 1.n”);
return;
finish
% Check to see that the given file exists and is readable
fID = fopen(imFileName, “r”);
if fID == -1
fprintf(“Error: File %s not discovered.n”, imFileName);
return;
finish
% Shut file take a look at
fclose(fID);
% Get the picture from the disk
I = imread(imFileName);
% Compute and extract the nook options
corners = detectHarrisFeatures(I, ‘MinQuality’, qFac);
[~,valid_corners] = extractFeatures(I,corners);
% Retailer the nook positions within the output array for the consumer and
% file the variety of factors. Set the legitimate flag to true.
X = valid_corners.Location;
legitimate = true;
finish
Producing Code and Making a Dynamic Library for Our Perform
- qFac: a single precision high quality issue
- imFile: a variable-sized character array for the filename (right here we use coder.typeof to specify this)
cfg = coder.config(‘dll’);
% Outline enter argument sorts for coder
imFile = coder.typeof(‘a’, [1,2048], [0, 1]);
% Name the coder to generate the dynamic library
codegen -config cfg cornerFinder.m -args {imFile, qFac} -nargout 2 -report
Trying into cornerFinder.h from the report reveals the consumer interface for the generated library. It has the prototype for the cornerFinder C operate. Be aware that the operate has three enter arguments, the title of the picture file, it is dimensions and the standard issue. There are two outputs, the array of factors and the legitimate flag. The vital a part of the interface code is right here:
extern void cornerFinder(const char imFileName_data[],
const int imFileName_size[2], float qFac,
emxArray_real32_T *X, boolean_T *legitimate);
We’ll use this data later to assemble our consumer code.
Packaging the Library and Its Dependencies
codebuild(‘./codegen/dll/cornerFinder’, ‘BuildMethod’, ‘CMake’, ‘BuildVariant’, ‘SHARED_LIBRARY’);
packNGo(‘./codegen/dll/cornerFinder’, ‘packType’, ‘hierarchical’, ‘fileName’, ‘myCornerFinderExample’);
This sequence of instructions generates what we’ll must go to our consumer. The codebuild command creates a CMakeLists.txt file which we’ll ultimately use to create a DLL from our code. The packNGo command generates a file referred to as myCornerFinderExample.zip. Inside this file are two different zip recordsdata, mlrFiles.zip and sDirFiles.zip. The primary of those accommodates all of the MATLAB runtime recordsdata and libraries wanted for the pc imaginative and prescient performance, and the second accommodates the generated code. On my machine, I now have this in my Present Folder:
Utilizing Our Archive to Construct a DLL with CMake
As soon as now we have our archive, we are able to hand it to our good friend who will compile it. They start by unzipping the archive and its subarchives. This produces a listing tree that appears like this:
C:work> dir /B c:workmyCornerFinderExample
mlrFiles
mlrFiles.zip
sDirFiles
sDirFiles.zip
C:work> cmake -G Ninja -B ./construct -DCMAKE_BUILD_TYPE=Launch -DMATLAB_ROOT=./myCornerFinderExample/mlrFiles/R2022b -S ./myCornerFinderExample/sDirFiles/codegen/dll/cornerFinder
C:work> cmake –build construct
This may create cornerFinder_win64.dll and cornerFinder_win64.lib. These can be within the .buildcodegendllcornerFinder listing.
Create a Easy Consumer Software
Now that our good friend has constructed our dynamic library, they will name it from their C++ program by way of the cornerFinder routine proven within the header above. On the coronary heart of this code, which is named myClientCode.cpp, is the decision to the cornerFinder operate. The traces of curiosity are right here:
/* Name the entry-point ‘cornerFinder’. */
emxInitArray_real32_T(&X, 2);
cornerFinder(imFileName_data, imFileName_size, qFac, X, &legitimate);
…
// Free the picture array
emxDestroyArray_real32_T(X);
The primary two traces create an array for storage of the corners and run the detector. The final line frees the purpose array.
Now our good friend creates a CMakeLists.txt file for his or her consumer code. It seems like this:
cmake_minimum_required(VERSION 3.12)
undertaking(MyClient)
set(CMAKE_CXX_STANDARD 14)
add_executable(myClient myClientCode.cpp)
target_include_directories(myClient PRIVATE
../myCornerFinderExample/sDirFiles/codegen/dll/cornerFinder
../myCornerFinderExample/mlrFiles/R2022b/extern/embrace )
target_link_directories(myClient PRIVATE ../construct/codegen/dll/cornerFinder
../myCornerFinderExample/mlrFiles/R2022b/extern/lib/win64/microsoft/
../myCornerFinderExample/mlrFiles/R2022b/toolbox/imaginative and prescient/builtins/src/ocvcg/opencv/win64/lib
)
target_link_libraries(myClient PRIVATE cornerFinder_win64.lib)
The consumer software is constructed equally to the DLL in earlier steps. On the command line within the ClientCode listing our good friend executes these instructions:
C:workClientCode> cmake -G Ninja -B construct -DCMAKE_BUILD_TYPE=Launch
C:workClientCode> cmake –build construct
This may generate myClient.exe. As soon as the generated and dependent DLLs are added to the PATH atmosphere variable, the consumer program will be executed to supply:
C:workClientCode> set PATH=C:workbuildcodegendllcornerFinder;%PATH%
C:workClientCode> set PATH=C:workmyCornerFinderExamplemlrFilesR2022bbinwin64;%PATH%
C:workClientCode> .buildmyClient.exe
Discovered some corners!
Variety of factors: 31 by 2
31.19 135.873
…
247.013 188.08
(Right here, we have eliminated all however the first and final rows of output information for brevity.)
Accessing This Instance
Now It is Your Flip
Inform us how you’ve got been utilizing the packNGo and codebuild instruments.
© 2023 The MathWorks, Inc.
