deep learning model deploy to raspberry pi using MATLAB
조회 수: 1 (최근 30일)
이전 댓글 표시
code
function PDR2()
%#codegen
raspiObj = raspi();
cam = cameraboard(raspiObj);
net = coder.loadDeepLearningNetwork('resnet50');
inputSize = [224, 224, 3];
textToDisplay = '......';
start = tic;
fprintf('Entering into while loop.\n');
while true
img = snapshot(cam);
elapsedTime = toc(start);
if elapsedTime > 1
imgSizeAdusted = imresize(img,inputSize(1:2));
[label,score] = net.classify(imgSizeAdusted);
maxScore = max(score);
labelStr = cellstr(label);
textToDisplay = sprintf('Label : %s \nScore : %f', labelStr{:}, maxScore);
start = tic;
end
img_label = insertText(img,[0,0],textToDisplay);
displayImage(raspiObj,img_label);
end
end
DEPLOY
board = targetHardware('Raspberry Pi');
board.CoderConfig.TargetLang = 'C++'
dlcfg = coder.DeepLearningConfig('arm-compute');
dlcfg.ArmArchitecture = 'armv7';
dlcfg.ArmComputeVersion = '19.05'
board.CoderConfig.DeepLearningConfig = dlcfg
deploy(board, 'PDR2')
error
Deploying code. This may take a few minutes.
------------------------------------------------------------------------
Error executing command "touch -c /home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2/*.*;make -f PDR2_rtw.mk all MATLAB_WORKSPACE="/home/pi/MATLAB_ws/R2020b" -C /home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2". Details:
STDERR: In file included from MWCNNLayerImpl.hpp:10,
from MWAdditionLayerImpl.hpp:8,
from MWAdditionLayer.cpp:4:
MWACLUtils.hpp:12:10: fatal error: arm_compute/runtime/NEON/NEFunctions.h: No such file or directory
#include "arm_compute/runtime/NEON/NEFunctions.h"
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [PDR2_rtw.mk:358: MWAdditionLayer.cpp.o] Error 1
STDOUT: make: Entering directory '/home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2'
g++ -c -MMD -MP -MF"MWAdditionLayer.cpp.dep" -MT"MWAdditionLayer.cpp.o" -fpermissive -O2 -std=c++11 -mfpu=neon -fopenmp -D_POSIX_C_SOURCE=199309L -DDISABLE_JPEG_ENCODING -DUSE_19_05_LIBRARY -D__MW_TARGET_USE_HARDWARE_RESOURCES_H__ -DMW_MATLABTARGET -D__linux__ -DARM_PROJECT -D_USE_TARGET_UDP_ -D_RUNONTARGETHARDWARE_BUILD_ -DSTACK_SIZE=200000 -DMODEL=PDR2 -I/home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2 -I/home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project -I/home/pi/MATLAB_ws/R2020b/C/Program_Files/MATLAB/R2020b/toolbox/eml/externalDependency/timefun -I/home/pi/MATLAB_ws/R2020b/C/ProgramData/MATLAB/SupportPackages/R2020b/toolbox/realtime/targets/raspi/server -I/opt/userland/host_applications/linux/libs/bcm_host/include -I/opt/userland -I/opt/userland/host_applications/linux/apps/raspicam -I/opt/userland/interface/vcos -I/opt/userland/interface/vcos/pthreads -I/opt/userland/interface/vmcs_host/linux -I -I/include -I/home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2/examples -I/home/pi/MATLAB_ws/R2020b/C/ProgramData/MATLAB/SupportPackages/R2020b/toolbox/realtime/targets/raspi/include -I/home/pi/MATLAB_ws/R2020b/C/Program_Files/MATLAB/R2020b/toolbox/coder/rtiostream/src/utils -I/home/pi/MATLAB_ws/R2020b/C/Program_Files/MATLAB/R2020b/extern/include -o "MWAdditionLayer.cpp.o" "MWAdditionLayer.cpp"
make: Leaving directory '/home/pi/MATLAB_ws/R2020b/C/Users/lenovo/Desktop/Project/codegen/exe/PDR2'
------------------------------------------------------------------------
??? Build error: C++ compiler produced errors. See the Build Log for further details.
More information
Code generation failed: View Error Report
what can I do?
please help me
Does it need GPU?
댓글 수: 1
Prasanth Sunkara
2021년 3월 26일
편집: Prasanth Sunkara
2021년 3월 26일
Along with Jason's suggetion, please check that you have arm_compute library installed on your pi.
답변 (1개)
Jason Ross
2021년 3월 18일
편집: Jason Ross
2021년 3월 18일
googling for "arm_compute/runtime/NEON/NEFunctions.h" returned this Answer
참고 항목
카테고리
Help Center 및 File Exchange에서 MATLAB Support Package for Raspberry Pi Hardware에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!