代码先锋网 代码片段及技术文章聚合

opencv已安装

下载源码
git clone –recursive https://github.com/rbgirshick/py-faster-rcnn.git

安装python依赖包
将caffe-fast-rcnn/python目录下的requirements下的依赖都装一遍

Cython>=0.19.2
numpy>=1.7.1
scipy>=0.13.2
scikit-image>=0.9.3
matplotlib>=1.3.1
ipython>=3.0.0
h5py>=2.2.0
leveldb>=0.191
networkx>=1.8.1
nose>=1.3.0
pandas>=0.12.0
python-dateutil>=1.4,<2
protobuf>=2.5.0
python-gflags>=2.0
pyyaml>=3.10
Pillow>=2.3.0
six>=1.1.0

命令：

for req in $(cat requirements.txt); do sudo pip install $req; done

修改py-faster-rcnn/lib/setup.py，注释掉GPU相关代码，如下

。。。
#CUDA = locate_cuda()
。。。
#            self.set_executable('compiler_so', CUDA['nvcc'])
。。。
#    Extension('nms.gpu_nms',
#        ['nms/nms_kernel.cu', 'nms/gpu_nms.pyx'],
#        library_dirs=[CUDA['lib64']],
#        libraries=['cudart'],
#        language='c++',
#        runtime_library_dirs=[CUDA['lib64']],
#        # this syntax is specific to this build system
#        # we're only going to use certain compiler args with nvcc and not with
#        # gcc the implementation of this trick is in customize_compiler() below
#        extra_compile_args={'gcc': ["-Wno-unused-function"],
#                            'nvcc': ['-arch=sm_35',
#                                     '--ptxas-options=-v',
#                                     '-c',
#                                     '--compiler-options',
#                                     "'-fPIC'"]},
#        include_dirs = [numpy_include, CUDA['include']]
#    ),
。。。

编译Cython命令：

~/py-faster-rcnn/lib$ make

注释：Cython可以把python代码编译成C代码，再用GCC编译，可以大幅度提高Python代码的执行速度。

修改py-faster-rcnn/caffe-fast-rcnn/Makefile.config，如下：

CPU_ONLY := 1 #将CPU_ONLY开关打开

BLAS_INCLUDE := /usr/include/atlas
BLAS_LIB := /usr/lib/atlas-base

PYTHON_INCLUDE := /usr/include/python2.7 \
                  /usr/lib/python2.7/site-packages/numpy/core/include

PYTHON_LIB := /usr/lib

WITH_PYTHON_LAYER := 1 #将WITH_PYTHON_LAYER开关打开

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib

修改py-faster-rcnn/caffe-fast-rcnn/Makefile，如下：

369行: LIBRARIES += satlas tatlas #新版atlas已经不用cblas atlas这两个lib了

第181行: LIBRARIES后面增加opencv_core opencv_highgui opencv_imgproc opencv_imgcodes

编译Caffe和Pycaffe命令：

~/py-faster-rcnn/caffe-fast-rcnn$ make -j8 && make pycaffe

还需修改相关配置，让其模型可以在CPU环境下运行
修改py-faster-rcnn/lib/fast_rcnn/config.py
__C.USE_GPU_NMS = False

修改py-faster-rcnn/tools/test_net.py和 py-faster-rcnn/tools/train_net.py

caffe.set_mode_gpu()修改为caffe.set_mode_cpu()

修改py-faster-rcnn/lib/setup.py

将含有'nms.gpu_nms’的部分注释掉

修改py-faster-rcnn/lib/fast_rcnn/nms_wrapper.py

#from nms.gpu_nms import gpu_nms

跑demo测试下
先下载作者训练好的模型：faster_rcnn_models，
然后命令：~/py-faster-rcnn$ python ./tools/demo.py --cpu

解决安装过程过出现的各种错误（我们的原则：遇山修路遇水搭桥）

问题：ImportError: No module named easydict
解决：sudo pip install easydict

问题：ImportError: No module named cv2
解决：sudo apt-get install python-opencv

问题：ImportError:No module named skimage.io
解决：sudo pip install scikit-image