03的开发日志
0%

Windows下编译OpenCV-3.1.0 + 扩展opencv_contrib

发表于 更新于 分类于
本文字数： 11k 阅读时长 ≈ 10 分钟
  1. 准备
  2. 配置编译生成VS工程文件
    1. 编译器的选择：VS2013 or VS2015？
    2. 第一个错误：FFMpeg下载失败
    3. 第二个错误：ippicv下载失败
  3. 引入opencv_contrib
  4. 使用VS生成dll和lib
    1. 第一个错误：没有生成“object”文件
    2. 可能出现的一个错误：缺少python27_d.lib
  5. 整合dll和lib

准备

使用OpenCV 3.1.0进行图像处理时，需要用到surf、sift等匹配算法，在官网下载的安装包解压得到的工具库中，不包含这些算法库（版权还是别的原因），需要自己编译。如果使用2.x版本，这些算法库包含在nonfree中，在3.0之后的版本中，这些算法库包含在xfeature2d中。

首先下载OpenCV-3.1.0的源代码，官网的下载页面： https://opencv.org/releases/page/4/ ，或者直接点击链接下载： https://github.com/opencv/opencv/archive/3.1.0.zip ，仓库地址：https://github.com/opencv/opencv/releases/tag/3.1.0

image-20200613224750516

然后我们需要下载OpenCV的额外模块（Extra Modules），仓库地址：https://github.com/opencv/opencv_contrib ，opencv_contrib的版本需要适配OpenCV的版本，在release页面（https://github.com/opencv/opencv_contrib/releases?after=3.4.1 ）找到3.1.0（https://github.com/opencv/opencv_contrib/releases/tag/3.1.0 ），下载源代码

image-20200613230931992

将下载的两个安装包解压到当前文件夹

image-20200613231301479

配置编译生成VS工程文件

编译器的选择：VS2013 or VS2015？

编译需要用到CMake（下载地址：https://cmake.org/download/ ），同时要求电脑上装有VS或Qt（作为编译器使用）。打开CMake，把opencv-3.1.0里边的CMakeLists.txt拖到CMake中，然后把构建目录的文件夹改为opencv-3.1.0-build，如图：

image-20200613231734871

点击Configure进行编译（提示创建文件夹，选“是”），编译器选择目标IDE（如VS2013、VS2015）。

image-20200613231925721

这影响的是最终生成的dll适配的平台版本，如果选择Visual Studio 12 2013，最终在install文件夹中将会生成vc12下的dll和lib，这个版本对应VS工程的平台工具集（vc12对应v120，vc14对应v140，可以在工程的属性页查看并配置）。

image-20200613232130239

image-20200613232325205

第一个错误：FFMpeg下载失败

解决参考：https://blog.csdn.net/yiyuehuan/article/details/52951574

不出意外的话此时会报错：Error in configuration process, project files may be invalid.

image-20200613232721767

检查错误发现：

image-20200613233023105

错误原因：FFMpeg下载失败，有可能是下载链接失效或者无法访问到远程仓库地址（编译旧版本C++库的时候可能存在类似问题），我们首先定位到第三方库的下载配置处：opencv-3.1.0\3rdparty\ffmpeg，打开ffmpeg.cmake

image-20200613233557849

我们先打开第三方库的远程仓库链接（https://github.com/opencv/opencv_3rdparty ），找到分支master_20150703（https://github.com/opencv/opencv_3rdparty/tree/ffmpeg/master_20150703 ），把需要的三个文件手动下载下来，并且将ffmpeg.cmake中的三条下载命令注释掉

image-20200613233920721

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
# Binary branch name: ffmpeg/master_20150703
# Binaries were created for OpenCV: e379ea6ed60b0caad4d4e3eea096e9d850cb8c86
set(FFMPEG_BINARIES_COMMIT "8aeefc4efe3215de89d8c7e114ae6f7a6091b8eb")
set(FFMPEG_FILE_HASH_BIN32 "89c783eee1c47bfc733f08334ec2e31c")
set(FFMPEG_FILE_HASH_BIN64 "35fe6ccdda6d7a04e9056b0d73b98e76")
set(FFMPEG_FILE_HASH_CMAKE "8606f947a780071f8fcce8cbf39ceef5")

set(FFMPEG_DOWNLOAD_URL ${OPENCV_FFMPEG_URL};$ENV{OPENCV_FFMPEG_URL};https://raw.githubusercontent.com/Itseez/opencv_3rdparty/${FFMPEG_BINARIES_COMMIT}/ffmpeg/)

# ocv_download(PACKAGE opencv_ffmpeg.dll
             # HASH ${FFMPEG_FILE_HASH_BIN32}
             # URL ${FFMPEG_DOWNLOAD_URL}
             # DESTINATION_DIR ${CMAKE_CURRENT_LIST_DIR})

# ocv_download(PACKAGE opencv_ffmpeg_64.dll
             # HASH ${FFMPEG_FILE_HASH_BIN64}
             # URL ${FFMPEG_DOWNLOAD_URL}
             # DESTINATION_DIR ${CMAKE_CURRENT_LIST_DIR})

# ocv_download(PACKAGE ffmpeg_version.cmake
             # HASH ${FFMPEG_FILE_HASH_CMAKE}
             # URL ${FFMPEG_DOWNLOAD_URL}
             # DESTINATION_DIR ${CMAKE_CURRENT_LIST_DIR})

include(${CMAKE_CURRENT_LIST_DIR}/ffmpeg_version.cmake)

下载完成并注释后，重新打开CMake，点击Configure

第二个错误：ippicv下载失败

第二个错误出现了：

image-20200613234117809

同样的方法，定位到opencv-3.1.0\3rdparty\ippicv，打开downloader.cmake，查看我们需要手动下载的文件。文件中的意思是，当编译的平台为WIN32且不为ARM时，获取包名为ippicv_windows_20151201.zip的文件，包中的子目录名为ippicv_win；ICV的解压目录为/unpack。

image-20200613235211128

来到远程仓库（https://github.com/opencv/opencv_3rdparty/tree/ippicv/master_20151201 ），在这里面发现了我们需要的文件：

image-20200613235057887

我们在downloader.cmake的同级目录下创建文件夹unpack（cmake文件中指定的解压目录），将下载的ippicv_windows_20151201.zip下载到这个文件夹中，然后解压到当前文件夹（根据cmake文件中的语句进行操作），得到如下目录结构：

image-20200614000241805

接着回头把downloader.cmake中的下载语句注释掉，注释后的文件文本内容为：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
#
# The script downloads ICV package
#
# On return this will define:
# OPENCV_ICV_PATH - path to unpacked downloaded package
#

function(_icv_downloader)
  # Commit SHA in the opencv_3rdparty repo
  set(IPPICV_BINARIES_COMMIT "81a676001ca8075ada498583e4166079e5744668")
  # Define actual ICV versions
  if(APPLE)
    set(OPENCV_ICV_PACKAGE_NAME "ippicv_macosx_20151201.tgz")
    set(OPENCV_ICV_PACKAGE_HASH "4ff1fde9a7cfdfe7250bfcd8334e0f2f")
    set(OPENCV_ICV_PLATFORM "macosx")
    set(OPENCV_ICV_PACKAGE_SUBDIR "/ippicv_osx")
  elseif(UNIX)
    if(ANDROID AND NOT (ANDROID_ABI STREQUAL x86 OR ANDROID_ABI STREQUAL x86_64))
      return()
    endif()
    set(OPENCV_ICV_PACKAGE_NAME "ippicv_linux_20151201.tgz")
    set(OPENCV_ICV_PACKAGE_HASH "808b791a6eac9ed78d32a7666804320e")
    set(OPENCV_ICV_PLATFORM "linux")
    set(OPENCV_ICV_PACKAGE_SUBDIR "/ippicv_lnx")
  elseif(WIN32 AND NOT ARM)
    set(OPENCV_ICV_PACKAGE_NAME "ippicv_windows_20151201.zip")
    set(OPENCV_ICV_PACKAGE_HASH "04e81ce5d0e329c3fbc606ae32cad44d")
    set(OPENCV_ICV_PLATFORM "windows")
    set(OPENCV_ICV_PACKAGE_SUBDIR "/ippicv_win")
  else()
    return() # Not supported
  endif()

  set(OPENCV_ICV_UNPACK_PATH "${CMAKE_CURRENT_LIST_DIR}/unpack")
  set(OPENCV_ICV_PATH "${OPENCV_ICV_UNPACK_PATH}${OPENCV_ICV_PACKAGE_SUBDIR}")

  # if(DEFINED OPENCV_ICV_PACKAGE_DOWNLOADED
       # AND OPENCV_ICV_PACKAGE_DOWNLOADED STREQUAL OPENCV_ICV_PACKAGE_HASH
       # AND EXISTS ${OPENCV_ICV_PATH})
    # # Package has been downloaded and checked by the previous build
    # set(OPENCV_ICV_PATH "${OPENCV_ICV_PATH}" PARENT_SCOPE)
    # return()
  # else()
    # if(EXISTS ${OPENCV_ICV_UNPACK_PATH})
      # message(STATUS "ICV: Removing previous unpacked package: ${OPENCV_ICV_UNPACK_PATH}")
      # file(REMOVE_RECURSE ${OPENCV_ICV_UNPACK_PATH})
    # endif()
  # endif()
  # unset(OPENCV_ICV_PACKAGE_DOWNLOADED CACHE)

  # set(OPENCV_ICV_PACKAGE_ARCHIVE "${CMAKE_CURRENT_LIST_DIR}/downloads/${OPENCV_ICV_PLATFORM}-${OPENCV_ICV_PACKAGE_HASH}/${OPENCV_ICV_PACKAGE_NAME}")
  # get_filename_component(OPENCV_ICV_PACKAGE_ARCHIVE_DIR "${OPENCV_ICV_PACKAGE_ARCHIVE}" PATH)
  # if(EXISTS "${OPENCV_ICV_PACKAGE_ARCHIVE}")
    # file(MD5 "${OPENCV_ICV_PACKAGE_ARCHIVE}" archive_md5)
    # if(NOT archive_md5 STREQUAL OPENCV_ICV_PACKAGE_HASH)
      # message(WARNING "ICV: Local copy of ICV package has invalid MD5 hash: ${archive_md5} (expected: ${OPENCV_ICV_PACKAGE_HASH})")
      # file(REMOVE "${OPENCV_ICV_PACKAGE_ARCHIVE}")
      # file(REMOVE_RECURSE "${OPENCV_ICV_PACKAGE_ARCHIVE_DIR}")
    # endif()
  # endif()

  # if(NOT EXISTS "${OPENCV_ICV_PACKAGE_ARCHIVE}")
    # if(NOT DEFINED OPENCV_ICV_URL)
      # if(DEFINED ENV{OPENCV_ICV_URL})
        # set(OPENCV_ICV_URL $ENV{OPENCV_ICV_URL})
      # else()
        # set(OPENCV_ICV_URL "https://raw.githubusercontent.com/Itseez/opencv_3rdparty/${IPPICV_BINARIES_COMMIT}/ippicv")
      # endif()
    # endif()

    # file(MAKE_DIRECTORY ${OPENCV_ICV_PACKAGE_ARCHIVE_DIR})
    # message(STATUS "ICV: Downloading ${OPENCV_ICV_PACKAGE_NAME}...")
    # file(DOWNLOAD "${OPENCV_ICV_URL}/${OPENCV_ICV_PACKAGE_NAME}" "${OPENCV_ICV_PACKAGE_ARCHIVE}"
         # TIMEOUT 600 STATUS __status
         # EXPECTED_MD5 ${OPENCV_ICV_PACKAGE_HASH})
    # if(NOT __status EQUAL 0)
      # message(FATAL_ERROR "ICV: Failed to download ICV package: ${OPENCV_ICV_PACKAGE_NAME}. Status=${__status}")
    # else()
      # # Don't remove this code, because EXPECTED_MD5 parameter doesn't fail "file(DOWNLOAD)" step
      # # on wrong hash
      # file(MD5 "${OPENCV_ICV_PACKAGE_ARCHIVE}" archive_md5)
      # if(NOT archive_md5 STREQUAL OPENCV_ICV_PACKAGE_HASH)
        # message(FATAL_ERROR "ICV: Downloaded copy of ICV package has invalid MD5 hash: ${archive_md5} (expected: ${OPENCV_ICV_PACKAGE_HASH})")
      # endif()
    # endif()
  # endif()

  # ocv_assert(EXISTS "${OPENCV_ICV_PACKAGE_ARCHIVE}")
  # ocv_assert(NOT EXISTS "${OPENCV_ICV_UNPACK_PATH}")
  # file(MAKE_DIRECTORY ${OPENCV_ICV_UNPACK_PATH})
  # ocv_assert(EXISTS "${OPENCV_ICV_UNPACK_PATH}")

  # message(STATUS "ICV: Unpacking ${OPENCV_ICV_PACKAGE_NAME} to ${OPENCV_ICV_UNPACK_PATH}...")
  # execute_process(COMMAND ${CMAKE_COMMAND} -E tar xz "${OPENCV_ICV_PACKAGE_ARCHIVE}"
                  # WORKING_DIRECTORY "${OPENCV_ICV_UNPACK_PATH}"
                  # RESULT_VARIABLE __result)

  # if(NOT __result EQUAL 0)
    # message(FATAL_ERROR "ICV: Failed to unpack ICV package from ${OPENCV_ICV_PACKAGE_ARCHIVE} to ${OPENCV_ICV_UNPACK_PATH} with error ${__result}")
  # endif()

  # ocv_assert(EXISTS "${OPENCV_ICV_PATH}")

  # set(OPENCV_ICV_PACKAGE_DOWNLOADED "${OPENCV_ICV_PACKAGE_HASH}" CACHE INTERNAL "ICV package hash")

  message(STATUS "ICV: Package successfully downloaded")
  set(OPENCV_ICV_PATH "${OPENCV_ICV_PATH}" PARENT_SCOPE)
endfunction()

_icv_downloader()

完成后打开CMake，点击Configure，不出意外的话此时不会报错了

引入opencv_contrib

找到变量OPENCV_EXTRA_MODULES_PATH，把opencv_contrib-3.1.0的绝对路径填进去，然后点击Generate，就开始生成了

image-20200614001554567

如果这边opencv_contrib的版本和opencv的版本不一样，可能会产生奇怪的错误。解决的最好办法就是版本适配，否则极有可能瞎折腾。生成结束后，CMake中提示如下：

image-20200614002630708

使用VS生成dll和lib

打开opencv-3.1.0-build，使用VS2013（与编译器对应）打开里边的OpenCV.sln文件，在“生成-批生成”中先把ALL_BUILD的两个勾打上，点击“生成”

image-20200614015333768

待其结束后，勾选INSTALL后的两个勾，进行生成

image-20200614011313362

第一个错误：没有生成“object”文件

错误代码：error C2220

提示信息：error C2220: 警告被视为错误 - 没有生成“object”文件

解决方法：把build文件夹清空，打开CMake重新点击Generate，再进行生成

可能出现的一个错误：缺少python27_d.lib

我进行了两次编译，第一次使用opencv-3.1.0.exe解压的source作为源码进行编译，第二次直接下载opencv-3.1.0.zip的源码进行编译。在第一次编译中，遇到一个问题，生成时提示缺少python27_d.lib文件，报错，后来通过修改工程文件解决了这个问题。但是在直接使用源码编译时，并没有出现这个错误。下面是解决这个错误的方法记录。

错误 24 error LNK1104: 无法打开文件“python27_d.lib”

参考：https://blog.csdn.net/pplxlee/article/details/78067927

在OpenCV项目的bindings中的opencv_python项目中，点击外部依赖项，在其中找到pyconfig.h和object.h文件，修改其中对python27_d.dll的使用

image-20200614034117774

在pyconfig.h中

将

1
2
3
4
5
#ifdef _DEBUG

#	define Py_DEBUG

#endif

修改为：

1
2
3
#ifdef _DEBUG
//#	define Py_DEBUG
#endif

将

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
/* For an MSVC DLL, we can nominate the .lib files used by extensions */
#ifdef MS_COREDLL
#	ifndef Py_BUILD_CORE /* not building the core - must be an ext */
#		if defined(_MSC_VER)
			/* So MSVC users need not specify the .lib file in
			their Makefile (other compilers are generally
			taken care of by distutils.) */
#			ifdef _DEBUG
#				pragma comment(lib,"python27_d.lib")
#			else
#				pragma comment(lib,"python27.lib")
#			endif /* _DEBUG */
#		endif /* _MSC_VER */
#	endif /* Py_BUILD_CORE */
#endif /* MS_COREDLL */

修改为

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
/* For an MSVC DLL, we can nominate the .lib files used by extensions */
#ifdef MS_COREDLL
#	ifndef Py_BUILD_CORE /* not building the core - must be an ext */
#		if defined(_MSC_VER)
			/* So MSVC users need not specify the .lib file in
			their Makefile (other compilers are generally
			taken care of by distutils.) */
#			ifdef _DEBUG
#				pragma comment(lib,"python27.lib")
#			else
#				pragma comment(lib,"python27.lib")
#			endif /* _DEBUG */
#		endif /* _MSC_VER */
#	endif /* Py_BUILD_CORE */
#endif /* MS_COREDLL */

在object.h中

将

1
2
3
4
/* Py_DEBUG implies Py_TRACE_REFS. */
#if defined(Py_DEBUG) && !defined(Py_TRACE_REFS)
#define Py_TRACE_REFS
#endif

修改为

1
2
3
4
/* Py_DEBUG implies Py_TRACE_REFS. */
#if defined(Py_DEBUG) && !defined(Py_TRACE_REFS)
//#define Py_TRACE_REFS
#endif

整合dll和lib

生成成功后，在opencv-3.1.0-build目录下生成了一个install文件夹，里边的include和x86文件夹是我们最终需要的东西

image-20200614035035123

首先过河拆桥，把3.1.0文件夹重命名为3.1.0-，然后新建3.1.0的空文件夹，把include和x86拷贝到新的3.1.0文件夹下

image-20200614035846255

接着需要区分一下debug和release版本，打开x86/vc12，把staticlib删掉，把lib重命名为lib-，并新建dll和lib文件夹

image-20200614040107577

首先进入dll，新建两个文件夹：debug和release，把bin文件夹中所有以310d.dll结尾的dll文件全部移动到debug文件夹下，以310.dll结尾的dll文件全部移动到release文件夹下。然后把bin删掉。

接着在新建的lib文件夹中同样创建两个文件夹：debug和release，把lib-中所有以310d.lib结尾的lib文件移动到debug文件夹下，……（操作同dll）。然后删掉lib-文件夹。

这样三件套就集齐了，可以把整个3.1.0-文件夹删掉了。

image-20200614040909636

还蛮占空间的

image-20200614041011136