ROS中的libtiff4错误

Question

在Ubuntu14.04下，我试图从一个有catkin的朋友那里编译一个ROS-包，并得到以下错误：

/usr/bin/ld: warning: libboost_system.so.1.49.0, needed by   
//usr/local/MATLAB/R2014a/bin/glnxa64/libut.so, may conflict with libboost_system.so.1.54.0
//usr/local/lib/libcvd.so: undefined reference to `TIFFWriteEncodedStrip@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFReadRGBAImageOriented@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFWriteScanline@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFGetField@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFClose@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFClientOpen@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFGetFieldDefaulted@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFStripSize@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFSetField@LIBTIFF_4.0'
//usr/lib/x86_64-linux-gnu/libharfbuzz.so.0: undefined reference to `FT_Face_GetCharVariantIndex'
//usr/local/lib/libcvd.so: undefined reference to `TIFFSetErrorHandler@LIBTIFF_4.0'
//usr/local/lib/libcvd.so: undefined reference to `TIFFReadScanline@LIBTIFF_4.0'
//usr/lib/x86_64-linux-gnu/libharfbuzz.so.0: undefined reference to `FT_Get_Advance'
collect2: error: ld returned 1 exit status

我已经安装了libcvd和libtiff4-dev。有谁知道如何解决这个问题吗？

非常感谢,

下雪

编辑:正如建议的那样，我包括了CMakeLists.txt：

cmake_minimum_required(VERSION 2.8.3)
project(test)
set (test_VERSION "0.0.1")


find_package(OpenCV REQUIRED)

find_package(catkin REQUIRED COMPONENTS
  test_core
  cv_bridge
  image_transport
  roscpp
)

find_package(tracker)

set (CMAKE_CXX_FLAGS "-DNDEBUG -DNTIMING -DNRUN_UNIT_TESTS -g -O0 -std=c++11")

catkin_package(
  INCLUDE_DIRS include
)


include_directories (include
  ${CMAKE_CURRENT_SOURCE_DIR}/include
  ${tracker_INCLUDE_DIRS}
  ${TRIANGULATION_INCLUDE_DIRS}
  ${OPENCV_INCLUDE_DIRS}
)

include_directories(/usr/local/MATLAB/R2014a/extern/include)

include_directories (SYSTEM
  ${catkin_INCLUDE_DIRS}
)

set (SOURCE
  src/test/main.cc
  src/test/rosbridge.cc
  src/test/core.cc
)

add_executable (test ${SOURCE})

target_link_libraries(test
  /lib/x86_64-linux-gnu/libssl.so.1.0.0
  /lib/x86_64-linux-gnu/libcrypt.so.1
  /lib/x86_64-linux-gnu/libcrypto.so.1.0.0
  /usr/local/MATLAB/R2014a/bin/glnxa64/libmx.so
  /usr/local/MATLAB/R2014a/bin/glnxa64/libeng.so
  /usr/local/MATLAB/R2014a/bin/glnxa64/libmat.so
  /usr/local/MATLAB/R2014a/bin/glnxa64/libut.so
  ${OpenCV_LIBS}
  ${tracker_LIBRARIES}
  cvd
  ${catkin_LIBRARIES}
  ${TRIANGULATION_LIBRARIES}
)

Answer 1

我修好了！

您必须链接到lib文件夹中的libtiff lib，如下所示：

target_link_libraries(test
  .
  .
  .
  /usr/lib/x86_64-linux-gnu/libtiff.so.5  
  .
  .
  .
)

Answer 2

在您的情况下，仅仅链接cvd似乎不起作用。不过，CMake附带了一个很棒的find_package特性，所以让我们使用它：

1. 在文件顶部添加find_package(CVD REQUIRED)
2. 将${CVD_INCLUDE_DIRS}添加到include_directories
3. 用cvd替换target_link_libraries中的${CVD_LIBRARIES}

这可能不会立即起作用，但会抛出一个类似于"FindCVD.cmake not“之类的错误。这是一个脚本，在您的文件系统中搜索这个库的实际位置，并存储到上面使用的变量的路径。许多库已经自己带来了这样的文件，但是如果不是这样，您必须手动提供它。但是，在大多数情况下，您不必自己编写这个文件，因为通常有许多开源项目已经创建了这样一个文件，您可以重用这个文件(例如这里)。只要谷歌"FindCVD.cmake“找到他们。

一旦你有了这个文件：

1. 在项目中创建一个名为"cmake“的新子目录，并将文件存储在那里。
2. 将set(CMAKE_MODULE_PATH ${PROJECT_SOURCE_DIR}/cmake)添加到CMakeLists.txt的顶部(在find_package调用之前！)

现在希望它能起作用:)