mac中加载C++动态库文件

前言

需要再mac系统下运行C++开发的程序，通过摸索，初步实现了一版，大致记录下

1. 前提准备

1. 安装OpenCV

• 使用Homebrew安装OpenCV：

brew install opencv

• 确认安装路径：

brew --prefix opencv

默认路径为/opt/homebrew/opt/opencv。

• 确保头文件和动态库路径正确，以便在后续编译中正确链接。

2. 将C++程序编译成动态库

2.1 编写C++代码

• 编写基于OpenCV的C++代码，通过C++类实现图像处理过程，通过
  链接函数。
• 示例代码：头文件

#ifndef IMAGE_PROCESS_H
#define IMAGE_PROCESS_H#include <opencv2/opencv.hpp>class ImageProcess
{
public:cv::Mat getImageProcess(const cv::Mat &image);
};#ifdef __cplusplus
extern "C"
{
#endif// 声明一个 C 链接的函数，返回原始指针extern "C" unsigned char *processImage(unsigned char *input, int width, int height, int bytesPerRow);#ifdef __cplusplus
}
#endif#endif

#include "image.h"using namespace cv;
cv::Mat ImageProcess::getImageProcess(const cv::Mat &image)
{// 获取图像的宽度和高度int width = image.cols;int height = image.rows;// 计算矩形的左上角和右下角坐标int rectWidth = 150;  // 矩形的宽度int rectHeight = 50;  // 矩形的高度int startX = (width - rectWidth) / 2;  // 矩形的左上角 x 坐标int startY = (height - rectHeight) / 2;  // 矩形的左上角 y 坐标int endX = startX + rectWidth;  // 矩形的右下角 x 坐标int endY = startY + rectHeight;  // 矩形的右下角 y 坐标// 绘制矩形rectangle(image, Point(startX, startY), Point(endX, endY), Scalar(0, 255, 0), 4);  // 使用绿色绘制矩形，线条宽度为 2return image;
}extern "C" unsigned char *processImage(unsigned char *input, int width, int height, int bytesPerRow)
{cv::Mat inputMat(height, width, CV_8UC4, input, bytesPerRow);ImageProcess imageProcess;cv::Mat outputMat = imageProcess.getImageProcess(inputMat);// 将输出 Mat 的数据复制到一个新的缓冲区unsigned char *outputData = new unsigned char[outputMat.total() * outputMat.elemSize()];std::memcpy(outputData, outputMat.data, outputMat.total() * outputMat.elemSize());return outputData;
}extern "C" int add(int a, int b)
{return a + b;
}

• 注意：使用extern "C"声明函数，避免C++的名称修饰问题，确保从Swift或其他语言调用时能够正确链接。

2.2 编译动态库

• 编写CMakeLists.txt文件（推荐）或直接使用g++命令编译。
• 示例CMakeLists.txt：

cmake_minimum_required(VERSION 3.10)
project(ImageLib)# 设置 C++ 标准
set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_STANDARD_REQUIRED True)# 设置头文件路径
include_directories(${PROJECT_SOURCE_DIR}/include)# 添加动态库
add_library(Image SHARED src/image.cpp)# 找到 OpenCV 库
find_package(OpenCV REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})# 链接 OpenCV 库
target_link_libraries(Image ${OpenCV_LIBS})

• 编译命令：通过执行 sh build.sh编译

#!/bin/bash# 设置项目根目录
PROJECT_DIR=$(pwd)
echo "Project directory: $PROJECT_DIR"# 创建构建目录
BUILD_DIR="${PROJECT_DIR}/build"
echo "Creating build directory: $BUILD_DIR"
mkdir -p "${BUILD_DIR}"# 清空构建目录
if [ -d "${BUILD_DIR}" ]; thenecho "Cleaning previous build directory..."rm -rf "${BUILD_DIR}"/*
elseecho "Build directory does not exist. Creating it..."mkdir -p "${BUILD_DIR}"
fi# 检查构建目录是否创建成功
if [ ! -d "${BUILD_DIR}" ]; thenecho "Failed to create build directory: $BUILD_DIR"exit 1
fi# 进入构建目录
echo "Changing directory to: $BUILD_DIR"
cd "${BUILD_DIR}"# 运行 CMake
echo "Running CMake..."
cmake ..# 检查 CMake 是否成功
if [ $? -ne 0 ]; thenecho "CMake failed. Exiting..."exit 1
fi# 编译项目
echo "Compiling project..."
make -j$(sysctl -n hw.ncpu)# 检查编译是否成功
if [ $? -ne 0 ]; thenecho "Compilation failed. Exiting..."exit 1
fi# 检查动态库是否生成
if [ ! -f "${BUILD_DIR}/libImage.dylib" ]; thenecho "Dynamic library not found: ${BUILD_DIR}/libImage.dylib"exit 1
fi# 输出构建结果
echo "Build completed. Dynamic library is in ${BUILD_DIR}/libImage.dylib"

• 生成的动态库文件为libImage.dylib（在macOS上）。

3. 在Swift中调用动态库

1. 加载动态库

• 使用dlopen加载动态库，并通过dlsym获取函数地址。
• 示例代码：

image_processimport Foundation
import AppKitfunc main() {// 1. 动态库加载模块typealias ProcessImageFunction = @convention(c) (UnsafeMutablePointer<UInt8>?, Int32, Int32, Int32) -> UnsafeMutablePointer<UInt8>?// 加载动态库let handle = dlopen("/Users/gongyong/Desktop/Keyi/test_ws/image_process/build/libImage.dylib", RTLD_LAZY)if handle == nil {print("Failed to load library: \(String(cString: dlerror()))")exit(1)} else {print("Library loaded successfully")}// 获取函数地址let processImageFunctionPointer = dlsym(handle, "processImage")if processImageFunctionPointer == nil {print("Failed to find function: \(String(cString: dlerror()))")dlclose(handle)exit(1)} else {print("Function 'processImage' found successfully")}// 将指针转换为函数指针类型let processImageFunction = unsafeBitCast(processImageFunctionPointer, to: ProcessImageFunction.self)// 2. 加载图像，并且调用动态库处理let filePath = "/Users/gongyong/Desktop/Keyi/test_ws/demo/Source/demo/face.png"let outputPath = "/Users/gongyong/Desktop/Keyi/test_ws/demo/Source/demo/output.png"if let (pixelData, width, height, bytesPerRow) = loadPNGImage(from: filePath) {print("Image loaded successfully with dimensions: \(width) x \(height), bytesPerRow: \(bytesPerRow)")// 调用 C+ 图像处理函数let processedPixelData = processImageFunction(pixelData, Int32(width), Int32(height), Int32(bytesPerRow))if processedPixelData == nil {print("Image processing failed: processImage returned nil")} else {print("Image processed successfully")// 保存处理后的图像savePNGImage(to: outputPath, pixelData: processedPixelData, width: width, height: height, bytesPerRow: bytesPerRow)// 打印输出图像的长宽print("Output image dimensions: \(width) x \(height)")}// 释放动态分配的内存free(processedPixelData)} else {print("Failed to load image from \(filePath)")}// 关闭动态库dlclose(handle)print("Library closed successfully")
}// 加载 PNG 图像的辅助函数
func loadPNGImage(from path: String) -> (UnsafeMutablePointer<UInt8>?, Int, Int, Int)? {print("Loading image from \(path)")guard let image = NSImage(contentsOfFile: path) else {print("Failed to load image from \(path)")return nil}guard let tiffData = image.tiffRepresentation,let bitmapRep = NSBitmapImageRep(data: tiffData) else {print("Failed to create bitmap representation from image")return nil}let width = Int(bitmapRep.size.width)let height = Int(bitmapRep.size.height)let bytesPerRow = bitmapRep.bytesPerRowguard let pixelData = malloc(bytesPerRow * height)?.assumingMemoryBound(to: UInt8.self) else {print("Failed to allocate memory for pixel data")return nil}// Copy the bitmap data to the allocated memorymemcpy(pixelData, bitmapRep.bitmapData, bytesPerRow * height)print("Image data copied to memory")// 打印输入图像的长宽print("Input image dimensions: \(width) x \(height)")return (pixelData, width, height, bytesPerRow)
}// 保存 PNG 图像的辅助函数
func savePNGImage(to path: String, pixelData: UnsafeMutablePointer<UInt8>?, width: Int, height: Int, bytesPerRow: Int) {guard let pixelData = pixelData else {print("Failed to save image: pixelData is nil")return}var pixelDataPointer: UnsafeMutablePointer<UInt8>? = pixelData// For RGBA imagelet bitmapRep = NSBitmapImageRep(bitmapDataPlanes: &pixelDataPointer,pixelsWide: width,pixelsHigh: height,bitsPerSample: 8,samplesPerPixel: 4,hasAlpha: true,isPlanar: false,colorSpaceName: .calibratedRGB,bitmapFormat: .alphaFirst,bytesPerRow: bytesPerRow,bitsPerPixel: 32)guard let bitmapRep = bitmapRep else {print("Failed to create NSBitmapImageRep")return}guard let pngData = bitmapRep.representation(using: .png, properties: [:]) else {print("Failed to generate PNG data")return}do {try pngData.write(to: URL(fileURLWithPath: path))print("Image saved successfully to \(path)")} catch {print("Failed to save image: \(error)")}
}main()

• 注意：确保动态库路径正确，且动态库中的函数签名与Swift中声明的类型一致。

2. 加载和处理图像

• 使用Swift的NSImage和NSBitmapImageRep加载和保存图像。
• 将图像数据传递给C++动态库进行处理，并接收处理后的数据。

4. 注意事项

1. 动态库路径

• 确保动态库路径正确，可通过dlopen加载指定路径的动态库。
• 如果动态库路径不正确，会报错“Failed to load library”。

2. 函数签名一致性

• C++动态库中的函数签名必须与Swift中声明的函数指针类型一致，否则会报错“Failed to find function”或运行时崩溃。

3. 内存管理

• 动态分配的内存需要手动释放，避免内存泄漏。
• 在Swift中使用malloc分配内存，并在处理完成后使用free释放。

4. 图像格式

• 确保输入和输出图像格式一致，例如RGBA或灰度图像。
• 如果格式不一致，可能导致图像处理失败或输出图像异常。

5. 调试

• 使用dlerror获取动态库加载和函数查找的错误信息，便于调试。
• 在开发过程中，建议逐步测试每个模块（如动态库加载、图像加载、图像处理、图像保存等），确保每个部分正常工作。

5. 效果演示

加载图像

调用C++处理后的图像

附件

先编译完image_process下的C++文件
然后xcode打开demo解压文件，更新动态库路径、加载图片和输出图片的绝对路径。

文件下载地址：

git clone [e-mail]:xjturmy/xcode_c-.git