如何为Flutter Windows应用集成原生OpenCV库？附动态库创建指南

准备工作 #

• 下载安装OpenCV Windows版（推荐选择包含opencv_worldxxx.dll的发行版，例如4.8.0），记录安装路径（如C:\opencv）
• 确保Flutter已配置Windows开发环境（执行flutter doctor验证）

步骤1：创建OpenCV封装动态库 #

1.1 新建原生项目结构 #

创建名为opencv_native的文件夹，添加以下两个文件：

opencv_wrapper.cpp（封装OpenCV接口）

#include <opencv2/core/version.hpp>
#include <cstdlib>
#include <cstring>

// 使用extern "C"避免C++名字修饰，确保Dart能正确找到函数
extern "C" {
    // 返回OpenCV版本字符串，内存由malloc分配，需调用free_memory释放
    const char* get_opencv_version() {
        const char* version_str = CV_VERSION;
        char* allocated_str = (char*)malloc(strlen(version_str) + 1);
        strcpy(allocated_str, version_str);
        return allocated_str;
    }

    // 供Dart调用释放内存的函数
    void free_memory(char* ptr) {
        free(ptr);
    }
}

CMakeLists.txt（编译配置）

cmake_minimum_required(VERSION 3.10)
project(opencv_native)

# 指定C++标准
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)

# 替换为你的OpenCV安装路径
set(OpenCV_DIR "C:/opencv/build")
find_package(OpenCV REQUIRED)

# 生成动态库
add_library(opencv_native SHARED opencv_wrapper.cpp)

# 链接OpenCV依赖库
target_link_libraries(opencv_native ${OpenCV_LIBS})

# 设置动态库输出目录，方便后续拷贝
set_target_properties(opencv_native PROPERTIES
    RUNTIME_OUTPUT_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/output
)

1.2 编译动态库 #

1. 打开CMake GUI，选择opencv_native为源代码目录，新建build文件夹作为编译目录
2. 点击Configure，选择对应版本的Visual Studio（如2022 x64），完成配置
3. 点击Generate，然后选择Open Project打开Visual Studio
4. 选择Release模式编译项目，编译完成后，output/Release目录会生成opencv_native.dll
5. 从OpenCV安装目录（如C:\opencv\build\x64\vc16\bin）拷贝对应版本的opencv_worldxxx.dll到output/Release目录

步骤2：Flutter项目集成与FFI调用 #

2.1 拷贝动态库到Flutter项目 #

将opencv_native.dll和opencv_worldxxx.dll拷贝到Flutter项目的windows/runner/Release目录（Debug模式则拷贝到windows/runner/Debug，需对应使用OpenCV的Debug版库）

2.2 编写Dart FFI代码 #

在lib目录下新建opencv_ffi.dart：

import 'dart:ffi';
import 'dart:io';
import 'package:ffi/ffi.dart';

// 加载动态库
final DynamicLibrary _opencvLib = Platform.isWindows
    ? DynamicLibrary.open('opencv_native.dll')
    : throw UnsupportedError('仅支持Windows平台');

// 定义函数签名
typedef _GetVersionFunc = Pointer<Utf8> Function();
typedef _GetVersion = Pointer<Utf8> Function();

typedef _FreeMemoryFunc = Void Function(Pointer<Utf8>);
typedef _FreeMemory = void Function(Pointer<Utf8>);

// 获取函数指针
final _getVersion = _opencvLib.lookupFunction<_GetVersionFunc, _GetVersion>('get_opencv_version');
final _freeMemory = _opencvLib.lookupFunction<_FreeMemoryFunc, _FreeMemory>('free_memory');

// 封装调用方法，自动处理内存释放
String getOpenCVVersion() {
  final versionPtr = _getVersion();
  final version = versionPtr.toDartString();
  _freeMemory(versionPtr); // 释放C层分配的内存
  return version;
}

2.3 在Flutter界面中调用 #

修改lib/main.dart：

import 'package:flutter/material.dart';
import 'opencv_ffi.dart';

void main() => runApp(const MyApp());

class MyApp extends StatelessWidget {
  const MyApp({super.key});

  @override
  Widget build(BuildContext context) {
    return MaterialApp(
      home: Scaffold(
        appBar: AppBar(title: const Text('OpenCV版本查询')),
        body: Center(
          child: FutureBuilder<String>(
            future: Future.value(getOpenCVVersion()),
            builder: (context, snapshot) {
              if (snapshot.hasData) {
                return Text('OpenCV版本：${snapshot.data}');
              } else if (snapshot.hasError) {
                return Text('错误：${snapshot.error}');
              }
              return const CircularProgressIndicator();
            },
          ),
        ),
      ),
    );
  }
}

步骤3：运行项目 #

执行以下命令启动Flutter Windows应用：

flutter run -d windows

界面将显示当前集成的OpenCV版本号。

内容的提问来源于stack exchange，提问作者Ibne