diff --git a/README_OHOS_zh.md b/README_OHOS_zh.md new file mode 100755 index 0000000000000000000000000000000000000000..221b2ab9fa171d8b002474f2cce82a265193b12e --- /dev/null +++ b/README_OHOS_zh.md @@ -0,0 +1,228 @@ +OpenCV +============== + +原始仓来源：https://github.com/opencv/opencv/tree/4.5.5 + +## 仓库说明 +本仓库是基于opencv官方仓库4.5.5分支进行了OpenHarmony编译适配和拓展，可构建支持在OpenHarmony设备上运行的计算机视觉程序。 + +## 构建说明 + +### 嵌入OHOS中构建 + +* 构建环境： +1. 目前支持在Linux ubuntu20.4及以上版本随OpenHarmony3.2版本一起用GN编译，也可以用OpenCV自带cmake脚本交叉编译可在OpenHarmony系统调用的动态库; +2. 请确保当前构建环境可以编译OpenHarmony。 + +* 构建步骤： +1. 构建基础环境：可参照[OpenHarmony官网](https://www.openharmony.cn/download)； + + 需要安装的基础软件或库，如果构建环境已安装下面软件可忽略该步骤： + - git客户端和git-lfs + - 码云repo工具，安装命令：curl https://gitee.com/oschina/repo/raw/fork_flow/repo-py3 -o /usr/local/bin/repo; chmod a+x /usr/local/bin/repo; pip3 install -i https://repo.huaweicloud.com/repository/pypi/simple reques + - 执行命令安装OpenHarmony编译依赖的软件：sudo apt-get update && sudo apt-get install binutils git git-lfs gnupg flex bison gperf build-essential zip curl zlib1g-dev gcc-multilib g++-multilib libc6-dev-i386 lib32ncurses5-dev x11proto-core-dev libx11-dev lib32z1-dev ccache libgl1-mesa-dev libxml2-utils xsltproc unzip m4 bc gnutls-bin python3-pip -y + - 执行如下命令：sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.8 1 + + +2. 下载OpenCV源代码： + + 进入OpenHarmony工程的third_party目录下（如没有OpenHarmony源代码，需先下载OpenHarmony源代码），执行如下命令下载OpenCV代码到third_party目录下。 + ``` + git clone https://gitee.com/openharmony-sig/third_party_opencv.git + mv third_party_opencv opencv + ``` + +3. 修改编译白名单文件，增加opencv的编译目标（OpenHarmony4.0及以上版本）：在OpenHarmony源代码的build/compile_standard_whitelist.json文件中，在"deps_added_external_part_module"和"third_deps_bundle_not_add"两个json数组中分别增加以下条目： + ``` + "//third_party/opencv/3rdparty/ade:opencv_ade_source", + "//third_party/opencv/3rdparty/carotene:opencv_carotene_source", + "//third_party/opencv/3rdparty/carotene:libtegra_hal", + "//third_party/opencv/3rdparty/ffmpeg/libavcodec:opencv_ffmpeg_avcodec", + "//third_party/opencv/3rdparty/ffmpeg/libavcodec:libopencv_avcodec", + "//third_party/opencv/3rdparty/ffmpeg/libavdevice:opencv_ffmpeg_avdevice", + "//third_party/opencv/3rdparty/ffmpeg/libavdevice:libopencv_avdevice", + "//third_party/opencv/3rdparty/ffmpeg/libavfilter:opencv_ffmpeg_avfilter", + "//third_party/opencv/3rdparty/ffmpeg/libavfilter:libopencv_avfilter", + "//third_party/opencv/3rdparty/ffmpeg/libavformat:opencv_ffmpeg_avformat", + "//third_party/opencv/3rdparty/ffmpeg/libavformat:libopencv_avformat", + "//third_party/opencv/3rdparty/ffmpeg/libswresample:opencv_ffmpeg_swresample", + "//third_party/opencv/3rdparty/ffmpeg/libswresample:libopencv_swresample", + "//third_party/opencv/3rdparty/ffmpeg/libswscale:opencv_ffmpeg_swscale", + "//third_party/opencv/3rdparty/ffmpeg/libswscale:libopencv_swscale", + "//third_party/opencv/3rdparty/ffmpeg/libavutil:opencv_ffmpeg_avutil", + "//third_party/opencv/3rdparty/ffmpeg/libavutil:libopencv_avutil", + "//third_party/opencv/3rdparty/libpng:opencv_libpng_source", + "//third_party/opencv/3rdparty/libpng:libopencv_png", + "//third_party/opencv/3rdparty/libtiff:opencv_libtiff_source", + "//third_party/opencv/3rdparty/libtiff:libopencv_tiff", + "//third_party/opencv/3rdparty/openjpeg:opencv_openjpeg_source", + "//third_party/opencv/3rdparty/openjpeg:libopencv_openjpeg", + "//third_party/opencv/3rdparty/libjpeg-turbo::opencv_libjpeg-turbo_source", + "//third_party/opencv/3rdparty/libjpeg-turbo:libjpeg-turbo", + "//third_party/opencv/3rdparty/zlib:libopencv_zlib", + "//third_party/opencv/3rdparty/zlib:opencv_zlib_source", + "//third_party/opencv/3rdparty/libwebp:libopencv_webp", + "//third_party/opencv/3rdparty/libwebp:opencv_libwebp_source", + "//third_party/opencv/3rdparty/ittnotify:opencv_ittnotify_source", + "//third_party/opencv/3rdparty/ittnotify:libopencv_ittnotify", + "//third_party/opencv/3rdparty/quirc:opencv_quirc_source", + "//third_party/opencv/3rdparty/quirc:libopencv_quirc", + "//third_party/opencv/modules/calib3d:opencv_calib3d_source", + "//third_party/opencv/3rdparty/protobuf:opencv_protobuf_source", + "//third_party/opencv/3rdparty/protobuf:libopencv_protobuf", + "//third_party/opencv/modules/calib3d:libopencv_calib3d", + "//third_party/opencv/modules/core:opencv_core_source", + "//third_party/opencv/modules/core:libopencv_core", + "//third_party/opencv/modules/dnn:opencv_dnn_source", + "//third_party/opencv/modules/dnn:libopencv_dnn", + "//third_party/opencv/modules/features2d:opencv_features2d_source", + "//third_party/opencv/modules/features2d:libopencv_features2d", + "//third_party/opencv/modules/flann:opencv_flann_source", + "//third_party/opencv/modules/flann:libopencv_flann", + "//third_party/opencv/modules/gapi:opencv_gapi_source", + "//third_party/opencv/modules/gapi:libopencv_gapi", + "//third_party/opencv/modules/highgui:opencv_highgui_source", + "//third_party/opencv/modules/highgui:libopencv_highgui", + "//third_party/opencv/modules/imgcodecs:opencv_imgcodecs_source", + "//third_party/opencv/modules/imgcodecs:libopencv_imgcodecs", + "//third_party/opencv/modules/imgproc:opencv_imgproc_source", + "//third_party/opencv/modules/imgproc:libopencv_imgproc", + "//third_party/opencv/modules/ml:opencv_ml_source", + "//third_party/opencv/modules/ml:libopencv_ml", + "//third_party/opencv/modules/objdetect:opencv_objdetect_source", + "//third_party/opencv/modules/objdetect:libopencv_objdetect", + "//third_party/opencv/modules/photo:opencv_photo_source", + "//third_party/opencv/modules/photo:libopencv_photo", + "//third_party/opencv/modules/stitching:opencv_stitching_source", + "//third_party/opencv/modules/stitching:libopencv_stitching", + "//third_party/opencv/modules/ts:opencv_ts_source", + "//third_party/opencv/modules/ts:libopencv_ts", + "//third_party/opencv/modules/video:opencv_video_source", + "//third_party/opencv/modules/video:libopencv_video", + "//third_party/opencv/modules/videoio:opencv_videoio_source", + "//third_party/opencv/modules/videoio:libopencv_videoio", + "//third_party/opencv/napi:opencv_napi" + ``` + +4. 增加子组件opencv：修改文件"productdefine/common/inherit/default.json"，在thirdparty子系统下新增组件opencv。 + + ``` + { + "component": "opencv", + "features": [] + } + ``` + +5. 依赖OpenCV的组件增加opencv依赖：在依赖于opencv的编译目标下增加dep，如果仅仅依赖OpenCV的部分模块，可选择"opencv_image", "opencv_objdetect", "opencv_video"和"opencv_AI"四个中的任意一个作为依赖项，它们分别表示图像基本操作与修复功能集，目标检测类功能集，视频类功能集和AI类功能集。 + ``` + deps += [ "//third_party/opencv:opencv" ] + ``` + +6. 编译：在OpenHarmony根目录下，执行如下命令开始编译(注：rk3568是编译出的产品名称，请根据实际编译产品替换)，编译完成后在"out/rk3568/packages/phone/system/lib"可查看编译生成的多个opencv动态so库. + ``` + ./build.sh --product-name rk3568 --ccache + ``` +### 单独构建 +本仓提供shell脚本和交叉编译脚本，用户可以在linux ubuntu环境单独编译OpenCV，编译出的动态so库拷贝到OpenHarmony设备的"/system/lib/"目录下可被其他程序调用。 +* 构建环境： +1. cmake 3.12或以上版本 +2. git客户端和git-lfs +3. repo +4. wget +5. OpenHarmony full sdk下载链接(OpenHarmony各发布版本README.md文档中提供full sdk下载链接) + +* 构建步骤： +1. 下载OpenCV源代码： + + 执行如下命令下载OpenCV代码到指定目录下。 + ``` + git clone https://gitee.com/openharmony-sig/third_party_opencv.git + ``` +2. 获取sdk下载地址： +OpenHarmony的每个发布的版本md文件中均提供了sdk的下载路径，如OpenHarmony4.0Release版本的说明链接为[OpenHarmony4.0 Release版本说明](https://gitee.com/openharmony/docs/raw/master/zh-cn/release-notes/OpenHarmony-v4.0-release.md); 进入该链接后在"从镜像站点获取"章节的"表2 获取源码路径"表中找到标准系统Public SDK包的下载站点。 +3. 执行构建脚本： + 进入到目录"platforms/ohos", 执行下面命令，开始下载sdk和编译OpenCV, 其中下面命令中的脚本参数"download_url"的值是OpenHarmony4.0 Release版本配套的sdk下载地址，请根据实际版本替换相应sdk下载地址。编译成功后的动态库在"build_ohos/lib/"目录下，可执行文件在"build_ohos/bin/"目录下。 + ```sh + bash build_auto_wget.sh --download_url=https://repo.huaweicloud.com/openharmony/os/4.0-Release/ohos-sdk-windows_linux-public.tar.gz + ``` + +## 应用示例 + * 应用示例简介 + 本应用示例基于Native c++模版构建，实现了人脸检测，轮廓识别，二维码识别，图像灰度图四个简单场景sample。 + + * 约束 + 安装应用示例之前，请先确认已安装应用开发IDE DevEco Studio3.1 Release及以上版本，并安装了sdk API 9. + + * 工程目录 + ``` + entry/src/main + |-- cpp + |   |-- CMakeLists.txt // 编译脚本 + |   |-- common + |   |   |-- common.cpp // 公共函数 + |   |   |-- init.cpp // napi初始化，注册函数 + |   |-- image_basic + |   |   |-- edgeDetection.cpp // 调用轮廓检测native接口 + |   |   |-- faceDetect.cpp // 调用人脸检测native接口 + |   |   |-- image2Gray.cpp // 调用图像转换为灰度图native接口 + |   |   |-- QRCode_identify.cpp // 调用二维码识别native接口 + |   |-- include + |   |   |-- common.h // 公共函数头文件 + |   |   |-- opencv_napi_api.h // native接口头文件 + |   |-- types + |   | |-- libohos_opencv_sample + |   | | |-- oh-package.json5 + |   | | |-- opencv_sample.d.ts // 接口导出 + |-- ets + |   |-- common + |   |   |-- AttributeModificationTool.ets // 属性修改工具组件 + |   |   |-- IntroductionTitle.ets // 介绍类标题自定义组件 + |   |   |-- TabContentNavigation.ets // 页签内容导航组件 + |   |   |-- TitleBar.ets // 页标题自定义组件 + |   |-- data + |   |   |-- CollectionCategory.ets // 页面内容数组 + |   |   |-- ResourceDataHandle.ets // 资源类数据转换普通数据类型接口 + |   |-- entryability + |   |   |-- EntryAbility.ts + |   |-- model + |   |   |── CategoricalDataType.ets // 页面内容导航数据结构定义 + |   |-- pages + |   |   |-- image + |   |   |   |-- basicSample + |   |   |   |   |-- edgeDetection.ets // 轮廓检测页面 + |   |   |   |   |-- faceDetect.ets // 人脸检测检测页面 + |   |   |   |   |-- image2Gray.ets // 图像转灰度图页面 + |   |   |   |   |-- qrcode.ets // 二维码识别页面 + |   |   |   |   |-- qrcode_result_web.ets // 二维码识别结果网页显示页面 + |   |   |   |-- ImageData.ets // 图像页签内容定义 + |   |   |-- Index.ets // 首页 + |   |   |-- machineLearning + |   |   |   |-- MachineLearningData.ets + |   |   |-- video + |   |   | |-- VideoData.ets + |   |-- util + |   | |-- Logger.ts // hilog封装 + |   | |-- ShowToast.ets // ShowToast封装 + |-- module.json5 + |-- resources + | |-- base + | |-- en_US + | |-- rawfile + | |   |-- haarcascade_frontalface_alt.xml // 人脸检测模型文件 + | |   |-- lena.jpg // 应用示例使用的图像文件 + | |   |-- qrcode2.jpg // 应用示例使用的图像文件 + | |   |-- qrcode.jpg // 应用示例使用的图像文件 + |-- |-- zh_CN + ``` + + * 使用方法 + 1. 下载：应用示例在本仓的路径"samples/ohos_samples/opencv_app_sample"。先下载本仓完成的代码，然后拷贝文件夹"opencv_app_sample"到windows环境下的应用目录下。 + 2. 拷贝opencv源代码到"opencv_app_sample/entry/src/main/cpp"目录下并修改文件夹名称为"opencv" + 3. 将应用示例工程"opencv_app_sample"导入DevEco Studio进行编译构建，安装到OpenHarmony设备上运行调试。 + + * 运行效果预览 +|首页|人脸检测|轮廓识别|灰度图|二维码识别| +|--------------------------------|--------------------------------|--------------------------------|--------------------------------|--------------------------------| +|![image](screenshots/homepage.jpeg)|![image](screenshots/objdetect.jpeg)|![image](screenshots/profile.jpeg)|![image](screenshots/gray.jpeg)|![image](screenshots/qrcode.jpeg)| + + diff --git a/screenshots/gray.jpeg b/screenshots/gray.jpeg new file mode 100755 index 0000000000000000000000000000000000000000..0dc27ff8b0d8bc371331716ddb56ce54951c8141 Binary files /dev/null and b/screenshots/gray.jpeg differ diff --git a/screenshots/homepage.jpeg b/screenshots/homepage.jpeg new file mode 100755 index 0000000000000000000000000000000000000000..e40a6ad87c65a77a0eaca7d84727567dcfa20e7e Binary files /dev/null and b/screenshots/homepage.jpeg differ diff --git a/screenshots/objdetect.jpeg b/screenshots/objdetect.jpeg new file mode 100755 index 0000000000000000000000000000000000000000..00b53c1037f1f2d97225d0c623f5aa9ae3de4400 Binary files /dev/null and b/screenshots/objdetect.jpeg differ diff --git a/screenshots/profile.jpeg b/screenshots/profile.jpeg new file mode 100755 index 0000000000000000000000000000000000000000..1f3d9025368a29d82921245c37cb0987fd135efb Binary files /dev/null and b/screenshots/profile.jpeg differ diff --git a/screenshots/qrcode.jpeg b/screenshots/qrcode.jpeg new file mode 100755 index 0000000000000000000000000000000000000000..bfa358c2a4e410ca2e3beb8afbf321892077cbb1 Binary files /dev/null and b/screenshots/qrcode.jpeg differ