opencv的使用（Ubuntu linux环境，AS jni，AS java）

最近要完成一个功能，就是把四个视频合成左右上下分布的一个视频。尝试很多方法，最终使用opencv来实现该功能。（通过opencv实现的视频好像没有声音。）研究的步骤，首先在Ubuntu环境测试，该功能是否实现。然后再将生成的库文件放到AS中，使用jni的方法调用，或者将源码放到AS中利用jni技术。在实现过程中遇到很多问题，下面记录。
一、在ubuntu linux环境使用opencv。
1、下载opencv安装包。
下载地址：官网：https://opencv.org/releases/

git地址：https://github.com/opencv/opencv/releases

2、将下载的安装包放到虚拟机根目录（新建software目录）。

3、安装包解压缩。

  unzip opencv-4.8.0.zip 

4、下载相关软件。
进入 opencv-4.8.0 文件夹。
1）更新软件

sudo apt-get update

2）安装cmake

sudo apt-get install cmake

3） 安装依赖库

sudo apt-get install build-essential libgtk2.0-dev libavcodec-dev libavformat-dev libjpeg.dev libtiff5.dev libswscale-dev libjasper-dev

4）在 opencv-4.8.0 文件夹下新建build文件夹。

 mkdir build

5）进入build文件夹，使用命令修改参数

 sudo cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local ..

6）使用make编译

sudo make

7）安装

sudo make install

8）配置环境

sudo gedit /etc/ld.so.conf.d/opencv.conf 

在新建的文档中添加：

/usr/local/lib

sudo ldconfig

配置路径生效。


9）配置bash

sudo gedit /etc/bash.bashrc 

在最末尾添加

PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig  
export PKG_CONFIG_PATH 

执行该步骤时，发现/usr/local/lib/路径下没有pkgconfig文件夹。并且执行命令

pkg-config --cflags openc

报一下错误。

解决：首先创建opencv.pc文件，这里要注意它的路径信息：

cd /usr/local/lib
sudo mkdir pkgconfig
cd pkgconfig
sudo touch opencv.pc

然后在opencv.pc中添加以下信息，注意这些信息需要与自己安装opencv时的库路径对应：（一下是我自己的内容）

prefix=/usr/local
exec_prefix=${prefix}
libdir=${exec_prefix}/lib
includedir=${prefix}/include/opencv4

Name: OpenCV
Description: Open Source Computer Vision Library
Version: 4.8.0
Libs: -L${exec_prefix}/lib -lopencv_highgui -lopencv_shape -lopencv_objdetect -lopencv_ml -lopencv_superres -lopencv_dnn -lopencv_stitching -lopencv_videostab -lopencv_calib3d -lopencv_videoio -lopencv_imgcodecs -lopencv_features2d -lopencv_video -lopencv_photo -lopencv_imgproc -lopencv_flann -lopencv_core
Libs.private: -ldl -lm -lpthread -lrt
Cflags: -I${includedir}

保存退出，添加环境：

export  PKG_CONFIG_PATH=/usr/local/lib/pkgconfig

执行.cpp编译命令：

 g++ videotest.cpp -o videotest `pkg-config --cflags --libs opencv`

报错：

再执行编译命令：

g++ videotest.cpp -o videotest `pkg-config --cflags --libs opencv` -std=gnu++11

报错：

查看/usr/local/lib/lib文件下没有libopencv_shape.so等这三个库，再次打开/usr/local/lib/pkgconfig/opencv.pc文件，将-lopencv_shape 等三个路径删除掉。保存退出。执行命令：

g++ videotest.cpp -o videotest `pkg-config --cflags --libs opencv` -std=gnu++11

编译成功。

最后的opencv.pc文件：

# Package Information for pkg-config

prefix=/usr/local
exec_prefix=${prefix}
libdir=${exec_prefix}/lib
includedir=${prefix}/include/opencv4

Name: OpenCV
Description: Open Source Computer Vision Library
Version: 4.8.0
Libs: -L${exec_prefix}/lib -lopencv_highgui -lopencv_objdetect -lopencv_ml -lopencv_dnn -lopencv_stitching -lopencv_calib3d -lopencv_videoio -lopencv_imgcodecs -lopencv_features2d -lopencv_video -lopencv_photo -lopencv_imgproc -lopencv_flann -lopencv_core
Libs.private: -ldl -lm -lpthread -lrt
Cflags: -I${includedir}

注意：1、该opencv源码编译的对应so库文件只是linux系统平台文件（我的是x86_64）,如果想将该库编译为其他平台，例如arm64需要交叉编译。（我没整明白）
注意：2、上述配置bash后，使用以下命令使得配置文件生效。若没有生效，重启电脑试试。（如果没有生效，执行 编译生成的可执行文件，报错 找不到opencv库）

source /etc/bash.bashrc  
sudo updatedb

注意：3 我这里编译的是opencv4.8.0版本，编译过程中使用python3.x版本，opencv.pc需要自己创建。链接http://www.taodudu.cc/news/show-3639538.html?action=onClick#google_vignette 编译的opencv3.4.1版本，使用python2.x版本，opencv.pc自动生成。按照链接中配置，就可以成功编译安装opencv。

最后是测试代码，功能将代码中的video.mp4合成上下左右分布的一个视频。生成视频为mergevideo.avi 或mergevideo.mkv

#include <iostream>
#include <opencv2/opencv.hpp>
#include <unistd.h>
#include <error.h>
#include <errno.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <sys/types.h>
#include <pthread.h>
#include <linux/videodev2.h>
#include <sys/mman.h>
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <iostream>
#include <iomanip>
#include <string>

using namespace std;
using namespace cv;
int main(int argc, char ** argv)
{
    std::string videoFile = "video.mp4";//视频的路径
    std::string videoFile1 = "video.mp4";//视频的路径
    std::string videoFile2 = "video.mp4";//视频的路径
    std::string videoFile3 = "video.mp4";//视频的路径

    /** 打开第一个视频文件 */
    VideoCapture cap; //视频句柄变量
    cap.open(videoFile);//打开视频
    if(!cap.isOpened())  //判断是否打开了
    {  
        printf("1cap.isOpened is error\n");
        return -1;  
    } 
    
    /** 打开第二个视频文件 */
    VideoCapture cap1; //视频句柄变量
    cap1.open(videoFile1);//打开视频
    if(!cap1.isOpened())  //判断是否打开了
    {  
        printf("2cap.isOpened is error\n");
        return -1;  
    } 
    
    /** 打开第三个视频文件 */
    VideoCapture cap2; //视频句柄变量
    cap2.open(videoFile2);//打开视频
    if(!cap2.isOpened())  //判断是否打开了
    {  
        printf("3cap.isOpened is error\n");
        return -1;  
    } 

    /** 打开第四个视频文件 */
    VideoCapture cap3; //视频句柄变量
    cap3.open(videoFile3);//打开视频
    if(!cap3.isOpened())  //判断是否打开了
    {  
        printf("4cap.isOpened is error\n");
        return -1;  
    } 


    /** 打开第一个视频文件的帧数 */
    int frame_num = cap.get(cv::CAP_PROP_FRAME_COUNT);
	std::cout << "videoFile total frame number is: " << frame_num << std::endl;
	
	/** 打开第二个视频文件的帧数 */
	int frame_num1 = cap1.get(cv::CAP_PROP_FRAME_COUNT);
	std::cout << "videoFile1 total frame number is: " << frame_num1 << std::endl;

    /** 打开第三个视频文件的帧数 */
    int frame_num2 = cap2.get(cv::CAP_PROP_FRAME_COUNT);
	std::cout << "videoFile2 total frame number is: " << frame_num2 << std::endl;

	/** 打开第四个视频文件的帧数 */
	int frame_num3 = cap3.get(cv::CAP_PROP_FRAME_COUNT);
	std::cout << "videoFile3 total frame number is: " << frame_num3 << std::endl;

	/** 打开第一个视频文件的帧率 */
    int fps = cap.get(cv::CAP_PROP_FPS);
	std::cout << "videoFile fps: " << fps << std::endl;

	/** 打开第二个视频文件的帧率 */
	int fps1 = cap1.get(cv::CAP_PROP_FPS);
	std::cout << "videoFile1 fps1: " << fps1 << std::endl;

    /** 打开第三个视频文件的帧率 */
    int fps2 = cap2.get(cv::CAP_PROP_FPS);
	std::cout << "videoFile fps2: " << fps2 << std::endl;

	/** 打开第四个视频文件的帧率 */
	int fps3 = cap3.get(cv::CAP_PROP_FPS);
	std::cout << "videoFile1 fps3: " << fps3 << std::endl;

	 /** 打开第一个视频文件的宽度 */
    int image_width = cap.get(cv::CAP_PROP_FRAME_WIDTH);
	std::cout << "videoFile image width is: " << image_width << std::endl;

	/** 打开第二个视频文件的宽度 */
	int image_width1 = cap1.get(cv::CAP_PROP_FRAME_WIDTH);
	std::cout << "videoFile1 image width is: " << image_width1 << std::endl;

    /** 打开第三个视频文件的宽度 */
    int image_width2 = cap2.get(cv::CAP_PROP_FRAME_WIDTH);
	std::cout << "videoFile2 image width is: " << image_width2 << std::endl;

	/** 打开第四个视频文件的宽度 */
	int image_width3 = cap3.get(cv::CAP_PROP_FRAME_WIDTH);
	std::cout << "videoFile3 image width is: " << image_width3 << std::endl;

	/** 打开第一个视频文件的高度 */
    int image_height = cap.get(cv::CAP_PROP_FRAME_HEIGHT);
	std::cout << "videoFile image height: " << image_height << std::endl;

	/** 打开第二个视频文件的高度 */
	int image_height1 = cap1.get(cv::CAP_PROP_FRAME_HEIGHT);
	std::cout << "videoFile1 image height: " << image_height1 << std::endl;

    /** 打开第三个视频文件的高度 */
    int image_height2 = cap2.get(cv::CAP_PROP_FRAME_HEIGHT);
	std::cout << "videoFile2 image height: " << image_height2 << std::endl;

	/** 打开第四个视频文件的高度 */
	int image_height3 = cap3.get(cv::CAP_PROP_FRAME_HEIGHT);
	std::cout << "videoFile3 image height: " << image_height3 << std::endl;

    /** 打开第一个视频文件的矩阵对象的格式*/
     int frame_format = cap.get(cv::CAP_PROP_FORMAT);
	std::cout << "videoFile frame format: " << frame_format << std::endl;

	/** 打开第二个视频文件的矩阵对象的格式 */
	int frame_format1 = cap1.get(cv::CAP_PROP_FORMAT);
	std::cout << "videoFile1 frame format: " << frame_format1 << std::endl;

    /** 打开第三个视频文件的矩阵对象的格式*/
     int frame_format2 = cap2.get(cv::CAP_PROP_FORMAT);
	std::cout << "videoFile2 frame format: " << frame_format2 << std::endl;

	/** 打开第四个视频文件的矩阵对象的格式 */
	int frame_format3 = cap3.get(cv::CAP_PROP_FORMAT);
	std::cout << "videoFile3 frame format: " << frame_format3 << std::endl;

	/** 合并视频初始化 */
	//std::string mergeVideoFile = "mergeVideo.avi";
    std::string mergeVideoFile = "mergeVideo.mkv";
    VideooHeight = 720;
	int mergeVideooWidth = 1280;
	float mergeVideooFps = 10.0;
	int mergeVideooFpsFrameFormat = CV_8UC3; // CV_8UC3代表每个像素占8位
	
	cv::VideoWriter track_writer;
	
    Mat img = cv::Mat::zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);
    
    //track_writer.open (mergeVideoFile, cv::VideoWriter::fourcc('m', 'p', '4', '2'), mergeVideooFps,       cv::Size(mergeVideooWidth, mergeVideooHeight));//.avi
    
   track_writer.open (mergeVideoFile, cv::VideoWriter::fourcc('m', 'p', '4', 'v'), mergeVideooFps, cv::Size(mergeVideooWidth, mergeVideooHeight));//.mkv
   
    if(!track_writer.isOpened())
    {
       std::cout << "!track_writer.isOpened(): " << std::endl;

        assert("track writer open failed!\n");
    }
    
    Mat frame; 
    Mat frame1;
    Mat frame2; 
    Mat frame3;
    while(1)
    {
        cap.read(frame);//从第一个视频获取一帧图片
        cap1.read(frame1);//从第二个视频获取一帧图片
        cap2.read(frame2);//从第三个视频获取一帧图片
        cap3.read(frame3);//从第四个视频获取一帧图片
        
        if(frame.empty()) break; //是否加载成功
        if(frame1.empty()) break; //是否加载成功
        if(frame2.empty()) break; //是否加载成功
        if(frame3.empty()) break; //是否加载成功

        // 设定ROI区域：截取一部分进行合并
        Mat imageROI= frame(Rect(0,0,frame.cols,frame.rows)); // Rect(左上角横坐标，左上角纵坐标，宽度，高度)
        Mat imageROI1= frame1(Rect(0,0,frame1.cols,frame1.rows));
        Mat imageROI2= frame2(Rect(0,0,frame2.cols,frame2.rows));
        Mat imageROI3= frame3(Rect(0,0,frame3.cols,frame3.rows));

        // 大小转换
        Mat imageROIdst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); 
        resize(imageROI, imageROIdst, imageROIdst.size());
        
        Mat imageROI1dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); 
        resize(imageROI1, imageROI1dst, imageROI1dst.size());
        
        Mat imageROI2dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); 
        resize(imageROI2, imageROI2dst, imageROI2dst.size());

        Mat imageROI3dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); 
        resize(imageROI3, imageROI3dst, imageROI3dst.size());

        // 视频写字
        // putText(imageROIdst, videoFile, Point(5, 55),FONT_HERSHEY_PLAIN,2.0,(0, 255, 255),2);

        //创建目标Mat
        Mat des;
        des.create(mergeVideooHeight,mergeVideooWidth, imageROI1dst.type()); // Mat.create(高，宽，像素编码类型这里是CV_8UC3)
        
        // 视频帧合并
        Mat r1 = des(Rect(0, 0, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI1dst.copyTo(r1);
        
        Mat r = des(Rect(mergeVideooWidth/2, 0, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROIdst.copyTo(r);

        Mat r2 = des(Rect(0, mergeVideooHeight/2, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI2dst.copyTo(r2);

        Mat r3 = des(Rect(mergeVideooWidth/2, mergeVideooHeight/2, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI3dst.copyTo(r3);
        
        // 格式化要保存的视频帧
        cv::resize(des, img , cv::Size(mergeVideooWidth, mergeVideooHeight));

        // 保存视频
        track_writer.write(img);
    }
    cap.release();//释放视频句柄
    cap1.release();//释放视频句柄
    cap2.release();//释放视频句柄
    cap3.release();//释放视频句柄

    track_writer.release();
    return 0;
}

编译指令：

g++ videotest.cpp -o videotest `pkg-config --cflags --libs opencv` -std=gnu++11

执行指令：

./videotest

路径：

运行结果：

二、AS中集成opencv（jni环境）
上述已经说过，在linux编译的库文件为x86_64平台，我需要arm64平台。因此上述不满足我的需求。使用其他办法。在AS中集成opencv。
1、下载opencv 使用android平台的sdk包。
下载地址不变：下载地址：官网：https://opencv.org/releases/

2、AS中集成opencv库。查看文章"android studio 3.6.3 ndk开发-cpp文件中加载第三方.so库并生成新.so文件（二）"
链接：https://blog.csdn.net/zhuowalun8427/article/details/114294901 后半部分“AS中添加opencv环境”。
3、修改videotest.cpp。与linux环境的videotest.cpp文件相同，只是添加一些jni库文件等。在代码中详细指出。

#include <iostream>

#include <opencv2/opencv.hpp>

#include <unistd.h>

#include <error.h>

#include <errno.h>

#include <fcntl.h>

#include <sys/ioctl.h>

#include <sys/types.h>

#include <pthread.h>

#include <linux/videodev2.h>

#include <sys/mman.h>

#include <opencv2/core/core.hpp>

#include <opencv2/highgui/highgui.hpp>

#include <stdio.h>

#include <stdlib.h>

#include <string.h>



#include <iostream>

#include <iomanip>

#include <string>

// 修改1：添加jni环境需要的头文件
#include <jni.h>
#include <android/log.h>

// 修改2：jni 日志
#define TAG "lilitest" // 这个是自定义的LOG的标识
#define LOGI(...) __android_log_print(ANDROID_LOG_INFO,TAG ,__VA_ARGS__)

using namespace std;

using namespace cv;

// 修改3：main() 方法改成接口，方便java层调用
int mergeVideo()
{
    // 修改4：路径改成android路径，我这里是外置U盘
    std::string videoFile = "/storage/udisk1/assets/video.mp4";//视频的路径

    std::string videoFile1 = "/storage/udisk1/assets/video.mp4";//视频的路径

    std::string videoFile2 = "/storage/udisk1/assets/video.mp4";//视频的路径

    std::string videoFile3 = "/storage/udisk1/assets/video.mp4";//视频的路径

    

    /** 打开第一个视频文件 */
    VideoCapture cap; //视频句柄变量
    cap.open(videoFile);//打开视频
    if(!cap.isOpened())  //判断是否打开了
    {
        // 修改5：所有打印日志改为LOGI
        LOGI("1cap.isOpened is error");
        return -1;  

    }
    

    /** 打开第二个视频文件 */

    VideoCapture cap1; //视频句柄变量
    cap1.open(videoFile1);//打开视频
    if(!cap1.isOpened())  //判断是否打开了
    {

        LOGI("2cap.isOpened is error\n");
        return -1;  

    } 

    /** 打开第三个视频文件 */
    VideoCapture cap2; //视频句柄变量
    cap2.open(videoFile2);//打开视频
    if(!cap2.isOpened())  //判断是否打开了
    {
        LOGI("3cap.isOpened is error\n");
        return -1;  

    } 

    /** 打开第四个视频文件 */
    VideoCapture cap3; //视频句柄变量
    cap3.open(videoFile3);//打开视频
    if(!cap3.isOpened())  //判断是否打开了
    {
        LOGI("4cap.isOpened is error\n");
        return -1;
    } 

    /** 打开第一个视频文件的帧数 */
    int frame_num = cap.get(cv::CAP_PROP_FRAME_COUNT);
	LOGI("videoFile total frame number is: %d",frame_num);

	/** 打开第二个视频文件的帧数 */
	int frame_num1 = cap1.get(cv::CAP_PROP_FRAME_COUNT);
	LOGI("videoFile1 total frame number is: %d",frame_num1);

    /** 打开第三个视频文件的帧数 */
    int frame_num2 = cap2.get(cv::CAP_PROP_FRAME_COUNT);
    LOGI("videoFile2 total frame number is: %d",frame_num2);

	/** 打开第四个视频文件的帧数 */
	int frame_num3 = cap3.get(cv::CAP_PROP_FRAME_COUNT);
    LOGI("videoFile3 total frame number is: %d",frame_num3);


	/** 打开第一个视频文件的帧率 */
    int fps = cap.get(cv::CAP_PROP_FPS);
    LOGI("videoFile fps: %d",fps);

	/** 打开第二个视频文件的帧率 */
	int fps1 = cap1.get(cv::CAP_PROP_FPS);
    LOGI("videoFile1 fps: %d",fps1);

    /** 打开第三个视频文件的帧率 */
    int fps2 = cap2.get(cv::CAP_PROP_FPS);
    LOGI("videoFile2 fps: %d",fps2);

	/** 打开第四个视频文件的帧率 */
	int fps3 = cap3.get(cv::CAP_PROP_FPS);
    LOGI("videoFile3 fps: %d",fps3);

	 /** 打开第一个视频文件的宽度 */
    int image_width = cap.get(cv::CAP_PROP_FRAME_WIDTH);
    LOGI("videoFile image width is: %d ",image_width);

	/** 打开第二个视频文件的宽度 */
	int image_width1 = cap1.get(cv::CAP_PROP_FRAME_WIDTH);
    LOGI("videoFile1 image width is: %d ",image_width1);

    /** 打开第三个视频文件的宽度 */
    int image_width2 = cap2.get(cv::CAP_PROP_FRAME_WIDTH);
    LOGI("videoFile2 image width is:  %d",image_width2);

	/** 打开第四个视频文件的宽度 */
	int image_width3 = cap3.get(cv::CAP_PROP_FRAME_WIDTH);
    LOGI("videoFile3 image width is: %d ",image_width3);

	/** 打开第一个视频文件的高度 */
    int image_height = cap.get(cv::CAP_PROP_FRAME_HEIGHT);
    LOGI( "videoFile image height:%d ",image_height);

	/** 打开第二个视频文件的高度 */
	int image_height1 = cap1.get(cv::CAP_PROP_FRAME_HEIGHT);
    LOGI( "videoFile1 image height: %d",image_height1);

    /** 打开第三个视频文件的高度 */
    int image_height2 = cap2.get(cv::CAP_PROP_FRAME_HEIGHT);
    LOGI( "videoFile2 image height:%d ",image_height2);

	/** 打开第四个视频文件的高度 */
	int image_height3 = cap3.get(cv::CAP_PROP_FRAME_HEIGHT);
    LOGI( "videoFile3 image height: %d",image_height3);


    /** 打开第一个视频文件的矩阵对象的格式*/
     int frame_format = cap.get(cv::CAP_PROP_FORMAT);
    LOGI("videoFile frame format = %d",frame_format);

	/** 打开第二个视频文件的矩阵对象的格式 */
	int frame_format1 = cap1.get(cv::CAP_PROP_FORMAT);
    LOGI("videoFile1 frame format = %d",frame_format1);

    /** 打开第三个视频文件的矩阵对象的格式*/
     int frame_format2 = cap2.get(cv::CAP_PROP_FORMAT);
    LOGI("videoFile2 frame format = %d",frame_format2);

	/** 打开第四个视频文件的矩阵对象的格式 */
	int frame_format3 = cap3.get(cv::CAP_PROP_FORMAT);
    LOGI("videoFile3 frame format = %d",frame_format3);

	
    // 修改6：合并视屏的路径改为android路径，后缀只能为.avi
	/** 合并视频初始化 */
	std::string mergeVideooFile = "/storage/udisk1/assets/mergeVideotest30_1280_720.avi";

	int mergeVideooHeight = 720;
	int mergeVideooWidth = 1280;
	float mergeVideooFps = 30.0;
	int mergeVideooFpsFrameFormat = CV_8UC3; // CV_8UC3代表每个像素占8位


	cv::VideoWriter track_writer;
    Mat img = cv::Mat::zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);
    // 修改7 对于android系统，fourcc只能是('M', 'J', 'P', 'G')，生成的视频格式只能是.avi否则报错
    // 补充：fourcc为('M', 'J', 'P', 'G')生成.avi在android10不能播放，改为('H', '2', '6', '4')生成的
    //.avi可以在android设备正常播放
    track_writer.open (mergeVideooFile, cv::VideoWriter::fourcc('M', 'J', 'P', 'G'), mergeVideooFps, cv::Size(mergeVideooWidth, mergeVideooHeight),true);//.avi

    if(!track_writer.isOpened())
    {
       LOGI("!track_writer.isOpened(): ");
        assert("track writer open failed!\n");
        return -2;
    }
    

    Mat frame; 

    Mat frame1;

    Mat frame2; 

    Mat frame3;


    while(1)

    {
        cap.read(frame);//从第一个视频获取一帧图片

        cap1.read(frame1);//从第二个视频获取一帧图片

        cap2.read(frame2);//从第三个视频获取一帧图片

        cap3.read(frame3);//从第四个视频获取一帧图片

        if(frame.empty()) break; //是否加载成功
        if(frame1.empty()) break; //是否加载成功
        if(frame2.empty()) break; //是否加载成功
        if(frame3.empty()) break; //是否加载成功

        Mat imageROI= frame(Rect(0,0,frame.cols,frame.rows)); // Rect(左上角横坐标，左上角纵坐标，宽度，高度)
        Mat imageROI1= frame1(Rect(0,0,frame1.cols,frame1.rows));
       Mat imageROI2= frame2(Rect(0,0,frame2.cols,frame2.rows));
       Mat imageROI3= frame3(Rect(0,0,frame3.cols,frame3.rows));
       

       // 大小转换
        Mat imageROIdst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); //我要转化为850*600大小的
	    resize(imageROI, imageROIdst, imageROIdst.size());

        Mat imageROI1dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); //我要转化为850*600大小的
	    resize(imageROI1, imageROI1dst, imageROI1dst.size());

        Mat imageROI2dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); //我要转化为850*600大小的
	    resize(imageROI2, imageROI2dst, imageROI2dst.size());

        Mat imageROI3dst = Mat::zeros(mergeVideooHeight/2,mergeVideooWidth/2 , CV_8UC3); //我要转化为850*600大小的
	    resize(imageROI3, imageROI3dst, imageROI3dst.size());

        // 视频写字
        // putText(imageROIdst, videoFile, Point(5, 55),FONT_HERSHEY_PLAIN,2.0,(0, 255, 255),2);

        //创建目标Mat
        Mat des;
        des.create(mergeVideooHeight,mergeVideooWidth, imageROI1dst.type()); // Mat.create(高，宽，像素编码类型这里是CV_8UC3)

        // 视频帧合并
        Mat r1 = des(Rect(0, 0, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI1dst.copyTo(r1);

        Mat r = des(Rect(mergeVideooWidth/2, 0, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROIdst.copyTo(r);

        Mat r2 = des(Rect(0, mergeVideooHeight/2, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI2dst.copyTo(r2);

        Mat r3 = des(Rect(mergeVideooWidth/2, mergeVideooHeight/2, mergeVideooWidth/2, mergeVideooHeight/2));
        imageROI3dst.copyTo(r3);

        // 格式化要保存的视频帧
        cv::resize(des, img , cv::Size(mergeVideooWidth, mergeVideooHeight));

        // 保存视频
        track_writer.write(img);


    }

    cap.release();//释放视频句柄
    cap1.release();//释放视频句柄
    cap2.release();//释放视频句柄
    cap3.release();//释放视频句柄


    track_writer.release();

	return 0;

}

extern "C"
JNIEXPORT jint JNICALL
Java_com_htkj_testproject_jni_opencvActivity_myMergeVideo(JNIEnv *env, jobject thiz) {
    return mergeVideo();
}

代码中，fourcc与文件后缀不匹配，报错如下：

java调用接口省略。

三、AS 中集成opencv（java环境）
在AS中新建一个android应用模块 “mergevideo”
1、下载sdk包，与上述相同。
2、导入sdk包。
1）在模块所在工程中导入opencv sdk模块。选择路径如下：


导入之后，在工程路径下的lib模块：

2）在mergevideo模块的build.gradle添加sdk模块，并同步：

implementation project(':sdk')

3、在java代码中引入adk库。

 // 没有这个加载不到opencv
        if (!OpenCVLoader.initDebug()) {
            Log.d(TAG, "Internal OpenCV library not found. Using OpenCV Manager for initialization");
            OpenCVLoader.initAsync(OpenCVLoader.OPENCV_VERSION_3_0_0, this, mLoaderCallback);
        } else {
            Log.d(TAG, "OpenCV library found inside package. Using it!");
            mLoaderCallback.onManagerConnected(LoaderCallbackInterface.SUCCESS);
        }

//openCV4Android 需要加载用到
    private BaseLoaderCallback mLoaderCallback = new BaseLoaderCallback(this) {
        @Override
        public void onManagerConnected(int status) {
            switch (status) {
                case LoaderCallbackInterface.SUCCESS: {
                    Log.i(TAG, "OpenCV loaded successfully");
//                    mOpenCvCameraView.enableView();
//                    mOpenCvCameraView.setOnTouchListener(ColorBlobDetectionActivity.this);
                }
                break;
                default: {
                    super.onManagerConnected(status);
                }
                break;
            }
        }
    };

4、编写合并视频的主代码，将原来的c++ 代码修改为java代码。

// merge.java
package com.htkj.mergevideo;

import static org.opencv.imgproc.Imgproc.resize;

import android.util.Log;

import org.opencv.core.CvType;
import org.opencv.core.Mat;
import org.opencv.core.Rect;
import org.opencv.core.Size;
import org.opencv.videoio.VideoCapture;
import org.opencv.videoio.VideoWriter;
import org.opencv.videoio.Videoio;

/**
 * @ClassName merge
 * @Description //TODO
 * @Author lili
 * @DATE 2023/8/28 9:34
 */
public class merge {
    private final String TAG = "mergetest";
    private String pathVido1 = "/storage/udisk1/assets/video.mp4";//视频的路径
    private String pathVido2 = "/storage/udisk1/assets/paomo.mp4";//视频的路径
    private String pathVido3 = "/storage/udisk1/assets/paomo.mp4";//视频的路径
    private String pathVido4 = "/storage/udisk1/assets/video.mp4";//视频的路径


    private VideoCapture mVideoCapture1;
    private VideoCapture mVideoCapture2;
    private VideoCapture mVideoCapture3;
    private VideoCapture mVideoCapture4;

    // 视屏拼接
    private String mergeVideooFile = "/storage/udisk1/assets/mergepaomovideo.avi";
    private int mergeVideooHeight = 640;
    private int mergeVideooWidth = 800;
    private double mergeVideooFps = 25.0;
    private int mergeVideooFpsFrameFormat = CvType.CV_8UC3; // CV_8UC3代表每个像素占8位
    private VideoWriter track_writer;


    public int openVideo() {
        Log.d(TAG, "openVideo...");

        // 打开第一个视频
        if (mVideoCapture1 == null) {
            mVideoCapture1 = new VideoCapture();
        }

        // Videoio.CAP_ANY 或者 Videoio.CAP_ANDROID才能正确加载视频
        boolean open1 = mVideoCapture1.open(pathVido1, Videoio.CAP_ANDROID);
        Log.d(TAG, "open1 = " + open1);
        if (!open1) {
            return -1;
        }

        // 打开第二个视频
        if (mVideoCapture2 == null) {
            mVideoCapture2 = new VideoCapture();
        }
        boolean open2 = mVideoCapture2.open(pathVido2);
        Log.d(TAG, "open2 = " + open2);
        if (!open2) {
            return -2;
        }

        // 打开第三个视频
        if (mVideoCapture3 == null) {
            mVideoCapture3 = new VideoCapture();
        }
        boolean open3 = mVideoCapture3.open(pathVido3);
        Log.d(TAG, "open3 = " + open3);
        if (!open3) {
            return -3;
        }

        // 打开第三个视频
        if (mVideoCapture4 == null) {
            mVideoCapture4 = new VideoCapture();
        }
        boolean open4 = mVideoCapture4.open(pathVido4);
        Log.d(TAG, "open4 = " + open4);
        if (!open4) {
            return -4;
        }
        getVideoFrameCount();
        getVideoFrameFps();
        getVideoHeight();
        getVideoWidth();
        getVidewFormat();
        merge();
        return 0;
    }

    // 获取视频帧数
    private int getVideoFrameCount() {

        // 第一个视屏
        double frameNum1 = mVideoCapture1.get(Videoio.CAP_PROP_FRAME_COUNT);
        Log.d(TAG, "videoFile1 total frame number is: " + frameNum1);
        return 0;
    }

    // 获取视频帧率
    private int getVideoFrameFps() {
        // 第一个视屏
        double fps1 = mVideoCapture1.get(Videoio.CAP_PROP_FPS);
        Log.d(TAG, "videoFile1 FPS number is: " + fps1);
        return 0;
    }

    // 获取视频高度
    private int getVideoHeight() {
        // 第一个视屏
        double height1 = mVideoCapture1.get(Videoio.CAP_PROP_FRAME_HEIGHT);
        Log.d(TAG, "videoFile1 height is: " + height1);
        return 0;
    }

    // 获取视频宽度
    private int getVideoWidth() {
        // 第一个视屏
        double width1 = mVideoCapture1.get(Videoio.CAP_PROP_FRAME_WIDTH);
        Log.d(TAG, "videoFile1 width is: " + width1);
        return 0;
    }

    // 获取视频文件的矩阵对象的格式
    private int getVidewFormat() {
        // 第一个视屏
        double format1 = mVideoCapture1.get(Videoio.CAP_PROP_FORMAT);
        Log.d(TAG, "videoFile1 format is: " + format1);
        return 0;
    }

    private int merge() {
        int fourcc = VideoWriter.fourcc('M', 'J', 'P', 'G');
        Mat img = Mat.zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);
        if (track_writer == null) {
            track_writer = new VideoWriter();
        }
        // 注意这里：https://zinyan.com/?p=498
        track_writer.open(mergeVideooFile, fourcc, mergeVideooFps, new Size(mergeVideooWidth, mergeVideooHeight), true);//.avi
        if (!track_writer.isOpened()) {
            Log.d(TAG, "the track is open faild!!!");
            return -5;
        }

        while (true) {
            Mat mFrame1 = Mat.zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);

            Mat mFrame2 = Mat.zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);

            Mat mFrame3 = Mat.zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);

            Mat mFrame4 = Mat.zeros(mergeVideooWidth, mergeVideooHeight, mergeVideooFpsFrameFormat);

            // 读取原始视频帧
            mVideoCapture1.read(mFrame1);
            if (mFrame1.empty()) {
                Log.d(TAG, "mFrame1 is empty!!!");
                break;
            }
            mVideoCapture2.read(mFrame2);
            if (mFrame2.empty()) {
                Log.d(TAG, "mFrame2 is empty!!!");
                break;
            }
            mVideoCapture3.read(mFrame3);
            if (mFrame3.empty()) {
                Log.d(TAG, "mFrame3 is empty!!!");
                break;
            }
            mVideoCapture4.read(mFrame4);
            if (mFrame4.empty()) {
                Log.d(TAG, "mFrame4 is empty!!!");
                break;
            }

            // 设定ROI区域：截取一部分进行合并

            Mat mImageROI1 = new Mat(mFrame1, new Rect(0, 0, mFrame1.cols(), mFrame1.rows())); // Rect(左上角横坐标，左上角纵坐标，宽度，高度)
            Mat mImageROI2 = new Mat(mFrame2, new Rect(0, 0, mFrame2.cols(), mFrame2.rows()));
            Mat mImageROI3 = new Mat(mFrame3, new Rect(0, 0, mFrame3.cols(), mFrame3.rows()));
            Mat mImageROI4 = new Mat(mFrame4, new Rect(0, 0, mFrame4.cols(), mFrame4.rows()));


            // 转换大小

            Mat mImageROIdst1 = Mat.zeros(mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooFpsFrameFormat);
            resize(mImageROI1, mImageROIdst1, mImageROIdst1.size());

            Mat mImageROIdst2 = Mat.zeros(mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooFpsFrameFormat);
            resize(mImageROI2, mImageROIdst2, mImageROIdst2.size());

            Mat mImageROIdst3 = Mat.zeros(mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooFpsFrameFormat);
            resize(mImageROI3, mImageROIdst3, mImageROIdst3.size());

            Mat mImageROIdst4 = Mat.zeros(mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooFpsFrameFormat);
            resize(mImageROI4, mImageROIdst4, mImageROIdst4.size());

            // 创建目标
            Mat des = Mat.zeros(mergeVideooHeight, mergeVideooWidth, mergeVideooFpsFrameFormat);

            // 视频帧合并
            Mat mMergeRoi1 = new Mat(des, new Rect(0, 0, mergeVideooWidth / 2, mergeVideooHeight / 2));
            mImageROIdst1.copyTo(mMergeRoi1);

            Mat mMergeRoi2 = new Mat(des, new Rect(mergeVideooWidth / 2, 0, mergeVideooWidth / 2, mergeVideooHeight / 2));
            mImageROIdst2.copyTo(mMergeRoi2);

            Mat mMergeRoi3 = new Mat(des, new Rect(0, mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooHeight / 2));
            mImageROIdst3.copyTo(mMergeRoi3);

            Mat mMergeRoi4 = new Mat(des, new Rect(mergeVideooWidth / 2, mergeVideooHeight / 2, mergeVideooWidth / 2, mergeVideooHeight / 2));
            mImageROIdst4.copyTo(mMergeRoi4);

            // 格式化要保存的视频帧
            resize(des, img, new Size(mergeVideooWidth, mergeVideooHeight));

            // 保存视频
            track_writer.write(img);

            if (mFrame1 != null) {
                mFrame1.release();
            }
            if (mFrame2 != null) {
                mFrame2.release();
            }
            if (mFrame3 != null) {
                mFrame3.release();
            }
            if (mFrame4 != null) {
                mFrame4.release();
            }

            if (mImageROI1 != null) {
                mImageROI1.release();
            }
            if (mImageROI2 != null) {
                mImageROI2.release();
            }
            if (mImageROI3 != null) {
                mImageROI3.release();
            }
            if (mImageROI4 != null) {
                mImageROI4.release();
            }

            if (mImageROIdst1 != null) {
                mImageROIdst1.release();
            }
            if (mImageROIdst2 != null) {
                mImageROIdst2.release();
            }
            if (mImageROIdst3 != null) {
                mImageROIdst3.release();
            }
            if (mImageROIdst4 != null) {
                mImageROIdst4.release();
            }

            if (mMergeRoi1 != null) {
                mMergeRoi1.release();
            }
            if (mMergeRoi2 != null) {
                mMergeRoi2.release();
            }
            if (mMergeRoi3 != null) {
                mMergeRoi3.release();
            }
            if (mMergeRoi4 != null) {
                mMergeRoi4.release();
            }
        }


        mVideoCapture1.release();
        mVideoCapture2.release();
        mVideoCapture3.release();
        mVideoCapture4.release();
        track_writer.release();

        return 0;
    }
}

主应用代码：

// MainActivity.java
package com.htkj.mergevideo;

import android.app.Activity;
import android.os.Bundle;
import android.util.Log;
import android.view.View;
import android.widget.Button;

import androidx.annotation.Nullable;

import org.opencv.android.BaseLoaderCallback;
import org.opencv.android.LoaderCallbackInterface;
import org.opencv.android.OpenCVLoader;

/**
 * @ClassName MainActivity
 * @Description //TODO
 * @Author lili
 * @DATE 2023/8/28 9:32
 */
public class MainActivity extends Activity {
    private final String TAG = MainActivity.class.getSimpleName();
    private Button mBtMerge;
    private merge mMerge;

    @Override
    protected void onCreate(@Nullable Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        setContentView(R.layout.main_activity);
        mMerge = new merge();
        mBtMerge = findViewById(R.id.bt_merge);
        mBtMerge.setOnClickListener(new View.OnClickListener() {
            @Override
            public void onClick(View view) {
                mergeVideoThread.start();
            }
        });
    }

    @Override
    protected void onResume() {
        super.onResume();

        // 没有这个加载不到opencv
        if (!OpenCVLoader.initDebug()) {
            Log.d(TAG, "Internal OpenCV library not found. Using OpenCV Manager for initialization");
            OpenCVLoader.initAsync(OpenCVLoader.OPENCV_VERSION_3_0_0, this, mLoaderCallback);
        } else {
            Log.d(TAG, "OpenCV library found inside package. Using it!");
            mLoaderCallback.onManagerConnected(LoaderCallbackInterface.SUCCESS);
        }
    }

    Thread mergeVideoThread = new Thread(new Runnable() {
        @Override
        public void run() {
            int result = mMerge.openVideo();
            Log.d(TAG,"result = " + result);
        }
    });

    //openCV4Android 需要加载用到
    private BaseLoaderCallback mLoaderCallback = new BaseLoaderCallback(this) {
        @Override
        public void onManagerConnected(int status) {
            switch (status) {
                case LoaderCallbackInterface.SUCCESS: {
                    Log.i(TAG, "OpenCV loaded successfully");
//                    mOpenCvCameraView.enableView();
//                    mOpenCvCameraView.setOnTouchListener(ColorBlobDetectionActivity.this);
                }
                break;
                default: {
                    super.onManagerConnected(status);
                }
                break;
            }
        }
    };
}

// build.gradle
plugins {
    id 'com.android.application'
}

android {
    compileSdk 33

    defaultConfig {
        applicationId "com.htkj.mergevideo"
        minSdk 21
        targetSdk 33
        versionCode 1
        versionName "1.0"

        testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
    }

    buildTypes {
        release {
            minifyEnabled false
            proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
        }
    }
    compileOptions {
        sourceCompatibility JavaVersion.VERSION_1_8
        targetCompatibility JavaVersion.VERSION_1_8
    }

    signingConfigs {
        release {
            storeFile file("../platform.jks")
            storePassword '123456'
            keyAlias 'platform'
            keyPassword '123456'
        }
        debug {
            storeFile file("../platform.jks")
            storePassword '123456'
            keyAlias 'platform'
            keyPassword '123456'
        }
    }
}

dependencies {
//    implementation fileTree(dir: 'libs', include: ['*.jar'])
    implementation 'androidx.appcompat:appcompat:1.3.0'
    implementation 'com.google.android.material:material:1.4.0'
    testImplementation 'junit:junit:4.13.2'
    androidTestImplementation 'androidx.test.ext:junit:1.1.3'
    androidTestImplementation 'androidx.test.espresso:espresso-core:3.4.0'
    implementation project(':sdk')
}

布局文件省略。
效果图：

四、AS中集成opencv(java + jni环境)
在学习opencv过程中，首先以opencv自带的人脸识别demo学习，这里主要记录opencv模块的加载以及，cmake文件的编写。
1、新建一个模块，并添加jni配置。
2、java代码引入opencv。添加opencv sdk库。与上述相同。
3、添加jni所需配置。
1）build.gradle中添加：

externalNativeBuild {
            cmake {
                cppFlags "-std=c++11 -frtti -fexceptions"
                arguments "-DOpenCV_DIR=" + "E:/work/xinchi/testProject/sdk/native/jni" // , "-DANDROID_ARM_NEON=TRUE"
            }
        }

 splits {
      abi {
          enable true
          universalApk false
          reset()
          include 'armeabi-v7a' // , 'x86', 'x86_64', 'arm64-v8a'
      }
  }


implementation fileTree(dir: 'libs', include: ['*.jar'])

完整build.gradle

plugins {
    id 'com.android.application'
}

android {
    compileSdk 33

    defaultConfig {
        applicationId "com.htkj.facedetection1"
        minSdk 21
        targetSdk 33
        versionCode 1
        versionName "1.0"

        testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
        externalNativeBuild {
            cmake {
                cppFlags "-std=c++11 -frtti -fexceptions"
                arguments "-DOpenCV_DIR=" + "E:/work/xinchi/testProject/sdk/native/jni" // , "-DANDROID_ARM_NEON=TRUE"
            }
        }

        splits {
            abi {
                enable true
                universalApk false
                reset()
                include 'armeabi-v7a' // , 'x86', 'x86_64', 'arm64-v8a'
            }
        }
    }

    buildTypes {
        release {
            minifyEnabled false
            proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
        }
    }
    compileOptions {
        sourceCompatibility JavaVersion.VERSION_1_8
        targetCompatibility JavaVersion.VERSION_1_8
    }
    externalNativeBuild {
        cmake {
            path file('src/main/cpp/CMakeLists.txt')
            version '3.18.1'
        }
    }

    signingConfigs {
        release {
            storeFile file("../platform.jks")
            storePassword '123456'
            keyAlias 'platform'
            keyPassword '123456'
        }
        debug {
            storeFile file("../platform.jks")
            storePassword '123456'
            keyAlias 'platform'
            keyPassword '123456'
        }
    }

//    ndkVersion = "21.4.7075529" // e.g.,  ndkVersion "21.3.6528147"
}

dependencies {
    implementation fileTree(dir: 'libs', include: ['*.jar'])
    implementation 'androidx.appcompat:appcompat:1.3.0'
    implementation 'com.google.android.material:material:1.4.0'
    testImplementation 'junit:junit:4.13.2'
    androidTestImplementation 'androidx.test.ext:junit:1.1.3'
    androidTestImplementation 'androidx.test.espresso:espresso-core:3.4.0'
    implementation project(':sdk')
}

2） 修改cmakeList.txt文件：

cmake_minimum_required(VERSION 3.18.1)

project("facedetection1" CXX)

# 导入头文件ok
include_directories(${CMAKE_SOURCE_DIR}/)

#set(OpenCV_DIR "E:/work/xinchi/testProject/sdk/native/jni")
# 设置OpenCV要加载的组件为opencv_java
#set(ANDROID_OPENCV_COMPONENTS "opencv_java" CACHE STRING "")
find_package(OpenCV 4.8.0 REQUIRED java)


add_library(
             facedetection1

             SHARED

             facedetection1.cpp )

find_library( # Sets the name of the path variable.
              log-lib

              log )


target_link_libraries(
                       facedetection1
                       ${log-lib}
                       ${OpenCV_LIBRARIES})

模块结构：
文章来源地址https://www.toymoban.com/news/detail-774890.html

到了这里，关于opencv的使用（Ubuntu linux环境，AS jni，AS java）的文章就介绍完了。如果您还想了解更多内容，请在右上角搜索TOY模板网以前的文章或继续浏览下面的相关文章，希望大家以后多多支持TOY模板网！