数据处理音视频像素入门AndroidYUV(像素数据数据处理音视频入门)「像素数据量」

这段时间学习音视频，在网上找了很多资料，翻看到雷神的博客，说的很是详细，准备按照雷神博客走一遍。
感谢前人的栽树。
本文记录RGB/YUV视频像素数据的处理方法，不使用任何第三方库，完全用C实现。
视频像素数据在视频播放器的解码流程中的位置如下图所示，RGB/YUV查看用yuvplayer.exe，已上传到项目git中。
包括如下几个处理函数：分离YUV420P像素数据中的Y、U、V分量分离YUV444P像素数据中的Y、U、V分量将YUV420P像素数据去掉颜色，变成灰度图将YUV420P像素数据的亮度减半将YUV420P像素数据的周围加上边框生成YUV420P格式的灰阶测试图计算两个YUV420P像素数据的PSNR分离RGB24像素数据中的R、G、B分量将RGB24格式像素数据封装为BMP图像将RGB24格式像素数据转换为YUV420P格式像素数据生成RGB24格式的彩条测试图1）分离YUV420P像素数据中的Y、U、V分量

/  Split Y, U, V planes in YUV420P file.  @param jstr_url Location of Input YUV file.  @param w    Width of Input YUV file.  @param h    Height of Input YUV file.  @param num  Number of frames to process.  分离YUV420P像素数据中的Y、U、V分量  如果视频帧的宽和高分别为w和h，那么一帧YUV420P像素数据一共占用wh3/2 Byte的数据。
其中前wh Byte存储Y，接着的wh1/4 Byte存储U，最后wh1/4 Byte存储V /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420Split(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h, jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_y.y");    FILE fp1 = fopen(input_url,"wb+");     strcpy(input_url+strlen(input_url)-4,"_u.y");    FILE fp2 = fopen(input_url,"wb+");     strcpy(input_url+strlen(input_url)-4,"_v.y");    FILE fp3 = fopen(input_url,"wb+");     unsigned char pic = (unsigned char )malloc(wh3/2);    for(int i = 0;i < num;i++) {        fread(pic, 1, w  h  3 / 2, fp);        //Y        fwrite(pic, 1, w  h, fp1);        //U        fwrite(pic + w  h, 1, w  h / 4, fp2);        //V        fwrite(pic + w  h  5 / 4, 1, w  h / 4, fp3);    }    free(pic);    fclose(fp);    fclose(fp1);    fclose(fp2);    fclose(fp3);    return 0;}

函数调用AVUtils.getInstance().yuv420Split(ROOT+"lena_256x256_yuv420p.yuv",256,256,1);运行上面代码后会在同目录下生成y、u、v三个文件，工具查看要改为对应格式和分辨率大小查看

lena_256x256_yuv420p_y.y： Y数据， 分辨率为256x256lena_256x256_yuv420p_u.y： U数据， 分辨率为128x128lena_256x256_yuv420p_v.y： V数据， 分辨率为128x128

原始图片：lena_256x256_yuv420p.yuv生成的y、u、v分量图片如下：lena_256x256_yuv420p_y.ylena_256x256_yuv420p_u.ylena_256x256_yuv420p_v.y2）分离YUV444P像素数据中的Y、U、V分量

/  Split Y, U, V planes in YUV444P file.  @param jstr_url Location of YUV file.  @param w    Width of Input YUV file.  @param h    Height of Input YUV file.  @param num  Number of frames to process.  分离YUV444P像素数据中的Y、U、V分量  如果视频帧的宽和高分别为w和h，那么一帧YUV444P像素数据一共占用wh3 Byte的数据。
其中前wh Byte存储Y，接着的wh Byte存储U，最后wh Byte存储V /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv444Split(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h, jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_y.y");    FILE fp1 = fopen(input_url,"wb+");    strcpy(input_url+strlen(input_url)-4,"_u.y");    FILE fp2 = fopen(input_url,"wb+");    strcpy(input_url+strlen(input_url)-4,"_v.y");    FILE fp3 = fopen(input_url,"wb+");    unsigned char pic = (unsigned char ) malloc(w  h  3);    for(int i = 0;i < num;i++) {        fread(pic, 1, w  h  3, fp);        //Y        fwrite(pic, 1, w  h, fp1);        //U        fwrite(pic + w  h, 1, w  h, fp2);        //V        fwrite(pic + w  h  2, 1, w  h, fp3);    }    free(pic);    fclose(fp);    fclose(fp1);    fclose(fp2);    fclose(fp3);    return 0;}

函数调用AVUtils.getInstance().yuv444Split(ROOT+"lena_256x256_yuv444p.yuv",256,256,1);运行上面代码后会在同目录下生成y、u、v三个文件

lena_256x256_yuv444p_u.y： Y数据， 分辨率为256x256lena_256x256_yuv444p_u.y： U数据， 分辨率为256x256lena_256x256_yuv444p_v.y： V数据， 分辨率为256x256

原始图片lena_256x256_yuv444p.yuv生成的y、u、v分量图片如下：lena_256x256_yuv444p_y.ylena_256x256_yuv444p_u.ylena_256x256_yuv444p_v.y3）将YUV420P像素数据去掉颜色，变成灰度图

/  Convert YUV420P file to gray picture  @param jstr_url Location of Input YUV file.  @param w       Width of Input YUV file.  @param h       Height of Input YUV file.  @param num     Number of frames to process.  将YUV420P像素数据去掉颜色，变成灰度图  如果想把YUV格式像素数据变成灰度图像，只需要将U、V分量设置成128即可。
  这是因为U、V是图像中的经过偏置处理的色度分量。
色度分量在偏置处理前的取值范围是-128至127，这时候的无色对应的是0值。
经过偏置后色度分量取值变成了0至255，因而此时的无色对应的就是128了 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420Gray(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h, jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_gray.yuv");    FILE fp1 = fopen(input_url,"wb+");    unsigned char pic=(unsigned char )malloc(wh3/2);    for(int i = 0;i < num;i++) {        fread(pic, 1, w  h  3 / 2, fp);        //Gray        memset(pic + w  h, 128, w  h / 2);        fwrite(pic, 1, w  h  3 / 2, fp1);    }    free(pic);    fclose(fp);    fclose(fp1);    return 0;}

函数调用AVUtils.getInstance().yuv420Gray(ROOT+"lena_256x256_yuv420p.yuv",256,256,1);原始图片lena_256x256_yuv420p.yuv生成的灰度图如下：lena_256x256_yuv420p_gray.yuv4）将YUV420P像素数据的亮度减半

/  Halve Y value of YUV420P file  @param jstr_url Location of Input YUV file.  @param w       Width of Input YUV file.  @param h       Height of Input YUV file.  @param num     Number of frames to process.  将YUV420P像素数据的亮度减半  如果打算将图像的亮度减半，只要将图像的每个像素的Y值取出来分别进行除以2就可以了 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420Halfy(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h, jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_halfy.yuv");    FILE fp1 = fopen(input_url,"wb+");    unsigned char pic=(unsigned char )malloc(wh3/2);    for(int i = 0;i < num;i++) {        fread(pic, 1, w  h  3 / 2, fp);        //Half        for (int j = 0; j < w  h; j++) {            unsigned char temp = pic[j] / 2;            pic[j] = temp;        }        fwrite(pic, 1, w  h  3 / 2, fp1);    }    free(pic);    fclose(fp);    fclose(fp1);    return 0;}

函数调用AVUtils.getInstance().yuv420Halfy(ROOT+"lena_256x256_yuv420p.yuv",256,256,1);原始图片lena_256x256_yuv420p.yuv生成的亮度减半图如下：lena_256x256_yuv420p_halfy.yuv5）将YUV420P像素数据的周围加上边框

/  Add border for YUV420P file  @param jstr_url Location of Input YUV file.  @param w       Width of Input YUV file.  @param h       Height of Input YUV file.  @param border  Width of Border.  @param num     Number of frames to process.  将YUV420P像素数据的周围加上边框  图像的边框的宽度为border，本程序将距离图像边缘border范围内的像素的亮度分量Y的取值设置成了亮度最大值255 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420Border(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h, jint border, jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_border.yuv");    FILE fp1 = fopen(input_url,"wb+");    unsigned char pic=(unsigned char )malloc(wh3/2);    for(int i = 0;i < num;i++){        fread(pic,1,wh3/2,fp);        //Y        for(int j = 0;j < h;j++){            for(int k = 0;k < w;k++){                if(k < border || k >= (w-border) || j < border || j >= (h-border)){                    pic[jw+k]=255;                    //pic[jw+k]=0;                }            }        }        fwrite(pic,1,wh3/2,fp1);    }    free(pic);    fclose(fp);    fclose(fp1);    return 0;}

函数调用AVUtils.getInstance().yuv420Border(ROOT+"lena_256x256_yuv420p.yuv",256,256,20,1);原始图片lena_256x256_yuv420p.yuv生成的周围加上边框图如下：lena_256x256_yuv420p_border.yuv6）生成YUV420P格式的灰阶测试图

/  Generate YUV420P gray scale bar.  @param width    Width of Output YUV file.  @param height   Height of Output YUV file.  @param ymin     Max value of Y  @param ymax     Min value of Y  @param barnum   Number of bars  @param jstr_url_out Location of Output YUV file.  生成YUV420P格式的灰阶测试图  通过灰阶测试图的亮度最小值ymin，亮度最大值ymax，灰阶数量barnum确定每一个灰度条中像素的亮度分量Y的取值。
  还要根据图像的宽度width和图像的高度height以及灰阶数量barnum确定每一个灰度条的宽度 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420GrayBar(        JNIEnv env,jobject / this /,        jstring jstr_url_out, jint width, jint height, jint ymin,jint ymax,jint barnum) {    int barwidth;    float lum_inc;    unsigned char lum_temp;    int uv_width,uv_height;    FILE fp=NULL;    unsigned char data_y = NULL;    unsigned char data_u = NULL;    unsigned char data_v = NULL;    int t = 0,i = 0,j = 0;    barwidth = width/barnum;    lum_inc = ((float)(ymax-ymin))/((float)(barnum-1));    uv_width = width/2;    uv_height = height/2;    data_y = (unsigned char )malloc(widthheight);    data_u = (unsigned char )malloc(uv_widthuv_height);    data_v = (unsigned char )malloc(uv_widthuv_height);    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url_out, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url_out,str);    if((fp=fopen(input_url,"wb+")) == NULL){        LOGE("Error: Cannot create file!");        return -1;    }    //Output Info    LOGI("Y, U, V value from picture's left to right:\n");    for(t = 0;t < barnum;t++){        lum_temp = ymin+(char)(tlum_inc);        LOGI("%3d, 128, 128\n",lum_temp);    }    //Gen Data    for(j = 0;j < height;j++){        for(i = 0;i < width;i++){            t = i/barwidth;            lum_temp = ymin+(char)(tlum_inc);            data_y[jwidth+i] = lum_temp;        }    }    for(j = 0;j < uv_height;j++){        for(i = 0;i < uv_width;i++){            data_u[juv_width+i] = 128;        }    }    for(j = 0;j < uv_height;j++){        for(i = 0;i < uv_width;i++){            data_v[juv_width+i] = 128;        }    }    fwrite(data_y,widthheight,1,fp);    fwrite(data_u,uv_widthuv_height,1,fp);    fwrite(data_v,uv_widthuv_height,1,fp);    fclose(fp);    free(data_y);    free(data_u);    free(data_v);    return 0;}

函数调用AVUtils.getInstance().yuv420GrayBar(ROOT+"graybar_640x360_yuv420p.yuv",640, 360,0,255,10);生成的灰阶测试图如下：graybar_640x360_yuv420p.yuv7）计算两个YUV420P像素数据的PSNR

/  Calculate PSNR between 2 YUV420P file  @param jstr_url1     Location of first Input YUV file.  @param jstr_url2     Location of another Input YUV file.  @param w        Width of Input YUV file.  @param h        Height of Input YUV file.  @param num      Number of frames to process.  计算两个YUV420P像素数据的PSNR  PSNR是最基本的视频质量评价方法  程序计算后得到的PSNR取值为26.693。
PSNR取值通常情况下都在20-50的范围内，取值越高，代表两张图像越接近，反映出受损图像质量越好 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_yuv420Psnr(        JNIEnv env,jobject / this /,        jstring jstr_url1,jstring jstr_url2, jint w, jint h,jint num) {    const char str1 = env->GetStringUTFChars(jstr_url1, NULL);    FILE fp1 = fopen(str1,"rb+");    env->ReleaseStringUTFChars(jstr_url1,str1);    if(fp1 == NULL)        return  -1;    const char str2 = env->GetStringUTFChars(jstr_url2, NULL);    FILE fp2 = fopen(str2,"rb+");    env->ReleaseStringUTFChars(jstr_url2,str2);    if(fp2 == NULL)        return  -1;    unsigned char pic1 = (unsigned char )malloc(wh);    unsigned char pic2 = (unsigned char )malloc(wh);    for(int i = 0;i < num;i++){        fread(pic1,1,wh,fp1);        fread(pic2,1,wh,fp2);        double mse_sum = 0,mse = 0,psnr = 0;        for(int j = 0;j < wh;j++){            mse_sum += pow((double)(pic1[j]-pic2[j]),2);        }        mse = mse_sum/(wh);        psnr = 10log10(255.0255.0/mse);        LOGI("%5.3f\n",psnr);        fseek(fp1,wh/2,SEEK_CUR);        fseek(fp2,wh/2,SEEK_CUR);    }    free(pic1);    free(pic2);    fclose(fp1);    fclose(fp2);    return 0;}

函数调用AVUtils.getInstance().yuv420Psnr(ROOT+"lena_256x256_yuv420p.yuv",ROOT+"lena_distort_256x256_yuv420p.yuv",256,256,1);对于8bit量化的像素数据来说，PSNR的计算公式PSNR的计算公式原始图和受损图如下：程序计算后得到的PSNR取值为26.693。
PSNR取值通常情况下都在20-50的范围内，取值越高，代表两张图像越接近，反映出受损图像质量越好。
8）分离RGB24像素数据中的R、G、B分量

/  Split R, G, B planes in RGB24 file.  @param jstr_url  Location of Input RGB file.  @param w    Width of Input RGB file.  @param h    Height of Input RGB file.  @param num  Number of frames to process.  分离RGB24像素数据中的R、G、B分量  RGB24格式的每个像素的三个分量是连续存储的。
一帧宽高分别为w、h的RGB24图像一共占用wh3 Byte的存储空间。
  RGB24格式规定首先存储第一个像素的R、G、B，然后存储第二个像素的R、G、B…以此类推。
  类似于YUV420P的存储方式称为Planar方式，而类似于RGB24的存储方式称为Packed方式 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_rgb24Split(        JNIEnv env,jobject / this /,        jstring jstr_url, jint w, jint h,jint num) {    char input_url[100] = {0};    const char str = env->GetStringUTFChars(jstr_url, NULL);    sprintf(input_url,"%s",str);    env->ReleaseStringUTFChars(jstr_url,str);    FILE fp = fopen(input_url,"rb+");    if(fp == NULL) return -1;    strcpy(input_url+strlen(input_url)-4,"_r.y");    FILE fp1 = fopen(input_url,"wb+");    strcpy(input_url+strlen(input_url)-4,"_g.y");    FILE fp2 = fopen(input_url,"wb+");    strcpy(input_url+strlen(input_url)-4,"_b.y");    FILE fp3 = fopen(input_url,"wb+");    unsigned char pic=(unsigned char )malloc(wh3);    for(int i = 0;i < num;i++){        fread(pic,1,wh3,fp);        for(int j = 0;j < wh3;j = j+3){            //R            fwrite(pic+j,1,1,fp1);            //G            fwrite(pic+j+1,1,1,fp2);            //B            fwrite(pic+j+2,1,1,fp3);        }    }    free(pic);    fclose(fp);    fclose(fp1);    fclose(fp2);    fclose(fp3);    return 0;}

函数调用AVUtils.getInstance().rgb24Split(ROOT+"cie1931_500x500.rgb",500,500,1);运行上面代码后会在同目录下生成r、g、b三个文件

cie1931_500x500_r.y： R数据，分辨率为500x500cie1931_500x500_g.y：G数据，分辨率为500x500cie1931_500x500_b.y：B数据，分辨率为500x500

输入的原图是一张标准的CIE 1931色度图。
该色度图右下为红色，上方为绿色，左下为蓝色cie1931_500x500.rgb生成的r、g、b分量图片如下：cie1931_500x500_r.ycie1931_500x500_g.ycie1931_500x500_b.y9）将RGB24格式像素数据封装为BMP图像

/  Convert RGB24 file to BMP file  @param jstr_rgb24path    Location of input RGB file.  @param width        Width of input RGB file.  @param height       Height of input RGB file.  @param jstr_bmppath      Location of Output BMP file.  将RGB24格式像素数据封装为BMP图像  BMP采用的是小端（Little Endian）存储方式。
这种存储方式中“RGB24”格式的像素的分量存储的先后顺序为B、G、R。
由于RGB24格式存储的顺序是R、G、B，所以需要将“R”和“B”顺序作一个调换再进行存储  该程序完成了主要完成了两个工作：  1)将RGB数据素的“B”和“前面加上文件头。
  2)将RGB数据中每个像R”的位置互换。
   BMP文件是由BITMAPFILEHEADER、BITMAPINFOHEADER、RGB像素数据共3个部分构成，它的结构如下所示。
  BITMAPFILEHEADER  BITMAPINFOHEADER  RGB像素数据 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_rgb24ToBmp(        JNIEnv env,jobject / this /,        jstring jstr_rgb24path, jint width, jint height,jstring jstr_bmppath) {    typedef struct    {        long imageSize;        long blank;        long startPosition;    }BmpHead;    typedef struct    {        long  Length;        long  width;        long  height;        unsigned short  colorPlane;        unsigned short  bitColor;        long  zipFormat;        long  realSize;        long  xPels;        long  yPels;        long  colorUse;        long  colorImportant;    }InfoHead;    int i = 0,j = 0;    BmpHead m_BMPHeader={0};    InfoHead  m_BMPInfoHeader={0};    char bfType[2] = {'B','M'};    int header_size = sizeof(bfType)+sizeof(BmpHead)+sizeof(InfoHead);    unsigned char rgb24_buffer = NULL;    FILE fp_rgb24 = NULL,fp_bmp = NULL;    const char rgb24path = env->GetStringUTFChars(jstr_rgb24path, NULL);    if((fp_rgb24=fopen(rgb24path,"rb")) == NULL){        LOGE("Error: Cannot open input RGB24 file.\n");        env->ReleaseStringUTFChars(jstr_rgb24path,rgb24path);        return -1;    }    env->ReleaseStringUTFChars(jstr_rgb24path,rgb24path);    const char bmppath = env->GetStringUTFChars(jstr_bmppath, NULL);    if((fp_bmp=fopen(bmppath,"wb")) == NULL){        LOGE("Error: Cannot open output BMP file.\n");        env->ReleaseStringUTFChars(jstr_bmppath,bmppath);        return -1;    }    rgb24_buffer = (unsigned char )malloc(widthheight3);    fread(rgb24_buffer,1,widthheight3,fp_rgb24);    m_BMPHeader.imageSize = 3widthheight+header_size;    m_BMPHeader.startPosition = header_size;    m_BMPInfoHeader.Length = sizeof(InfoHead);    m_BMPInfoHeader.width = width;    //BMP storage pixel data in opposite direction of Y-axis (from bottom to top).    m_BMPInfoHeader.height = -height;    m_BMPInfoHeader.colorPlane = 1;    m_BMPInfoHeader.bitColor = 24;    m_BMPInfoHeader.realSize = 3widthheight;    fwrite(bfType,1,sizeof(bfType),fp_bmp);    fwrite(&m_BMPHeader,1,sizeof(m_BMPHeader),fp_bmp);    fwrite(&m_BMPInfoHeader,1,sizeof(m_BMPInfoHeader),fp_bmp);    //BMP save R1|G1|B1,R2|G2|B2 as B1|G1|R1,B2|G2|R2    //It saves pixel data in Little Endian    //So we change 'R' and 'B'    for(j = 0;j < height;j++){        for(i = 0;i < width;i++){            char temp = rgb24_buffer[(jwidth+i)3+2];            rgb24_buffer[(jwidth+i)3+2] = rgb24_buffer[(jwidth+i)3+0];            rgb24_buffer[(jwidth+i)3+0] = temp;        }    }    fwrite(rgb24_buffer,3widthheight,1,fp_bmp);    fclose(fp_rgb24);    fclose(fp_bmp);    free(rgb24_buffer);    LOGI("Finish generate %s!\n",bmppath);    env->ReleaseStringUTFChars(jstr_bmppath,bmppath);    return 0;}

函数调用AVUtils.getInstance().rgb24ToBmp(ROOT+"lena_256x256_rgb24.rgb",256,256,ROOT+"lena_256x256.bmp");原始图片lena_256x256_rgb24.rgb生成的bmp可以用自带看图软件打开：lena_256x256.bmp10）将RGB24格式像素数据转换为YUV420P格式像素数据

unsigned char clip_value(unsigned char x,unsigned char min_val,unsigned char  max_val){    if(x > max_val){        return max_val;    }else if(x < min_val){        return min_val;    }else{        return x;    }}//RGB to YUV420bool RGB24_TO_YUV420(unsigned char RgbBuf,int w,int h,unsigned char yuvBuf){    unsigned charptrY, ptrU, ptrV, ptrRGB;    LOGI("pic_rgb24 [ ");    ptrRGB = RgbBuf;    for(int ll = 0;ll < wh3;ll++){        LOGI("%c ",ptrRGB++);        if(ll%50 == 0){LOGI("\n");}    }    LOGI("]\n");    memset(yuvBuf,0,wh3/2);    ptrY = yuvBuf;    ptrU = yuvBuf + wh;    ptrV = ptrU + (wh1/4);    unsigned char y, u, v, r, g, b;    for (int j = 0; j < h;j++){        ptrRGB = RgbBuf + wj3 ;        //ptrRGB = RgbBuf + w  (h - 1 - j)  3 ;        for (int i = 0;i < w;i++){            r = (ptrRGB++);            g = (ptrRGB++);            b = (ptrRGB++);            /y = (unsigned char)( ( 66  r + 129  g +  25  b + 128) >> 8) + 16  ;            u = (unsigned char)( ( -38  r -  74  g + 112  b + 128) >> 8) + 128 ;            v = (unsigned char)( ( 112  r -  94  g -  18  b + 128) >> 8) + 128 ;/            y = (unsigned char)( ( 66  r + 129  g +  25  b) >> 8) + 16  ;            u = (unsigned char)( ( -38  r -  74  g + 112  b) >> 8) + 128 ;            v = (unsigned char)( ( 112  r -  94  g -  18  b) >> 8) + 128 ;            (ptrY++) = clip_value(y,0,255);            if (j%2==0 && i%2 ==0){                (ptrU++) = clip_value(u,0,255);            }            else{                if (i%2==0){                    (ptrV++) = clip_value(v,0,255);                }            }        }    }    return true;}/  Convert RGB24 file to YUV420P file  @param url_in  Location of Input RGB file.  @param w       Width of Input RGB file.  @param h       Height of Input RGB file.  @param num     Number of frames to process.  @param url_out Location of Output YUV file.  RGB24格式像素数据转换为YUV420P格式像素数据  RGB到YUV的转换公式：            Y = 0.257R' + 0.504G' + 0.098B' + 16            U = -0.148R' - 0.291G' + 0.439B' + 128            V = 0.439R' - 0.368G' - 0.071B' + 128  转换的过程中有以下几点需要注意：  1) RGB24存储方式是Packed，YUV420P存储方式是Planar。
  2) U，V在水平和垂直方向的取样数是Y的一半 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_rgb24ToYuv420(        JNIEnv env,jobject / this /,        jstring jstr_urlin, jint w, jint h, jint num,jstring jstr_urlout) {    const char url_in = env->GetStringUTFChars(jstr_urlin, NULL);    LOGI("url_in : %s\n",url_in);    FILE fp = fopen(url_in,"rb+");    env->ReleaseStringUTFChars(jstr_urlin,url_in);    if(fp == NULL){        return -1;    }    const char urlout = env->GetStringUTFChars(jstr_urlout, NULL);    LOGI("urlout : %s\n",urlout);    FILE fp1 = fopen(urlout,"wb+");    env->ReleaseStringUTFChars(jstr_urlout,urlout);    if(fp1 == NULL){        return -1;    }    unsigned char pic_rgb24 = (unsigned char )malloc(wh3);    unsigned char pic_yuv420 = (unsigned char )malloc(wh3/2);    for(int i = 0;i < num;i++){        fread(pic_rgb24,1,wh3,fp);        RGB24_TO_YUV420(pic_rgb24,w,h,pic_yuv420);        fwrite(pic_yuv420,1,wh3/2,fp1);    }    free(pic_rgb24);    free(pic_yuv420);    fclose(fp);    fclose(fp1);    return 0;}

函数调用AVUtils.getInstance().rgb24ToYuv420(ROOT+"lena_256x256_rgb24.rgb",256,256,1,ROOT+"lena_256x256_rgb24ToYuv420.yuv");原始图片lena_256x256_rgb24.rgb转换后的YUV420P格式的图片：lena_256x256_rgb24ToYuv420.yuv11）生成RGB24格式的彩条测试图

/  Generate RGB24 colorbar.  @param width    Width of Output RGB file.  @param height   Height of Output RGB file.  @param url_out  Location of Output RGB file.  生成RGB24格式的彩条测试图 /extern "C" JNIEXPORT jint JNICALLJava_com_xohn_ffmpeg_AVUtils_rgb24ColorBar(        JNIEnv env,jobject / this /,        jint width, jint height,jstring jstr_urlout) {    unsigned char data = NULL;    int barwidth;    FILE fp = NULL;    int i = 0,j = 0;    data = (unsigned char )malloc(widthheight3);    barwidth = width/8;    const char url_out = env->GetStringUTFChars(jstr_urlout, NULL);    fp = fopen(url_out,"wb+");    env->ReleaseStringUTFChars(jstr_urlout,url_out);    if(fp == NULL){        return -1;    }    for(j = 0;j < height;j++){        for(i = 0;i < width;i++){            int barnum = i/barwidth;            switch(barnum){                case 0:{                    data[(jwidth+i)3+0] = 255;                    data[(jwidth+i)3+1] = 255;                    data[(jwidth+i)3+2] = 255;                    break;                }                case 1:{                    data[(jwidth+i)3+0] = 255;                    data[(jwidth+i)3+1] = 255;                    data[(jwidth+i)3+2] = 0;                    break;                }                case 2:{                    data[(jwidth+i)3+0] = 0;                    data[(jwidth+i)3+1] = 255;                    data[(jwidth+i)3+2] = 255;                    break;                }                case 3:{                    data[(jwidth+i)3+0] = 0;                    data[(jwidth+i)3+1] = 255;                    data[(jwidth+i)3+2] = 0;                    break;                }                case 4:{                    data[(jwidth+i)3+0] = 255;                    data[(jwidth+i)3+1] = 0;                    data[(jwidth+i)3+2] = 255;                    break;                }                case 5:{                    data[(jwidth+i)3+0] = 255;                    data[(jwidth+i)3+1] = 0;                    data[(jwidth+i)3+2] = 0;                    break;                }                case 6:{                    data[(jwidth+i)3+0] = 0;                    data[(jwidth+i)3+1] = 0;                    data[(jwidth+i)3+2] = 255;                    break;                }                case 7:{                    data[(jwidth+i)3+0] = 0;                    data[(jwidth+i)3+1] = 0;                    data[(jwidth+i)3+2] = 0;                    break;                }            }        }    }    fwrite(data,widthheight3,1,fp);    fclose(fp);    free(data);    return 0;}

函数调用AVUtils.getInstance().rgb24ColorBar(640, 360,ROOT+"rgb24colorbar_640x360.rgb");生成“白黄青绿品红蓝黑”8种颜色的彩条，如下图：rgb24colorbar_640x360.rgb参考文章：雷神文章 https://blog.csdn.net/leixiaohua1020/article/details/50534150RGB到YUV的转换公式 https://blog.csdn.net/xiaoyafang123/article/details/82153279项目git地址：https://gitee.com/xohn/FFmpeg.git 中的FFMpeg1