c++ - OpenCV 编译问题

Question

通过编译这个项目，我有几个问题:

由于“找不到文件”，我不得不更改标题
来自:

#include <cv.h>
#include <highgui.h>

到:

#include <opencv.hpp>
#include <highgui.hpp>

解决了包含问题但编译给出了更多错误:


我做错了什么？
下面的整个代码:

#include <stdio.h>
#include <unistd.h>
#include <opencv.hpp>
#include <highgui.hpp>

int main(int argc, char **argv)
{
    // get command line parameters
    if(argc < 6)
    {
        printf("Usage: %s <video> <vga window x> <vga window y> <vga width> <vga height>\n", argv[0]);
        printf("Usage: %s <camera id> <vga window x> <vga window y> <vga width> <vga height>\n", argv[0]);
        exit(0);
    }
    const char *filename = argv[1];
    int vgaX = atoi(argv[2]);
    int vgaY = atoi(argv[3]);
    int vgaWidth = atoi(argv[4]);
    int vgaHeight = atoi(argv[5]);
    int outputPixelCount = vgaWidth * vgaHeight;

    // opencv initializations
    CvCapture* cap = cvCaptureFromFile(filename);
    int isVideoFile = cap != 0;
    if (!cap)
        cap = cvCaptureFromCAM(atoi(filename));
    if (!cap)
    {
        printf("Could not open file/camera!\n");
        exit(1);
    }
    IplImage* frame = cvQueryFrame(cap); // get first frame for size
    if (!frame)
    {
        printf("The Video is empty!\n");
        cvReleaseCapture(&cap);
        exit(1);
    }
    IplImage* edges = cvCreateImage(cvGetSize(frame), IPL_DEPTH_8U, 1);
    IplImage* lines = cvCreateImage(cvGetSize(frame), IPL_DEPTH_8U, 3);
    IplImage* out = cvCreateImage(cvSize(vgaWidth, vgaHeight), IPL_DEPTH_8U, 3);
    int outStep = out->widthStep;
    int outChannels = out->nChannels;
    unsigned char *outData = (unsigned char*)out->imageData;

    // position windows // TODO: make debug output windows optional?
    cvNamedWindow("frame", CV_WINDOW_AUTOSIZE);
    cvMoveWindow("frame", 0, 32);
    cvNamedWindow("edges", CV_WINDOW_AUTOSIZE);
    cvMoveWindow("edges", frame->width, 32);
    cvNamedWindow("lines", CV_WINDOW_AUTOSIZE);
    cvMoveWindow("lines", 2 * frame->width, 32);
    cvNamedWindow("out", CV_WINDOW_AUTOSIZE);
    cvMoveWindow("out", vgaX, vgaY);

    while (42)
    {
        frame = cvQueryFrame(cap);
        if (!frame || (cvWaitKey(1) & 0xff) == 'q')
            break;
        cvShowImage("frame", frame);

        // edge detection
        cvCanny(frame, edges, 128.0, 130.0, 3); // TODO: tweakable parameters?
        cvShowImage("edges", edges);

        // get contours
        CvMemStorage *storage = cvCreateMemStorage(0);
        CvSeq *contours;
        int contourCount = cvFindContours(
            edges, storage, &contours, sizeof(CvContour),
            CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE, cvPoint(0, 0));
        cvZero(lines);
        cvDrawContours(lines, contours, cvScalar(32, 255, 32, 255), cvScalarAll(0), 100, 1, 8, cvPoint(0, 0));
        cvShowImage("lines", lines);

        // calculate total length over all contours
        float contourLengthSum = 0.0f;
        for(CvSeq *c = contours; c; c = c->h_next)
        {
            for(int i = 0; i < c->total - 1; i++)
            {
                CvPoint *p0 = CV_GET_SEQ_ELEM(CvPoint, c, i);
                CvPoint *p1 = CV_GET_SEQ_ELEM(CvPoint, c, i + 1);
                int dx = p1->x - p0->x;
                int dy = p1->y - p0->y;
                contourLengthSum += sqrtf(dx * dx + dy * dy);
            }
        }
        float factor = (float)outputPixelCount / contourLengthSum;

        // write output image
        int cx = 0, cy = 0;
        float xScale = 255.0f / frame->width;
        float yScale = 255.0f / frame->height;
        unsigned char *dp = &outData[1];
        for(CvSeq *c = contours; c; c = c->h_next)
        {
            CvPoint *p0 = CV_GET_SEQ_ELEM(CvPoint, c, 0);
            for(int i = 1; i < c->total; i++)
            {
                CvPoint *p1 = CV_GET_SEQ_ELEM(CvPoint, c, i);
                float x0x1 = p1->x - p0->x;
                float y0y1 = p1->y - p0->y;
                int n = (int)(sqrtf(x0x1 * x0x1 + y0y1 * y0y1) * factor);
                float x =          (float)p0->x * xScale;
                float y = 255.0f - (float)p0->y * yScale;
                float dt = 1.0f / (float)(n - 1);
                float dx = dt *  x0x1 * xScale;
                float dy = dt * -y0y1 * yScale;
                for (int j = 0; j < n; j++)
                {
                    dp[0] = (unsigned char)x;
                    dp[1] = (unsigned char)y;
                    x += dx;
                    y += dy;
                    dp += outChannels;
                    if (++cx == vgaWidth)
                    {
                        cx = 0;
                        dp = &outData[++cy * outStep + 1];
                        if (cy == vgaHeight)
                            goto full;
                    }
                }
                p0 = p1;
            }
        }

        // fill last few pixels with last pixel value, if there are any left
        for (; cy < vgaHeight; cy++)
        {
            for (;cx < vgaWidth; cx++)
            {
                outData[cy * outStep + cx * outChannels + 1] = 0;
                outData[cy * outStep + cx * outChannels + 2] = 0;
            }
            cx = 0;
        }

        full:
        cvReleaseMemStorage(&storage);
        cvShowImage("out", out);

        //if (isVideoFile)
        //  usleep(8000); // TODO: proper synchronization
    }
    cvReleaseImage(&out);
    cvReleaseImage(&lines);
    cvReleaseImage(&edges);
    cvReleaseCapture(&cap);
    return 0;
}

Answer 1

看起来您正在使用旧的 OpenCV API。它们被新方法取代。您的代码不适用于最近的 OpenCV 4.3.0。 cvCaptureFromFile , cvCaptureFromCAM和其他一些存在于 4.0.0-rc OpenCV 文档。

https://docs.opencv.org/4.0.0-rc/dd/d01/group__videoio__c.html

之后 4.0.0-rc OpenCV 的文档指向 videoio .

https://docs.opencv.org/master/dd/de7/group__videoio.html

进一步引用:

https://answers.opencv.org/question/55344/undeclared-indentifier-opencv-cvcapturefromcam-and-cvqueryframe/

此外，如果您在新的 OpenCV 版本中包含问题，请尝试，

#include <opencv2/opencv.hpp>
#include <opencv2/highgui.hpp>