问题比比皆是：连接ip摄像头，即覆盆子pi摄像头，从覆盆子pi流到我的电脑 . 利用openCV;试图进行光学跟踪 . 我在网上发现了一些我已经修改过的开源代码，它在C中，而之前我在Python中编写了更多的OSource .

所以这里是工作的块，通知字节被解析为python中的jpg图像 . 记下urllib .

import cv2
import urllib
import numpy as np

stream=urllib.urlopen('http://<ip>:<port>/?action=stream')
bytes=''
while True:
    bytes+=stream.read(1024)
    a = bytes.find('\xff\xd8')
    b = bytes.find('\xff\xd9')
    if a!=-1 and b!=-1:
        jpg = bytes[a:b+2]
        bytes= bytes[b+2:]
        i = cv2.imdecode(np.fromstring(jpg, dtype=np.uint8),cv2.CV_LOAD_IMAGE_COLOR)
        cv2.imshow('Robo-View',i)
        if cv2.waitKey(1) == 27:
            exit(0)

现在对于C代码片段：我相信从已分配的IP地址捕获流 . 我需要找到\ xff \ xd8＆\ xff \ xd9来识别jpg帧 .

int main(int argc, char* argv[])
{
    //some boolean variables for different functionality within this
    //program
    bool trackObjects = true;
    bool useMorphOps = true;
    calibrationMode = true;
    //Matrix to store each frame of the webcam feed
    Mat cameraFeed;
    //matrix storage for HSV image
    Mat HSV;
    //matrix storage for binary threshold image
    Mat threshold;
    //x and y values for the location of the object
    int x = 0, y = 0;
    //video capture object to acquire webcam feed
    VideoCapture capture;
    //Video located at this address: 
    const std::string vidStreamAddress = "http://<ip>:<port>/?action=stream";
    //open capture object at vidstreamaddress
    capture.open(vidStreamAddress);
    //set height and width of capture frame
    capture.set(CV_CAP_PROP_FRAME_WIDTH, FRAME_WIDTH);
    capture.set(CV_CAP_PROP_FRAME_HEIGHT, FRAME_HEIGHT);

帮助将不胜感激 . 在识别解析片段的位置或者是否有可用的方法时 . 谢谢

python c++ opencv