Node.js的 - MJPEG TCP流为base64图片 [英] Node.js - MJPEG TCP stream to base64 images

问题描述

根据 paparazzo.js lib中，我试图从MJPEG得到的base64图片流（流过在GStreamer TCP）在Node.js的服务器上，并通过WebSockets向它们发送到客户端。

Based on paparazzo.js lib, I'm trying to get base64 images from a MJPEG stream (streamed over TCP with GStreamer) in a Node.js server, and to send them to the clients via websockets.

我想我是pretty接近，但我的图片已损坏。这里是code我使用的是：

I think I'm pretty close, but my images are corrupted. Here is the code I'm using :

var boundary = "----videoboundary";
var data = "";

var tcpServer = net.createServer(function (socket) {

    socket.on('data', function(chunk) {

        var boundaryIndex = chunk.toString().indexOf(boundary);

        if (boundaryIndex !== -1) {

            // Get the image's last piece of data :
            data += chunk.toString().substring(0, boundaryIndex);

            // Convert the data to a base64 image and broadcast it :
            var image = new Buffer(data).toString('base64');
            io.sockets.emit('image', image);

            // Reset the data :
            data = '';

            // Get the remaining data (with new image's headers) :
            var remaining = chunk.toString().substring(boundaryIndex);

            // Remove the new image's headers and add the remaining data :
            var contentTypeMatches   = remaining.match(/Content-Type:\s+image\/jpeg\s+/);
            var contentLengthMatches = remaining.match(/Content-Length:\s+(\d+)\s+/);

            if(contentLengthMatches != null && contentLengthMatches.length > 1) {

                var newImageBeginning = remaining.indexOf(contentLengthMatches[0]) + contentLengthMatches[0].length;
                data += remaining.substring(newImageBeginning);
            }
            else if(contentTypeMatches != null) {

                var newImageBeginning = remaining.indexOf(contentTypeMatches[0]) + contentTypeMatches[0].length;
                data += remaining.substring(newImageBeginning);
            }
            else {

                var newImageBeginning = boundaryIndex + boundary.length;
                data += remaining.substring(newImageBeginning);

                io.sockets.emit('error', { message: 'Could not find beginning of next image' });
            }
        }
        else {
            data += chunk;
        }
    });
});

任何想法？

感谢

推荐答案

chunk.toString（）二进制缓冲区转换为UTF8-CN codeD字符串（默认情况下），因此对于二进制图像数据可能会导致你的一些问题。

chunk.toString() converts the binary Buffer to a utf8-encoded string (by default), so for binary image data that will probably cause you some problems.

这可能有助于简化事情对你另一种选择是使用迪塞尔模块。就这样，你的code可能看起来像：

Another option that might help simplify things for you is to use the dicer module. With that, your code may look like:

var Dicer = require('dicer');
var boundary = '----videoboundary';

var tcpServer = net.createServer(function(socket) {
    var dice = new Dicer({ boundary: boundary });

    dice.on('part', function(part) {
      var frameEncoded = '';
      part.setEncoding('base64');
      part.on('header', function(header) {
        // here you can verify content-type, content-length, or any other header
        // values if you need to
      }).on('data', function(data) {
        frameEncoded += data;
      }).on('end', function() {
        io.sockets.emit('image', frameEncoded);
      });
    }).on('finish', function() {
      console.log('End of parts');
    });
    socket.pipe(dice);
});

这篇关于Node.js的 - MJPEG TCP流为base64图片的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！