Acquisition and Streaming - Multiple Streams

<< Acquisition and Streaming

Some devices may be able to provide multiple independent data streams with potentially different data types. An example for such a device is a multispectral device providing multiple streams transferring different spectral information or a device streaming a depth image and a RGB color image at the same time.

The only modification required to support multi-stream devices is to create and start all streams, then check on each stream for more data in a loop.

C++C#Python

C++

using namespace Cvb;
 
auto devices = Cvb::DeviceFactory::Discover(DiscoverFlags::IgnoreVins);
auto device = DeviceFactory::Open<GenICamDevice>(devices[0].AccessToken(), AcquisitionStack::GenTL);
 
std::vector<ImageStreamPtr> streams;
for (int i = 0; i < device->StreamCount(); ++i) {
  streams.push_back(device->Stream<ImageStream>(i)) // (1)
}
 
 
 
for (int i = 0; i < streams.size(); ++i) {
  streams[i]->Start(); // (2)
}
 
int streamIndex = 0;
while(true)
{
  std::chrono::milliseconds TIMEOUT(3000);
 
  ImagePtr image;
  WaitStatus status;
  NodeMapEnumerator nodeMaps;
  std::tie(image, status, nodeMaps) = streams[streamIndex]->WaitFor(TIMEOUT); // (3)
  streamIndex = (streamIndex + 1) % streams.size();
 
  if (status == WaitStatus::Timeout) {
    // No data on this stream, try the next stream
    continue;
  }
 
  //Process image
}
 
for (int i = 0; i < streams.size(); ++i) {
  streams[i]->Abort(); // (4)
}

C#

using Stemmer.Cvb;
using Stemmer.Cvb.Driver;
 
using (var devices = DeviceFactory.Discover(DiscoverFlags.IgnoreVins))
{
  using (var device = (GenICamDevice)DeviceFactory.Open(devices[0], AcquisitionStack.GenTL))
  {
    var streams = new List<ImageStream>;
    for (int i = 0; i < device.StreamCount(); ++i) {
      streams.Add(device->GetStream<ImageStream>(i)) // (1)
    }
 
    foreach (ImageStream stream in streams) {
      stream.Start(); // (2)
    }
 
    int streamIndex = 0;
    while(true) {
      UsTimeSpan TIMEOUT(3000);
 
      WaitStatus status;
      using (var image = streams[streamIndex].WaitFor(TIMEOUT, out status)) // (3)
      {
        streamIndex = (streamIndex + 1) % streams.size();
        if (status = WaitStatus.Timeout) {
          continue;
        }
      
        //Process image
      }
    }
 
    foreach (ImageStream stream in streams) {
      stream.Abort(); // (4)
    }
  }
}

Python

import cvb
 
devices = cvb.DeviceFactory.discover_from_root(cvb.DiscoverFlags.IgnoreVins)
with cvb.DeviceFactory.open(devices[0].access_token, cvb.AcquisitionStack.GenTL) as device:
 
    streams = []
    for i in range(0, device.StreamCount()):
      streams.append(device.stream(cvb.ImageStream, i)) # (1)
 
    for stream in streams:
      stream.start() # (2)
 
    streamIndex = 0
    while True:
      TIMEOUT = 3000
      image, status, node_maps = streams[streamIndex].wait(TIMEOUT) # (3)
      streamIndex = (streamIndex + 1) % streams.size()
 
      if status == WaitStatus.Timeout:
        continue
       
      # Process image
 
    for stream in streams:
      stream.abort() # (4)

Note: Error handling has been omitted from the above example.

Note: Waiting for data is sequential for simplicity; using concurrent waits (e.g. using async or thread) is more likely.

1. All data streams are instantiated. Note: The order does not matter.
2. Acquisition is started for all data streams. Note: The order does not matter.
3. Each composite will need to be proactively waited for by means of a call to wait on the stream. In this case, to prevent blocking indefinitely if one stream does not have data, the Cvb::Stream::WaitFor with timeout is used. If the status is Cvb::WaitStatus::Timeout, no data was received and the next stream can checked.
4. All streams are aborted. Note: The order does not matter.