search

amazon-archives / aws-robomaker-sample-application-persondetection

Use AWS RoboMaker and demonstrate the use of Amazon Rekognition to recognize people's faces and Amazon Polly to synthesize speech.
MIT No Attribution
25 stars 18 forks source link

Use launch files for all AWS Cloud Extensions #4

Closed timrobotson closed 5 years ago

timrobotson commented 5 years ago

Instead of running the cloud extension nodes directly, include their launch files. This is the recommended way to use them in an application.

To ensure nothing changed I saved the list of PARAMETERS displayed when running roslaunch person_detection_robot person_detection.launch for both master and this branch and compared them and they are identical.

This PR requires https://github.com/aws-robotics/kinesisvideo-encoder-ros1/pull/4 to be merged first to allow the user to change the image transport type. After it is merged it will need to be tagged and then the ~/robot_ws/.rosinstall file in this sample application must be updated.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

samuelgundry commented 5 years ago

Ah, Tim pointed out the namespace was implicit before too. kinesis_video_streamer.launch[1] specifies the namespace that PersonDetection needs to be consistent with for its stream name rosparams. I'm okay with these being a separate PR.

[1] https://github.com/aws-robotics/kinesisvideo-ros1/blob/master/kinesis_video_streamer/launch/kinesis_video_streamer.launch#L25

timrobotson commented 5 years ago

I tested this by changing the .rosinstall file to use this branch for the to https://github.com/aws-robotics/kinesisvideo-encoder-ros1/pull/4, then building and running it in RoboMaker. The Kinesis video stream worked correctly and Rekognition detected faces correctly.