Skip to content

Releases: igorrendulic/video-edge-ai-proxy

Added remote application install and bug fixes

01 Mar 22:29
Compare
Choose a tag to compare
  • Implementation of adding an application remotely from ChrysCloud
  • Improved device status change reporting (e.g. Camera offline) for Slack and Webhook notifications

Bug fixes and integration with ChrysCloud

09 Feb 22:23
Compare
Choose a tag to compare
  • Changed the LatestVideoImage interface to Unary gRPC call
  • Improved response times of LatestVideoImage to be ale to analyze even more connected cameras
  • Added In memory buffer probe (start, end and length of the in memory buffer)
  • MQTT integration for camera installation and removal from Chrysalis Cloud
  • Reporting monitoring to Chrysalis Cloud (memory, cpu, camera online, offline, ...)
  • Bug fixes

In-memory buffer, video probe, corresponding examples and object detection example

19 Jan 00:12
Compare
Choose a tag to compare

Implementation of in-memory buffer

Added gRPC methods to access in-memory buffer:
rpc VideoBufferedImage(VideoFrameBufferedRequest) returns (stream VideoFrame) {}

In memory buffer stores compressed video, packet by packet. It can be queried between two timestamps (from, to). To query the complete in memory buffer define: timestamp_from = 0 and timestamp_to = sys.maxsize

Added gRPC to return basic information about a video stream:
rpc VideoProbe(VideoProbeRequest) returns (VideoProbeResponse) {}

Returns VideoCodec object:

message VideoCodec {
    string name = 1;
    int32 width = 2;
    int32 height = 3;
    string pix_fmt = 4;
    bytes extradata = 5;
    int32 extradata_size = 6;
    string long_name = 7;
}

Added gRPC method to return system time:
rpc SystemTime(SystemTimeRequest) returns (SystemTimeResponse) {}

Returns current system time in the Chrysalis Edge Proxy.

Examples on how to use in-memory buffer

  • opencv_inmemory_display.py (single process consumer)
  • opencv_inmemory_display_advanced.py (multiple process consumer)
  • yolo_object_detection.py example how to do object detection in live video stream from RTSP camera

BugFix mp4 segment creation

18 Jan 23:30
Compare
Choose a tag to compare

Mainly fixing mp4 file segmenting.

v0.0.5

04 Nov 21:18
Compare
Choose a tag to compare
  • Web UI/UX bug fixes (error reporting on creation of new RTSP stream)
  • Manual upgrade of existing edgeproxy containers when new version available
  • Other minor bug fixes

Bug fixes and on disk circular buffer upgrade

02 Nov 22:50
ccb31bb
Compare
Choose a tag to compare
  • Few minor bug fixes
  • moved circular buffer management to individual running container, separating concerns to specific camera process instead of global management

v0.0.3

21 Sep 16:10
Compare
Choose a tag to compare
  • store key-frame segmented mp4 files to local disk
  • running mp4 segment cleanup cron job

v0.0.2

26 Aug 00:08
Compare
Choose a tag to compare
  • Added settings to store Chrysalis Edge Key/Secret
  • Added gRPC service to turn storage on/off on Chrysalis Cloud
  • Removed ProxyStart/ProxyStop methods and created a single method Proxy with on/off functionality
  • Added Streaming API connector for Computer Vision Annotations

v0.0.1-alpha

07 Aug 17:56
Compare
Choose a tag to compare
v0.0.1-alpha Pre-release
Pre-release

Initial alpha release. More testing required.