github ↗

The eyes of the distributed chess system. CameraServer opens a webcam and streams what it sees over the network — the board, the players, any feed — so the visual data can flow into the neural layers downstream.

It is a small Java socket server. WebcamServer opens the camera and serves frames over TCP; a thin WebcamClient connects and receives them. From here the pixels feed the SOM layers (Test3 → Test4), or trainingGround's neural processing layer for interpretation.

how it works

Frames are grabbed with OpenCV (VideoCapture) and pushed as encoded images over a socket (port 5000, with a second stream on 5050); the client just opens a Socket to the host and reads the stream. Running it needs the OpenCV Java bindings plus the native library on the path.

compile & run
# build — needs OpenCV (org.opencv) jar + native lib on the path mvn compile # run the server, then the client in a second terminal java -cp target/classes cam.WebcamServer java -cp target/classes cam.WebcamClient
camera sockets vision server java
active
01 screenshot
02 screenshot
03 screenshot

1. e4light steps forward and the eye opens — nothing is computed until something is first seen.

Sxf?the scout captures the entire board and removes not a single piece.

0-0?!I never move; I only shift the lens — and a truth you weren't watching delivers a discovered check.

∞ on the long diagonalthe open file belongs to whoever is looking down it.