First pass at experimental code to split and threshold image.

Image captured by webcam is split into three components, and the green channel is
thresholded to identify the green LEDs.

detector.py

@@ -0,0 +1,36 @@
+import cv2 as cv
+
+cv.namedWindow("preview")
+vc = cv.VideoCapture(-1)
+
+if vc: # try to get the first frame
+    vc.set(cv.cv.CV_CAP_PROP_FRAME_WIDTH, 640)
+    vc.set(cv.cv.CV_CAP_PROP_FRAME_HEIGHT, 480)
+    retval, frame = vc.read()
+
+while retval:
+    # Split the frame into components.
+    green_img = cv.split(frame)[1]
+
+    # Find the maximum pixel value in the green image.
+    [minVal, maxVal, minLoc, maxLoc] = cv.minMaxLoc(green_img)
+
+    # Don't show anything below this fraction of the max pixel value. Trial and error.
+    margin = 0.92
+
+    # Threshold value in pix value to be extracted
+    thresh = int( maxVal * margin)
+
+    # Do the thresholding.
+    retval, thresh_img = cv.threshold(green_img, thresh, 255, cv.cv.CV_THRESH_BINARY)
+
+    # Show the result
+    cv.imshow("preview", thresh_img)
+
+    # Get the next frame
+    retval, frame = vc.read()
+
+    key = cv.waitKey(20)
+    if key != -1: # exit on any key
+        break
+    # Loop forever...