JUNENG GLASS ENGINEERS, INC.
|
Automatically Control Tweel with Digital
Image Analysis
The
forming of the float glass is conducted in the high-temperature tin bath. Molten
glass is fed into the tin bath through a canal which links the melting tank and
the tin bath. Glass flow volume is regulated by a tweel on the canal. Upon
entering the float tin bath, molten glass spreads out on the surface of liquid
tin and forms an “onion head”. Final glass ribbon is drawn from this onion
head. Because
of the difficulty in measuring the flow volume of the high-temperature glass
stream with normal methods, operators observe the width change of
the glass “ onion” via cameras, and then manually control the extent
of opening of the canal tweel to adjust the flow volume of the glass stream. In
this way, operators are likely to get tired and inadvertent. To solve this
problem, we have developed the Digital Image Analysis System(DIAS) which can control
the canal tweel and the fed volume of the glass stream automatically. The system
observes the edge positions of the glass onion via cameras, digitalizes the
observed image, and instructs the tweel mechanism in real time. 2.
Principle Figure 1 is the view of the hot end of a tin bath. In Figure 1, We can easily obtain: L
= H + S - SR - S L Where: L
— the width of the glass Onion H
— the central optic axis distance of the left and the right camera S
— visual field’s width of the camera
SR — departure of the onion’s right edge from the edge of
the visual field
SL — departure of the onion’s left edge from the edge of
the visual field
we
can infer: If
the distances from the corresponding image point of the bright spot on the left
and right plate edges projecting to the photoconductive elements of the camera to the border image point are respectively X L and XR, the
magnification coefficient of lens is K, then: SR
=K×XR
S
L =K×
X L To
substitute K×XR
and K×
X L for SR and S L in the formula (1):
L=H+S-K ×
XR- K ×X
L (2)
Since
the width of image points on the CCD photoconductive elements are fixed value
determined by the specifications of the camera, and the magnification
coefficient K and the width of the visual field are also fixed values if
specifications and focus of lens and visual angle of camera are decided, if A=H+S,
then formula (2) could be: L=A-K(XR
+ X L)
(3) Formula
(3) is for calculating the onion width. K and A are constants and can be
measured and set during the installation and commissioning. XR and X
L are variable according to onion edge. As soon as the computer image
processing system recognizes the position of the corresponding image point of
the bright spot of the glass edge, the values of X L and XR
could be worked out. And then the real width of the glass onion will be worked out.
The structure of the system is as the following chart 2. Its components and
functions are as follows: Two
CCD black and white camera, GDS-16 aperture lens,
high-temperature protective water jacket, used to take a photograph of the two
edges of the glass onion. A
computer with display, keyboard, hard disk and floppy drive. It will be used
to recognize the edge and calculate the width. A
special picture-collecting control board inserted into the expansion slot of the
computer. It will be used for sampling picture’s signals, quantification,
synchronic control and storing. A
special relay output control board inserted into the expansion slot of the
computer. It will be used for magnifying and outputting control signals. A
14”
RGB monitor used to monitor the actuality on the spot.
|
|