discussed after the factors that contribute to the cor- 
relator performance are described. 
Table I lists the significant performance factors of 
the two correlators. They include bandwidth, internal 
and external noise-sources, input power, and dynamic 
range. Here it should be mentioned briefly that the 
performance of certain components of the electronic 
flying-spot scanner system degrades with use; speci- 
fically, the CRT loses brightness, and the phosphor 
noise increases. These components must be replaced 
periodically. Such degradation effects are not present 
in the optical system. 
The bandwidths of the optical and the electronic 
correlators are defined in terms of the effective aper- 
ture of the optics; in the case of the electronic corre- 
lator, they are defined also in terms of size of the 
scanner CRT spot. Here we will assume that the elec- 
tronic circuits have sufficient bandwidth to handle 
the video signals from the flying-spot scanner system. 
The graphs in Figure 11 show how the optics and 
CRT spot affect the system bandwidth in both cases. 
The bandpass of the coherent optical correlator is flat 
out to a spatial frequency that is related directly to 
the f-number of the lens and the wavelength A of the 
light used. This cutoff spatial frequency is 1/2 Af, or 
280 line-pairs-per-millimeter for a f/2.8 lens and 
0.6328 10-3 mm laser light. The bandwidth of the 
flying-spot scanner system is related directly to the 
size of the CRT spot, the minification of the transfer 
lens system, and the modulation transfer function of 
the transfer lens. The transfer function of a diffrac- 
tion-limited f/2.8 lens is illustrated in the graph along 
with the effective bandpass of the CRT spot. The 
CRT spot is assumed to have a Gaussian shape with 
an effective diameter of 50 microns, defined as the 
distance between the half-amplitude points. The 
transfer lens is assumed to be working at a minifica- 
tion of 20. The curve labeled as the product of the 
lens and spot transfer functions is the bandpass of the 
electronic correlator. A comparison of the two band- 
Table | Correlator Performance Factors 
Optical Electronic 
Bandwidth Bo-1;|v| $ 4A) BE = EXP [v2 2 (WM)? 
Bo =0;|v|> CAD 
Internal Noise None 1. CRT phosphor 
grain noise 
2. CRT phosphor 
wear noise 
3. PMT shot noise 
1. PMT shot noise 
2. Correlator noise 
External Noise Correlator noise 
Light Source Power Laser-unlimited Phosphor-limited 
THERMAL MAPPING 
1.0 
  
  
  
  
  
N Coherent Light 
0.9 £— s Lens Bandpass 
08 A (Amplitude) 
.8 fo \ 
0.7 = A Gaussian CRT Spot 
À L—— Transfer Function 
0.6 + X 
\ 
05 +— à Lens Modulation Transfer Function 
; X — (Intensity) 
0.4 + 
03 = p : Ty One-Channel 
3 EE A Flying Spot Scanner 
: x Response 
TOM 
N 
m > 
0m NS 
200 300 400 500 600 700 800 900 
  
Spatial Frequency — Line Pairs per Millimeter 1 
Mno 
Figure 11 Optical and Electronic System 
Bandwidth Comparison 
widths indicates that, for a given lens, the bandpass of 
the optical system surpasses that of the flying-spot 
scanner. If the spot could be made infinitely small, 
the scanner bandwidth would be limited by the lens 
modulation transfer function. 
Since the two correlators can be treated as measur- 
ing systems, their ultimate accuracy is a function of 
system noise. The sources of noise can be categorized 
into three groups: physical, electronic, and correlator, 
ignoring the photographic grain noise since it is a part 
of the input that cannot be controlled. The flying- 
spot scanner correlator has all three sources of noise. 
The CRT phosphor grain and, after a time, the wear 
pattern in the phosphor are the physical sources of 
noise. As the scanning spot moves over the phosphor, 
its intensity fluctuates in proportion to the size of the 
grain in the path of the scanning electron beam. The 
rms fluctuation of the intensity can be as high as 30% 
in some CRT tubes. In fine-grain micro-spot CRT 
tubes, the rms intensity fluctuation is around 10%. 
Since this noise has a high-frequency spectrum, it can 
usually be filtered out in the video amplifier without 
degrading the performance of the correlator. In sys- 
tems designed to work in the megacycle range, how- 
ever, phosphor noise is the predominant noise in the 
scanner. 
The next source of noise that affects both the 
optical and the electronic systems is the shot noise 
generated by the light signal in the phototubes. The 
shot noise in the electronic correlator is greater than 
that in the optical correlator because it. is a function 
of the total light hitting the phototubes. In each scan- 
ner, the light hitting the tubes is composed of a signal 
and dc component due to the average image transmit- 
tance. The dc component is usually large and, there- 
fore, causes a proportionally large shot noise, reduc- 
ing the signal-to-noise performance of the scanner. In 
the optical correlator, the light detected by the 
phototube is all signal, therefore, the shot noise is 
only a function of the signal strength. 
79