You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Proceedings of the Symposium on Progress in Data Processing and Analysis

and being rather uniform on the picture is shown in Fig. 2.
— -*-S, A4
It is seen that along with the decrease in the radio
brightness temperature the fragment of the record corres
ponding to the boggy terrain is, on average, characterized
by stronger fluctuations of the signal.
If we use the above-described approach, the probabi
lity of the correct identification of the boundary of such
objects shall be low because of low radio-brightness con
trast comparable with the fluctuations of the radio-ther
mal signal within the limits of the uniform object. More
reliable results are obtained in the cases where the mean
squared deviation of the radio-brightness temperature is
taken as an informative parameter. The corresponding
assessment is sampled and the samples are independent.
Let us denote the mean squared deviations of radio
brightness temperatures of different objects ascf,, andcf^.
For increasing the accuracy of identification it is expe
dient to use also the sampling mean of the radio-bright
ness temperatures. For instance, the decision function can
be determined as follows
Te K
Fig. 2. Registergram
of radio-thermal
signal of boggy
24 0
where A - A",| T 9 , - Taal + K £ |<3" i -cr a | > K, and
j- ( Ts ,G) = K, IT„ - T^jt K ¿ I, consta:
and K 2 are
constants ;
T* - is the mean value of the radio-brightness tem
perature. Indices 1, 2 denote the objects of
the first and second types, respectively.