In the quantitative analysis of PCR gelation images system

in order to detect PCR correctly and make more accurate diagnosis of pathological situation of the person detected

a method to extract information automatically from PCR images is presented in this paper. Firstly

Gaussian filter and horizontal projection of PCR images' gray information are used to resist the effect of non uniform distribution of brightness and gain the position of the pervasive PCR. Here

a key step is to select a adaptable parameter value for the Gaussian filter. Then

the method uses smooth filtering and vertical projection to reduce the influence caused by the non uniform and moving illegibility of the PCR bands

and then extract PCR band areas accurately. Finally

it makes use of local rectification to improve the precision of the enclosed band. In practice

this method is applied to the PCR image processing and promising results are obtained. The experiment result indicates that with the PCR image been detected accurately

subsequent diagnosis and analysis can achieve fairly efficiency

so the method is of very significant practicality.