Analysis of the interference line and noise (Java) in the removal of the captcha image

(original article, reprint please add reprint address) Copyright notice: When reproduced, please indicate the original source of the article and the author's information and this statement in the form of a hyperlink

The most important thing in image processing is the feature analysis of the target image, and the image preprocessing method is designed through these features (points). For the verification code with interference line, noise characteristics of the picture, naturally have its own processing method, the following is a bit of personal humble opinion, image processing is commonly used in MATLAB and C + + to do, because they have a strong image processing library, in Java about the image processing is relatively less.

Picture pixels by the 24-bit binary machine code representation, can be expressed as ARGB, here and color RGB does not conflict, here A is the transparency, there are a lot of on-line on this aspect of the exposition, here do not repeat, I want to do here is about the pure background color, the foreground color does not need to do the verification code picture. The method is based on pixel-level operation, the foreground and background content on the pixel also have certain requirements, the interference line than the foreground content in the stroke to be thin, this is very important, is the main essence of the method, through the operation of the pixel can be achieved to remove the interference line and noise effect, Believe that someone who knows a bit of image processing should already know what's going on.

These are just a little bit of my original humble opinion, out of interest in the image processing, wrote a deal with this picture of the Java Small Project, I hope the great god many advice, thank you, leaving qq:484538927

There are many friends asked how to identify these verification code pictures, in fact, to identify these verification code pictures, the most difficult is the verification code image processing without any interference of two value image (black and white figure, this step is the focus, key, but also difficult), and then use the image segmentation to cut a single character, make a quantitative font, will be able to achieve the purpose of recognition.

Nonsense not to say, there is a picture of the truth, the following are their own processed pictures and:

Sample diagram:

After the treatment:

Analysis of the interference line and noise (Java) in the removal of the captcha image