Method for processing faulty wire of X-ray flat panel detector image and device thereof

Abstract

The invention provides a method for processing a faulty wire of an x-ray flat panel detector image and a device thereof. The method comprises the following steps of: detecting the position of the faulty wire in the x-ray flat panel detector image; in the x ray flat panel detector image to be processed, estimating detailed information of image pixels at the faulty wire by using image data in faulty wire-contained image sub-area; and in the image after the detail estimation, calculating compensation amount by using the image data of the image pixels of a non-faulty wire in the neighborhood of the faulty wire and compensating the image data of the image pixels at the faulty wire. Due to the adoption of the method, the faulty wire can be automatically identified and the position of the faulty wire can be saved through a software method by using the gray scale abnormity at the faulty wire, and then grey scale compensation is performed on the image detailed information at the faulty wire by using normal pixels (namely the non-faulty wire image pixels) in the neighbourhood of the faulty wire, so the image gray scale at the faulty wire is naturally transited.

Description

technical field [0001] The invention relates to an image analysis technology of an X-ray flat panel detector, in particular to a method and a device for processing bad lines of an X-ray flat panel detector image. Background technique [0002] At present, the X-ray flat panel detector has many advantages compared with the previous film imaging, such as high density resolution, easy storage, flexible post-processing, etc., but due to its complicated manufacturing process, the original image collected from the detector generally cannot meet the requirements actual testing requirements. The output of different pixels on the detector is not consistent under the same X-ray dose radiation. The reasons for this phenomenon include random noise, offset error, inconsistent response of pixels, bad pixels and so on. Except for bad pixels, the output of other pixels contains some useful information, which can be adjusted to the normal gray value through gray compensation; while the outpu...

AI Technical Summary

Problems solved by technology

Therefore, in order to obtain a good interpolation effect, it is necessary to iteratively perform Fourier transform and inverse transform multiple times, and this Fourier transform is a relatively time-consuming calculation method, so this kind of interpolation generally cannot meet the real-time requirements in terms of computing time

It can be seen that the details of the image correspond to the high-frequency information of the image, and the frequency domain processing can effectively control the number of details, but the frequency domain interpolation method has two main disadvantages: one is slow; the other is that the ideal image bandwidth needs to be known

The reason for the slow speed is that for the band-limited method, multiple iterations of Fourier transform / inverse transform are required to effectively compensate for the gray information at the bad line

When the image size is very large, the amount of calculation cannot meet the real-time requirements

For example, for an image with a size of 2048*2048, a fast Fourier transform takes a second-level processing time on a general PC. If it is performed multiple times, and considering that there are many bad lines in the image, the processing time will be longer

In addition, the frequency band limit here requires knowing the bandwidth of the image, otherwise the image details may be lost, which is also difficult to achieve in practical applications, because the processing process does not know what the object being photographed is

Images