A film interference point erasing method

Abstract

The present invention provides a method for erasing interference points on a film, including S1), using a scanner to scan the film to obtain a color image with three RGB channels, each pixel point and each channel has a corresponding pixel value; S2), after the film is irradiated with radiation, the color of the coating sensitive to radiation will change. If there is no interference point in the film, the pixel values ​​of all points on the film will be distributed as a curve with a certain width in the RGB space; if it is not on the curve, it is an interference point; S3), after the interference point is detected, the inpaint function of opencv is used to take the interference point as a mask, and then the pixel value of the interference point position is calculated by the pixel value of the normal point near the interference point to achieve the erasing of the interference point. The present invention uses an abnormal point detection algorithm to detect interference points or a scale and a dose response curve to determine interference points, and accurately determines interference points by determining whether they are on a straight line to avoid accidental deletion.

Description

Technical Field

[0001] The invention relates to the field of medical technology, in particular to a method for erasing film interference points. Background Art

[0002] In the fields of non-destructive detection, medical diagnosis and treatment, radiotherapy quality control, etc., it is necessary to irradiate the film with radiation and analyze the irradiated film. In actual operation, words or marks are often made on the film, and the film (especially the edges) is inevitably stained, which brings a lot of interference to the subsequent film analysis. The traditional processing methods are as follows:

[0003] 1. Filtering, such as median filtering, can remove scattered abnormal points, but it is powerless for blocky stains or words.

[0004] 2. Threshold: Set a pixel threshold and remove points that exceed the threshold. This method can remove interference points with particularly dark colors, but the threshold must be adjusted for different films and different interferences. When the normal signal value itself is relatively large, it is easy to be deleted by mistake, and it cannot process points with lighter colors. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method for erasing film interference points.

[0006] The technical solution of the present invention is: a method for erasing film interference points, comprising the following steps:

[0007] S1), using a scanner to scan the film to obtain a color image with three channels of RGB, each pixel point and each channel has a corresponding pixel value p (R, G, B);

[0008] S2) After the film is exposed to radiation, the color of the radiation-sensitive coating will change, which is expressed as R(D), G(D), B(D), where RGB is the three channels of red, green and blue, and D is the dose size. The larger the dose D, the darker the color;

[0009] If there are no interference points in the film, due to noise, the pixel values ​​of all points on the film will be distributed as a curve with a certain width in the RGB space; if they are not on this curve, they are interference points;

[0010] S3), after the interference point is detected, the inpaint function of opencv is used to take the interference point as a mask, and then the pixel value of the interference point position is inferred from the pixel value of the normal point near the interference point, so as to achieve the wiping of the interference point.

[0011] Preferably, in step S2), if no scale is set on the film, since the number of interference points is less than the number of normal points, the interference points are detected by an abnormal point detection algorithm or a classification algorithm.

[0012] Preferably, in step S2), if a scale is set on the film, that is, by setting a corresponding scale for the pixel value and dose of the film, the scale on the film and the dose response curve are used:

[0013]

[0014] Among them, k refers to the RGB three channels, a, b, c are fitting parameters, OD is the optical density, Among them I 0 is the incident light intensity, I t is the intensity of transmitted light. When scanning, pay attention to the effective range of the scanner light intensity. If the range is exceeded, the pixel value will be saturated. Fit a curve that the normal signal points on the film satisfy, and use points outside a certain range of the curve as interference points.

[0015] Preferably, the method for detecting interference points by the outlier detection algorithm is the IsolationForest algorithm of scipy, which is performed by inputting all the pixels of the film as three-dimensional data points.

[0016] Preferably, all pixel values ​​of the film are projected relative to the dose response curve, and the pixel points on the curve are projected onto the central axis of the curve, thereby achieving denoising of the image.

[0017] The beneficial effects of the present invention are:

[0018] 1. The present invention uses a scanner to scan the film to obtain a color image with three channels of RGB, and then uses the scale on the film and the dose response curve to determine the interference point. By determining whether it is on a straight line, the interference point can be accurately determined to avoid accidental deletion;

[0019] 2. The present invention can also use an outlier detection algorithm or a classification algorithm to detect interference points. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the structure of the point distribution on the film of the present invention; DETAILED DESCRIPTION

[0021] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings:

[0022] The present invention provides a method for erasing film interference points, comprising the following steps:

[0023] S1), using a scanner to scan the film to obtain a color image with three channels of RGB, each pixel point and each channel has a corresponding pixel value p (R, G, B);

[0024] S2) After the film is exposed to radiation, the color of the radiation-sensitive coating will change. The degree of change affected by the radiation dose is expressed as R(D), G(D), and B(D), where RGB is the red, green, and blue channels, and D is the dose size. The larger the dose D, the darker the color.

[0025] If there are no interference points in the film, due to noise, the pixel values ​​of all points on the film will be distributed in the RGB space as a curve with a certain width; if they are not on the curve, they are interference points; in this embodiment, since it is impossible for all points to fall exactly on the curve, the curve of this embodiment has a certain width, such as Figure 1 As shown, therefore, interference points will not fall on this curve and their color will not be particularly dark.

[0026] S3), after the interference point is detected, the inpaint function of opencv is used to take the interference point as a mask, and then the pixel value of the interference point position is inferred from the pixel value of the normal point near the interference point, so as to achieve the wiping of the interference point.

[0027] As a preferred embodiment of this invention, in step S2), there are two methods for detecting interference points, wherein:

[0028] If there is no scale on the film, since the number of interference points is less than the number of normal points, the interference points are detected by an abnormal point detection algorithm or a classification algorithm. In this embodiment, the method used is the IsolationForest algorithm of scipy, which can be input by taking all the pixel points of the film as three-dimensional data points.

[0029] As a preferred embodiment of this invention, in step S2), if a scale is set on the film, that is, a scale corresponding to the pixel value and the dose of the film is set, and the scale on the film and the dose response curve are used:

[0030]

[0031] Among them, k refers to the RGB three channels, a, b, c are fitting parameters, OD is the optical density, Among them I 0 is the incident light intensity, I t is the intensity of transmitted light. When scanning, pay attention to the effective range of the scanner light intensity. If the range is exceeded, the pixel value will be saturated. Fit a curve that the normal signal points on the film satisfy, and use points outside a certain range of the curve as interference points.

[0032] As a preferred embodiment of the present invention, all pixel values ​​of the film are projected relative to the dose response curve, and the pixel points on the curve are projected onto the central axis of the curve, thereby achieving denoising of the image.

[0033] The above embodiments and descriptions are only for illustrating the principles and best embodiments of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, all of which fall within the scope of the present invention to be protected.

Claims

1. A film interference point erasing method, It is characterized in that The following steps are involved: S1), using a scanner to scan the film to obtain a color image with three channels of RGB, each pixel point and each channel has a corresponding pixel value p (R, G, B); S2) After the film is exposed to radiation, the color of the radiation-sensitive coating will change. The degree of change affected by the radiation dose is expressed as R(D), G(D), B(D), where RGB is the red, green and blue channels, and D is the dose size. The larger the dose D, the darker the color. If there are no interference points in the film, the pixel values ​​of all points on the film will be distributed as a curve with a certain width in the RGB space; if they are not on this curve, they are interference points; If there is no scale on the film, since the number of interference points is less than the number of normal points, the interference points are detected by an abnormal point detection algorithm or a classification algorithm; If a scale is set on the film, the scale corresponding to the pixel value and dose of the film is set, and the scale on the film and the dose response curve are in the form of: Among them, k refers to the RGB three channels, D k refers to the dose of the kth channel, are fitting parameters, OD is the optical density, , where I 0 is the incident light intensity, I t is the intensity of transmitted light; when scanning, it is necessary to pay attention to the effective range of the scanner light intensity. If the pixel value exceeds the range, it will be saturated. Fit the curve that the normal signal points on the film meet, and use the points outside a certain range of the curve as interference points; S3), after the interference point is detected, the inpaint function of opencv is used to take the interference point as a mask, and then the pixel value of the interference point position is inferred from the pixel value of the normal point near the interference point, so as to achieve the wiping of the interference point.

2. A method for erasing film interference points according to claim 1, Features: The algorithm used is scipy's Isolation Forest algorithm, which is done by inputting all the pixels of the film as three-dimensional data points.

3. A film interference point erasing method according to claim 1, Features: Through the dose response curve, all the pixel values ​​of the film are projected relative to the curve, and the pixel points on the curve are projected onto the central axis of the curve, thereby achieving denoising of the image.

Images