CN104616257A - Recovery evidence obtaining method for blurred degraded digital images in administration of justice - Google Patents

Abstract

The restoration and evidence collection method of fuzzy and degraded digital images in judicial practice of the present invention includes: a). Judgment of motion or defocus blur; b). Using the formula Carry out sharpening processing; c). Judgment whether there is blurring limited by shooting conditions; d). Sharpening processing of blurred images; e). Judgment of low-resolution blurring; f). Super-resolution using deblurring and noise reduction The high-rate image reconstruction method obtains the super-resolution image of the blurred image; executes steps g); g). After the graphics processing is completed, the acquired image is a scientifically and reasonably restored image of the original blurred image. The restoration and evidence collection method of fuzzy and degraded digital images in judicial practice of the present invention can eliminate image blur caused by motion or defocus blur, poor shooting conditions, and image blur caused by low resolution, scientifically and reasonably realizing the original The realistic restoration of blurred images without image tampering provides a strong guarantee for images to be used as judicial basis.

Description

A Restoration and Evidence Collection Method of Fuzzy and Degraded Digital Image in Judiciary

technical field

The present invention relates to a method for recovering and obtaining evidence of fuzzy and degraded digital images in judicial practice, and more specifically, to a kind of fuzzy and degraded digital images that can effectively eliminate image blur caused by movement, poor shooting environment and low resolution. Restorative Evidence Methods in Justice.

Background technique

Digital image content forensics technology mainly judges the accuracy, integrity and originality of image content by analyzing the statistical characteristics of digital images. All kinds of images may be images after one or more kinds of tampering. Although they cannot be directly perceived by human vision, various tampering will inevitably change some data characteristics of the image. It is the key to digital image forensics to use scientific and reasonable image processing methods to detect the change of data characteristics caused by human tampering.

In judicial review, it is often encountered that the provided digital images are blurred and not enough to provide sufficient evidence for the content of the corroboration. At this time, it is necessary to restore the blurred and degraded images to make them become Valid evidence in the judicial process. In a nutshell, there are three main reasons for image blur and degradation: (1) one is that the image is blurred due to the bad shooting environment (such as heavy rain, heavy fog, heavy snow or haze); (2) the other is that the image is blurred. The image fineness caused by the low resolution of the acquisition equipment is not enough; (3) The third is that the image is blurred due to the relative motion between the image acquisition equipment and the acquisition object.

For the first type of situation, if the key target in the image is blurred, the details of the image cannot be observed clearly, and convincing and useful information cannot be provided, resulting in insufficient proof of the image. For the second type of situation, the details of the target of interest that can be displayed by the digital image with low resolution are limited. If the key objects in the image cannot be clearly displayed by scientific means, it may cause the original content of effective proof. The image loses its application effect. In some cases, in order to make the clear and recognizable images lose their role as evidence, the other party may artificially reduce the resolution of the images so that illegal or criminal evidence cannot be obtained. The clear restoration of the image is also very important. For the third type of situation, due to the sudden shaking or fast movement of the object or image acquisition equipment at the moment of digital image acquisition, a digital image with motion blur or defocus blur is caused, and the image with motion blur or defocus blur Clarity is also the key to digital image forensics.

Contents of the invention

In order to overcome the disadvantages of the above-mentioned technical problems, the present invention provides a restoration and evidence collection method of fuzzy and degraded digital images in judicial practice.

The fuzzy degraded digital image recovery and evidence collection method in the judiciary of the present invention is special in that it is realized through the following steps: a). For the judgment of motion or defocus blur, for the fuzzy degraded digital image to be cleared, first artificially Determine whether there is motion blur or defocus blur, if there is, then perform step b); if not, then perform step c); b). Clearing of motion or defocus blur, for the digital image to be cleared , using formula (1) to clarify it:

(1)

for the original image an estimate of is the degradation function of the image, is the additive noise of the image; and The more accurate you can get, the closer ; Determined by the blurred motion direction and motion length; use the obtained blurred motion direction and length to perform clear processing on the image, and remove the additive noise of the blurred image to obtain a clear image; perform step c); c). Determine whether there is Blur limited by shooting conditions, judge whether the image to be cleared has image blur caused by bad shooting conditions, if it exists, then perform step d); if it does not exist, then perform step e); d). Clear processing, set the pixel points in the image to be processed The R, G, and B values of , , , the value ranges of R, G, and B values of all pixels in the image are , , , 0≤ ≤255, 0≤ ≤255, 0≤ ≤255, use the formula (2) to clarify it:

(2)

Take advantage of , , value to replace the pixel The R, G, B values of R, G, and B can realize the clear processing of the graphics; among them, , , ; Execute step e); e). Judgment of low resolution and blur, whether there is artificially judging whether there is a blurred image caused by low image resolution, if not, then perform step g); if it exists, then perform Step f); f). The sharpening process of the low-resolution image is realized by the following steps: f-1). The acquisition of multiple low-resolution images, if the image to be cleared is a video frame image, through the video frame processing to obtain multiple low-resolution images with consistent frames; if the image to be cleared is a single digital image, use interpolation to obtain multiple low-resolution images with consistent frames; set the obtained low-resolution images The number is N; f-2). The super-resolution reconstruction of the image, according to the N low-resolution images obtained in the step f-1), adopts the super-resolution image reconstruction method of deblurring and noise reduction to obtain blurred The super-resolution image of the image; execution step g); g). The graphics processing is completed, and the restoration processing of the blurred and degraded digital graphics is completed. The blurring of the image can be eliminated, and the final obtained image is a scientific and reasonable restored image of the original blurred image, and the details displayed can be used as evidence.

In the judiciary restoration and evidence collection method of blurred and degraded digital images of the present invention, the clearing process of motion or defocus blur described in step b) is realized through the following steps: b-1). The difference operation of the image is based on rotation The motion blur direction of the differential operator is used to estimate the motion direction of the blurred image, and the differential operator and the blurred image are linearly convolved using formula (3) to realize the differential operation of the blurred image:

(3)

in, represents the original blurred image, represents the directional difference operator, Indicates the differential result after being acted on by the differential operator; b-2). To determine the direction of motion, let according to = The order of changes in turn, and find the corresponding , and according to formula (4) to find each Sum of absolute values of pixels in :

(4)

During the operation, if the inequality (5) is satisfied Appear,

(5)

then make The angle value with a minimum value is the motion direction of the blurred image, let the motion direction be ; b-3). Obtaining the motion blur parameters, using the fuzzy image point spread function parameter estimation method to estimate the motion length of the motion blur image;

Use formula (6) to find the motion blur parameters :

(6)

is the initial modulus motion parameter, is the motion step, D is the number of searches, 1≤ ≤D; b-4). Obtain the blurred image again, using the blur parameter to the original blurred image Blur to get a blurred image again , Find it by formula (7):

(7)

b-5). Determination of fuzzy motion length and calculation error ,make with minimum That is, the blur motion step size.

The restoration and evidence collection method of fuzzy and degraded digital images in judicial practice of the present invention, additive noise in step b) The removal of is realized by mean filtering method, median filtering method or frequency domain low-pass filtering method.

In the judiciary restoration and evidence collection method of fuzzy and degraded digital images of the present invention, in step d), if the acquired image is a gray value image, set the pixel The grayscale value of , the gray value range of all pixels in the image is ,0≤ ≤255; then the sharpening processing of the blurred image is realized by the formula (8):

(8)

Take advantage of value to replace the pixel of value, the clear processing of the graphics can be realized; among them, .

The beneficial effects of the present invention are: the restoration and evidence collection method of blurred and degraded digital images in judicial practice of the present invention, in the case of motion or defocus blur in the original blurred image, through the degradation function and additive noise Realize the maximum estimation of the original image to maximize the restoration of the image; when the blurred image has image blur caused by the poor shooting environment, such as the image acquired under heavy rain, heavy fog or heavy snow weather conditions, use The pixel value expansion method is used to obtain a clear image; when the blurred image has image blur caused by low resolution, the super-resolution image reconstruction method is used to obtain the super-resolution image of the blurred image, and the final obtained image is the original The fuzzy image is scientifically and reasonably restored, and the details displayed can be used as evidence.

The restoration and evidence collection method of fuzzy and degraded digital images in judicial practice of the present invention can eliminate image blur caused by motion or defocus blur, poor shooting conditions, and image blur caused by low resolution, scientifically and reasonably realizing the original The realistic restoration of blurred images without image tampering provides a strong guarantee for images to be used as judicial basis.

Description of drawings

Fig. 1 is the flow chart of the recovery evidence collection method of the fuzzy degraded digital image in the judiciary of the present invention;

Fig. 2 is original fuzzy image among the present invention;

Fig. 3 is the image of the original blurred image in the present invention after being processed by motion blur clearing;

Fig. 4 is the image after pixel equalization processing of the image in Fig. 3;

FIG. 5 is an image after super-resolution image reconstruction processing of the image in FIG. 4 .

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, the flow chart of the restoration and evidence collection method of blurred and degraded digital images in judicial practice of the present invention is provided, which is realized by the following methods:

a). Judgment of motion or defocus blur. For the blurred and degraded digital image to be cleared, first artificially judge whether there is motion blur or defocus blur. If yes, perform step b); if not, perform step c);

As shown in Figure 2, an original blurred image is given. The blurred image shown not only has motion blur, but also image blur caused by poor shooting environment and image blur caused by low resolution. Due to the blurring of the original image, it is impossible to obtain the license plate number in the monitoring screen.

b). Clearing of motion or defocus blur. For the digital image to be cleared, formula (1) is used to clear it:

(1)

for the original image an estimate of is the degradation function of the image, is the additive noise of the image; and The more accurate you can get, the closer ; Determined by blur motion direction and motion length;

Perform clearing processing on the image by using the obtained blurred motion direction and length, and remove the additive noise of the blurred image to obtain a clear image; perform step c);

In this step, the process of clearing motion or defocus blur is realized through the following steps:

b-1). Image differential operation, using the motion blur direction based on the rotation differential operator to estimate the motion direction of the blurred image, and performing linear convolution operation between the differential operator and the blurred image using formula (3) to realize the blurred image The difference operation:

(3)

in, represents the original blurred image, represents the directional difference operator, Indicates the difference result after being acted on by the difference operator;

b-2). To determine the direction of motion, let according to = The order of changes in turn, and find the corresponding , and according to formula (4) to find each Sum of absolute values of pixels in :

(4)

During the operation, if the inequality (5) is satisfied Appear,

(5)

then make The angle value with a minimum value is the motion direction of the blurred image, let the motion direction be ;

b-3). Calculation of the motion blur parameters, using the fuzzy image point spread function parameter estimation method to estimate the motion length of the motion blur image;

Use formula (6) to find the motion blur parameters :

(6)

is the initial modulus motion parameter, is the motion step, D is the number of searches, 1≤ ≤D;

b-4). Obtain the blurred image again, using the blur parameters to the original blurred image Blur to get a blurred image again , Find it by formula (7):

(7)

b-5). Determination of fuzzy motion length and calculation error ,make with minimum That is, the blur motion step size.

In this step, the additive noise The removal of is realized by mean filtering method, median filtering method or frequency domain low-pass filtering method.

As shown in Figure 3, the image processed by step b) is given. It can be seen that after the processing of step b), the motion blur of the image is eliminated, but the acquired image is blurred due to poor shooting conditions. Still can't recognize the license plate number.

c). Judging whether there is blur limited by shooting conditions, judging whether the image to be cleared has image blur caused by bad shooting conditions, if it exists, then perform step d); if not, then perform step e);

d). The clear processing of blurred images, set the pixel points in the image to be processed The R, G, and B values of , , , the value ranges of R, G, and B values of all pixels in the image are , , , 0≤ ≤255, 0≤ ≤255, 0≤ ≤255, use the formula (2) to clarify it:

(2)

Take advantage of , , value to replace the pixel The R, G, B values of R, G, and B can realize the clear processing of the graphics; among them, , , ; Execute step e);

If the acquired image is a grayscale image, set the pixel The grayscale value of , the gray value range of all pixels in the image is ,0≤ ≤255; then the sharpening processing of the blurred image is realized by the formula (8):

(8)

Take advantage of value to replace the pixel of value, the clear processing of the graphics can be realized; among them, .

As shown in Figure 4, the image in Figure 3 after pixel equalization processing is given. Compared with the monitoring screen in Figure 3, the image in Figure 4 is caused by the absence of motion and the harsh shooting environment. The image is blurry, but the license plate number cannot be obtained due to the low resolution of the image itself.

e). Judgment of low resolution and fuzzy, artificially judge whether there is a situation of blurred image caused by low image resolution, if not, then perform step g); if it exists, then perform step f);

f). The clear processing of the low-resolution image is realized through the following steps:

f-1). The acquisition of multiple low-resolution images, if the image to be cleared is a video frame image, through the processing of the video frame, multiple low-resolution images with the same picture are obtained; if the image to be cleared is For a single digital image, an interpolation method is used to obtain a plurality of consistent low-resolution images; the number of low-resolution images obtained is N;

f-2). The super-resolution reconstruction of the image, according to the N low-resolution images obtained in the step f-1), adopt the super-resolution image reconstruction method of deblurring and noise reduction to obtain the super-resolution of the blurred image image; perform step g);

As shown in Figure 5, the image in Figure 4 after super-resolution image reconstruction is given. Due to the elimination of the motion blur of the original image, the poor shooting environment blur, and the low-resolution blur, the license plate in the monitoring screen The numbers are clearly legible.

g). After the graphics processing is completed, the restoration of the blurred and degraded digital graphics is completed. The image blur caused by motion or defocus blur, poor shooting conditions, and image blur caused by low resolution can be eliminated. The finally acquired image is The original blurred image is scientifically and reasonably restored, and the details displayed can be used as evidence.

Claims (4)

1. A method for recovering and collecting evidence of fuzzy and degraded digital images in justice, characterized in that, it is realized through the following steps: a). Judgment of motion or defocus blur. For the blurred and degraded digital image to be cleared, first artificially judge whether there is motion blur or defocus blur. If yes, perform step b); if not, perform step c); b). Clearing of motion or defocus blur. For the digital image to be cleared, formula (1) is used to clear it: (1) for the original image an estimate of is the degradation function of the image, is the additive noise of the image; and The more accurate you can get, the closer ; Determined by blur motion direction and motion length; Perform clearing processing on the image by using the obtained blurred motion direction and length, and remove the additive noise of the blurred image to obtain a clear image; perform step c); c). Judging whether there is blur limited by shooting conditions, judging whether the image to be cleared has image blur caused by bad shooting conditions, if it exists, then perform step d); if not, then perform step e); d). The clear processing of blurred images, set the pixel points in the image to be processed The R, G, and B values of , , , the value ranges of R, G, and B values of all pixels in the image are , , , 0≤ ≤255, 0≤ ≤255, 0≤ ≤255, use the formula (2) to clarify it: (2) Take advantage of , , value to replace the pixel The R, G, B values of R, G, and B can realize the clear processing of the graphics; among them, , , ; Execute step e); e). Judgment of low resolution and fuzzy, artificially judge whether there is a situation of blurred image caused by low image resolution, if not, then perform step g); if it exists, then perform step f); f). The clear processing of the low-resolution image is realized through the following steps: f-1). The acquisition of multiple low-resolution images, if the image to be cleared is a video frame image, through the processing of the video frame, multiple low-resolution images with the same picture are obtained; if the image to be cleared is For a single digital image, an interpolation method is used to obtain a plurality of consistent low-resolution images; the number of low-resolution images obtained is N; f-2). The super-resolution reconstruction of the image, according to the N low-resolution images obtained in the step f-1), adopt the super-resolution image reconstruction method of deblurring and noise reduction to obtain the super-resolution of the blurred image image; perform step g); g). After the graphics processing is completed, the restoration of the blurred and degraded digital graphics is completed. The image blur caused by motion or defocus blur, poor shooting conditions, and image blur caused by low resolution can be eliminated. The finally acquired image is The original blurred image is scientifically and reasonably restored, and the details displayed can be used as evidence. 2. The method for recovering and obtaining evidence of fuzzy and degraded digital images in justice according to claim 1, wherein the process of clearing motion or defocus blur described in step b) is realized by the following steps: b-1). Image differential operation, using the motion blur direction based on the rotation differential operator to estimate the motion direction of the blurred image, and performing linear convolution operation between the differential operator and the blurred image using formula (3) to realize the blurred image The difference operation: (3) in, represents the original blurred image, represents the directional difference operator, Indicates the difference result after being acted on by the difference operator; b-2). To determine the direction of motion, let according to = The order of changes in turn, and find the corresponding , and according to formula (4) to find each Sum of absolute values of pixels in : (4) During the operation, if the inequality (5) is satisfied Appear, (5) then make The angle value with a minimum value is the motion direction of the blurred image, let the motion direction be ; b-3). Calculation of the motion blur parameters, using the fuzzy image point spread function parameter estimation method to estimate the motion length of the motion blur image; Use formula (6) to find the motion blur parameters : (6) is the initial modulus motion parameter, is the motion step, D is the number of searches, 1≤ ≤D; b-4). Obtain the blurred image again, using the blur parameters to the original blurred image Blur to get a blurred image again , Find it by formula (7): (7) b-5). Determination of fuzzy motion length and calculation error ,make with minimum That is, the blur motion step size. 3. according to claim 1 and 2 described fuzzy degraded digital image in judicial restoration method for forensics, it is characterized in that: in step b) additive noise The removal of is realized by mean filtering method, median filtering method or frequency domain low-pass filtering method. 4. according to claim 1 or 2 described fuzzy degraded digital image in judicial restoration method for forensics, it is characterized in that: in step d), if the acquired image is a gray value image, set the pixel point The grayscale value of , the gray value range of all pixels in the image is ,0≤ ≤255; then the sharpening processing of the blurred image is realized by the formula (8): (8) Take advantage of value to replace the pixel of value, the clear processing of the graphics can be realized; among them, .