CN114419006A - A method and system for removing text watermark from grayscale video that changes with background - Google Patents

Abstract

The invention discloses a method and a system for removing grayscale video text watermarks that change with background, including acquiring grayscale video to be processed, preprocessing original images in the grayscale video to be processed, and obtaining a watermark area for preliminary positioning; The watermark area of the preliminary location is subjected to watermark detection processing on the watermark area of the preliminary location by using a two-way watermark detection processing method, and two intermediate results are obtained; the intermediate results of the two channels are calculated to generate a watermark template, and a background is obtained. region and precisely positioned watermark region; using the background region to fill the precisely positioned watermark region to obtain an image after watermark removal. The invention solves the problem of removing grayscale video text watermarks that change with the background, removes the watermarks at specific positions in real time for grayscale imaging videos, and overcomes the changes of the watermarks that occur according to the light and dark changes of the images. The present invention can provide clean data for data labeling and training in the field of artificial intelligence.

Description

A method and system for removing text-like watermark from grayscale video that changes with background

technical field

The invention relates to the technical field of watermark removal, in particular to a method and system for removing grayscale video text watermarks that change with background.

Background technique

In the field of video image applications, grayscale imaging is often used, and various grayscale images have watermarks such as video annotations, dates, and times. Such watermarks are collectively referred to as text watermarks herein. These watermarks have several main characteristics: the first watermark is changing, such as time changes; the second watermark is light and dark, such as the background is dark, the watermark is light; the background is light, the watermark is dark ; The third watermark has slender lines, and the watermark and the background blend with each other; the gray value of the fourth watermark itself remains the same.

In recent years, it has become more and more common to use artificial intelligence algorithms to process videos and images, such as object detection and object classification. Artificial intelligence algorithms, especially the supervised deep learning algorithms that have become popular in recent years, work based on large amounts of data. First build a neural network, then input the labeled training set data into the network for training to obtain a trained model, and finally use the trained model for real-time work.

Data acquisition is very important for deep learning, but it is not easy, especially in some professional and niche fields, it is difficult to find relevant data resources on the Internet. It is a very sensible and effective way to collect some existing videos and images and use them, for example, for model training. However, many times the images or videos we obtain have watermarks. These watermarks interfere with the labeling of the training set and later training, reducing the efficiency of project advancement.

Existing watermarking techniques can be divided into two categories, one is based on machine learning algorithms, and the other is traditional algorithms. Most machine learning-based algorithms use training data. Most of the existing traditional watermark removal algorithms are based on static images, and need to manually identify the watermark, or use a more complex algorithm to detect the watermark and then fill the watermark area with the area near the watermark. The template matching detection watermark is aimed at the situation that the watermark style is fixed and unchanged. When the watermark style is changed, the traditional template matching can not be used to detect the watermark. Another method for removing watermarks is relatively straightforward, directly covering the watermark part to add a new watermark, or blurring the watermark part, which is not helpful for the model training mentioned in the background art. The text watermark mentioned in this article is blended with the background, that is to say, the hollow part of the text is the background. You cannot simply erase the watermark together with the hollow background, which will make the image after removing the watermark look very abrupt.

Then, the above existing watermark removal technologies are not suitable for the removal of text watermarks in grayscale videos that change with the background, and cannot solve the problem of removing watermarks at specific locations in real time for grayscale imaging videos; and it is difficult to quickly remove video or images. Batch processing.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is that the existing watermark removal technology is not suitable for the grayscale video text watermark removal that changes with the background, and cannot solve the problem of real-time removal of watermarks in specific positions for grayscale imaging videos. The object of the present invention is to provide a method and system for removing grayscale video text watermarks that change with the background, through which the problem of removing grayscale video text watermarks that change with the background can be solved, and the grayscale imaging video can be removed in real time at a specific location. Watermark, and overcome the change of watermark according to the light and dark of the image; In addition, clean data with watermark removed can be obtained in batches.

The present invention is achieved through the following technical solutions:

In a first aspect, the present invention provides a method for removing a grayscale video text watermark that changes with the background, the method comprising:

Obtain the grayscale video to be processed, and preprocess the original image in the grayscale video to be processed to obtain the watermark area for preliminary positioning;

According to the preliminarily positioned watermark area, a two-way watermark detection processing method is used to perform watermark detection processing on the preliminarily positioned watermark area (that is, precise positioning) to obtain two-way intermediate results; Seek the intersection), generate a watermark template, extract the pixel with the largest number of identical pixel values in the watermark template as a watermark area for precise positioning, and obtain a background area simultaneously; wherein, the area other than the precisely positioned watermark area is the background area;

The precisely positioned watermark area is filled with the background area to obtain an image after watermark removal.

The working principle is: the existing traditional watermark removal technology is not suitable for the removal of grayscale video text watermarks that change with the background, and cannot solve the problem of real-time removal of watermarks in specific locations for grayscale imaging videos; and most algorithms based on machine learning require Use training data. Therefore, the present invention designs a grayscale video text watermark removal method that changes with the background, including: first, acquiring a watermark image to be processed, and preprocessing the watermark image to be processed to obtain a preliminary positioned watermark area; wherein Image preprocessing includes preliminary positioning of the watermark area and noise filtering; the primary function of the preliminary positioning of the watermark area is to reduce the amount of calculation, and secondly, it can reduce the interference of the background to the threshold segmentation; the noise filtering is to reduce the interference to the edge detection. Secondly, the watermark is precisely positioned in the preliminary positioned watermark area, and a watermark template is generated; then, the watermark area is filled according to the watermark template to remove the watermark; finally, the image post-processing is performed to improve the watermark appearing at the boundary of the watermark area after the watermark is removed. Fragmentation.

The method of the invention has a reasonable process, solves the problem of removing the grayscale video text watermark that changes with the background, removes the watermark in a specific position in real time for the grayscale imaging video, and overcomes the change of the watermark according to the light and dark changes of the image; in addition, it can be batched Get clean data with watermark removed.

Further, the preprocessing of the watermark image to be processed includes preliminary positioning of the watermark area and noise filtering.

Further, the preliminary positioning of the watermark area is to perform preliminary positioning of the watermark, provided that the position of the watermark does not change. The watermark in a specific position only needs to draw a rectangular frame based on the center of the watermark. The long side of the rectangular frame should exceed the overall length of the watermark, and the short side should be 3 times the width of the watermark, so as to ensure that the proportion of the background in the initial positioning is higher than that of the watermark. , and at the same time reduce the amount of computation as much as possible.

Further, the two-way watermark detection processing method is used to perform watermark detection processing on the preliminary positioned watermark area, which specifically includes:

The first watermark detection processing: the sobel edge detection method is used to detect the edges in the horizontal and vertical directions respectively for the watermark area of the preliminary positioning, so as to obtain the horizontal edge and the vertical edge; and superimpose the horizontal edge and the vertical edge to obtain the superimposed edge. edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and perform an operation on the expansion result and the second intermediate result , get the first intermediate result, that is, the first watermark template;

The second watermark detection process: detecting the grayscale value of the watermark in the initially positioned watermark area, and using the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; The image is morphologically expanded to obtain the second intermediate result, that is, the second watermark template.

The above two-way watermark detection processing method is based on the following design considerations: the first watermark detection processing is to solve the watermark area according to the edge. This is considering that the watermark is obviously different from the background, and the edge of the watermark is easy to find. Therefore, first use the first method according to the edge. One watermark detection process obtains the watermark area, but the background may interfere with the watermark in this process, so the second watermark detection process is combined next; the second watermark detection process is to find the watermark according to the grayscale value of the watermark This is based on the same gray value of the watermark, but the individual gray values of the background area will be mistaken for a watermark, so the second watermark detection processing effect based on the gray value of the watermark alone is not so good. Therefore, the present invention considers two detection methods combining edge solution and watermark gray value solution to increase the watermark detection confidence and make the detected watermark area more accurate.

Morphological expansion is added to the two watermark detection processing methods. The reason for the morphological expansion of the edge image of the first channel is that the watermark has a certain width. The lost edges should be included as much as possible, following the principle that it is better to mistake the background as a watermark than to mistake the watermark as the background. The reason for expanding the segmentation result of the second channel also follows the principle that it is better to mistake the background as a watermark and not mistake the watermark as the background, and use the watermark as the background as little as possible.

Further, the watermark area of the first intermediate result is 1 or 255, and the background area is 0.

In the first watermark detection process, after adaptive segmentation and morphological expansion, the center point of the binary image is selected as the anchor point, and a small rectangular frame is drawn according to the length and width of the preliminary positioned watermark area by 3 times. Draw a small rectangular box of the same size at the same position on the binary image of the intermediate result of the two-way. Find the intersection of these two small rectangular boxes twice: the pixel value of the small rectangular area of the intermediate result of the second road in the two intersections remains unchanged, and the pixel value of the small rectangular area of the first road is intersected in the first time. It remains unchanged when the second intersection is performed, and the pixel value is inverted when the second intersection is performed. Count the number of pixels with a pixel value of 1 or 255 after the two intersections, and check whether the pixel value of the small rectangular area of the first binary image corresponding to the intersection with the largest number is reversed. The pixel value of the binary image after the morphological expansion of the first channel is negated to obtain the intermediate result of the first channel.

In the second watermark detection process, after obtaining the grayscale value of the watermark, threshold segmentation is performed using the grayscale value as the threshold value. Obtain the average value of the gray value of the image in the preliminary positioned watermark area, compare it with the gray value of the watermark, and select the corresponding threshold segmentation parameters to ensure that the value of the watermark after segmentation is 1 or 255, and the value of the background is 0.

Further, the detection of the gray value of the watermark is determined in a sliding window manner.

Further, after the watermark is removed, the method also includes post-processing of the image, using the alpha fusion method to perform alpha fusion on the watermark area initially positioned before the watermark removal and the corresponding watermark area after the watermark removal, and superimpose it on the image in the background area. Wherein, the corresponding watermark area after the watermark removal is an area that is completely consistent with the watermark area initially positioned before the watermark removal after the watermark removal.

Further, a Gaussian coefficient is used as the weight for fusion performed by the alpha fusion method.

In a second aspect, the present invention further provides a background-changing grayscale video text watermark removal system, the system supports the background-changing grayscale video text watermark removal method; the system includes an acquisition unit , a preprocessing unit, a watermark detection unit, a watermark removal unit and an image post-processing unit;

The obtaining unit is used to obtain the grayscale video to be processed;

The preprocessing unit is used to preprocess the original image in the grayscale video to be processed to obtain a watermark area that is initially positioned;

The watermark detection unit is configured to perform watermark detection processing (that is, precise positioning) on the preliminary positioned watermark region by using a two-way watermark detection processing method according to the preliminary positioned watermark region, and obtain two-way intermediate results; The above-mentioned two-way intermediate results are operated (for intersection) to generate a watermark template; the pixels with the largest number of identical pixel values in the watermark template are extracted as the watermark area of precise positioning, and the background area is obtained simultaneously; wherein, outside the watermark area of precise positioning The area is the background area;

The watermark removing unit is used to fill the precisely positioned watermark area by using the background area to obtain an image after watermark removal;

The image post-processing unit is used to perform alpha fusion between the watermark area preliminarily positioned before watermark removal and the corresponding watermark area after watermark removal by using the alpha fusion method, and superimpose it on the image in the background area.

Further, the watermark detection unit includes a first watermark detection processing subunit, a second watermark detection processing subunit and a calculation subunit;

Described first road watermark detection processing subunit, is used to adopt sobel edge detection method to carry out edge detection to horizontal and vertical directions respectively to the watermark area of described preliminary positioning, obtains horizontal edge and vertical edge; superimpose to obtain the superimposed edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and compare the expansion result with the first binary image. Perform operations on the intermediate results of the two channels to obtain the intermediate results of the first channel, that is, the first channel watermark template;

The second watermark detection processing sub-unit is used to detect the grayscale value of the watermark in the preliminary positioned watermark area, and use the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; and The obtained second binary image is morphologically expanded to obtain the second intermediate result, that is, the second watermark template;

The calculation subunit is configured to perform an operation on the intermediate result of the first path and the intermediate result of the second path (seek an intersection) to generate a watermark template.

Compared with the prior art, the present invention has the following advantages and beneficial effects:

1. The process of the present invention is reasonable, solves the problem of removing the grayscale video text watermark that changes with the background, removes the watermark at a specific position in real time for the video of grayscale imaging, and overcomes the change that the watermark will occur according to the light and dark changes of the image; In addition, Clean data with watermark removed can be obtained in batches.

2. Grayscale imaging is usually used in many fields. Artificial intelligence technology is used more and more frequently in these fields. Data collection and labeling are crucial to artificial intelligence. The present invention can provide clean data for data labeling and training.

Description of drawings

The accompanying drawings described herein are used to provide further understanding of the embodiments of the present invention, and constitute a part of the present application, and do not constitute limitations to the embodiments of the present invention. In the attached image:

FIG. 1 is a flowchart of a method for removing a grayscale video text watermark that changes with the background of the present invention.

FIG. 2 is a flow chart of two-way watermark detection in a method for removing text watermark from grayscale video that changes with background according to the present invention.

FIG. 3 is a schematic diagram of a sliding window for detecting grayscale values of watermarks according to the present invention.

FIG. 4 is a schematic diagram of the fusion template of the alpha fusion method of the present invention.

FIG. 5 is a schematic diagram of superimposing a fusion template on an image by an alpha fusion method of the present invention.

FIG. 6 is a schematic diagram of statistics of the number of pixel values of the sliding window according to the present invention.

FIG. 7 is a schematic structural diagram of a system for removing grayscale video text watermarks that changes with the background of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and the accompanying drawings. as a limitation of the present invention.

Example 1

As shown in FIG. 1 to FIG. 6 , a method for removing a grayscale video text watermark that changes with the background of the present invention, as shown in FIG. 1 , includes:

Obtain the grayscale video to be processed, and preprocess the original image in the grayscale video to be processed to obtain the watermark area for preliminary positioning;

According to the preliminarily positioned watermark area, a two-way watermark detection processing method is used to perform watermark detection processing on the preliminarily positioned watermark area (that is, precise positioning) to obtain two-way intermediate results; Find the intersection), generate a watermark template; extract the pixel with the largest number of identical pixel values in the watermark template as the watermark area of the precise positioning, and obtain the background area simultaneously; wherein, the area other than the watermark area of the precise positioning is the background area;

The precisely positioned watermark area is filled with the background area to obtain an image after watermark removal.

Specifically, the preprocessing of the to-be-processed watermark image includes preliminary positioning of the watermark area and noise filtering processing; the main function of the preliminary positioning of the watermark area is to reduce the amount of calculation, and secondly, it can reduce the interference of the background to the threshold segmentation. Noise filtering is to reduce interference with edge detection.

Wherein, the preliminary positioning of the watermark area is to perform preliminary positioning of the watermark, and the premise is that the watermark appears in a fixed area of the picture, such as the center, corner and other positions.

Specifically, watermark detection is one of the cores of this method. As shown in FIG. 2 , the pre-processed pre-positioned watermark area is input, and then a two-way watermark detection processing method is used to perform watermark detection processing on the pre-positioned watermark area. The two-way watermark detection processing method is based on the following design considerations: the first watermark detection process is to solve the watermark area according to the edge. This is considering that the watermark is obviously different from the background, and the edge of the watermark is easy to find. Therefore, first use the first watermark according to the edge. The watermark detection process obtains the watermark area, but the background may interfere with the watermark in this process, so the second watermark detection process is combined next; the second watermark detection process is to find the watermark area according to the watermark gray value. , which is based on the same gray value of the watermark, but individual gray values in the background area will be mistaken as watermarks, so the second watermark detection processing effect based on the gray value of the watermark alone is not so good. Therefore, the present invention considers two detection methods combining edge solution and watermark gray value solution to increase the watermark detection confidence and make the detected watermark area more accurate.

specifically:

The first watermark detection processing: the sobel edge detection method is used to detect the edges in the horizontal and vertical directions respectively for the watermark area of the preliminary positioning, so as to obtain the horizontal edge and the vertical edge; and superimpose the horizontal edge and the vertical edge to obtain the superimposed edge. edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and perform an operation on the expansion result and the second intermediate result , get the first intermediate result, that is, the first watermark template;

The second watermark detection process: detecting the grayscale value of the watermark in the initially positioned watermark area, and using the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; The image is morphologically expanded to obtain the second intermediate result, that is, the second watermark template. Specific implementation: First, the gray value of the watermark is detected, and the center position of the preliminary positioned watermark area is set as the anchor point, such as the center point of the dashed box as shown in FIG. 3 . A small sliding window containing the watermark is set centered on the anchor point. The height of the sliding window should not be higher than the height of the watermark to ensure that the detected edges do not include the edges of objects in the background to the greatest extent. Sobel edge detection is performed on the sliding window area centered on the anchor point, horizontal and vertical edges are detected and superimposed, and then a sliding window is slid left and right respectively with the step size of the width of the sliding window, and the edge is obtained by the same method, Get 3 edge templates. Perform the morphological closing operation on the three edge templates, stitch the small gaps and small breaks between the edges, and then count the gray values of the original image covered by the edges respectively, and count the number of pixels corresponding to a gray value as As shown in Figure 6, the number num is then sorted from large to small. The three lists are denoted L1, L2, L3. The maximum number of pixel values and pixel value pairs of the corresponding list are expressed as (NumMax1, VMax1), (NumMax2, VMax 2), (NumMax3, VMax 3), and NumMaxi represents the maximum number of pixels with the same pixel value in the list Li, The corresponding pixel value is VMax i. If VMax 1, VMax 2, and VMax 3 are greater than or equal to the same two numbers, the value is considered to be the pixel value of the watermark. If the same value is less than two, that is, VMax 1, VMax 2, and VMax 3 are not the same, continue to slide the slider to the left and right to count the number of pixel values covered by the edge corresponding to the slider, until the generation of The same number of pixel values VMax i corresponding to NumMaxi in the list Li is greater than or equal to 3, then the same pixel value corresponding to the list with the largest number is taken as the gray value of the edge. For example, there are 5 lists now, and the VMax corresponding to NumMax in the 3 lists is the same, then the pixel value VMax is considered to be the pixel value of the watermark. If the generated list has two or more lists with the same pixel value corresponding to the maximum number of pixels, for example, there are 3 lists with the pixel value v1 corresponding to the maximum number of pixels NumMax, and the maximum number of pixels NumMax corresponds to the list. There are also 3 lists with pixel value v2, then continue to slide the slider until there is a unique and maximum number of lists that meet the requirements. From a practical point of view, basically three sliding windows can be used to robustly determine the watermark gray value.

Wherein, the watermark area of the first intermediate result is 1 or 255, and the background area is 0.

On the basis of the above processing, an intersection operation is performed on the intermediate results obtained by the two paths (the first intermediate result and the second intermediate result), and the intersection is a further reduced watermark outline, that is, a watermark template.

Because the watermark changes with the background, there is an obvious difference between the watermark and the background. Therefore, the first watermark detection process can detect the edge of the watermark, and also detect the edge of the object in the background. The interference of the background edge can be reduced by adaptive threshold segmentation, because the edge gradient of the watermark is often larger than that of the background. The second watermark detection process performs threshold segmentation, and its principle is to maximize the variance between classes, which is divided into two categories: watermark and background. In the preprocessing stage, the watermark is preliminarily positioned, and the background area is reduced to the surrounding of the watermark, which can largely avoid the failure of watermark segmentation due to the large area of the background belonging to the same category as the watermark. The intersection of the results of the two paths can remove most of the background interference on the edge and inter-class variance, and at the same time increase the robustness of the obtained watermark.

Specifically, the use of the background area to fill the precisely positioned watermark area, that is, to remove the watermark; this is to expand the watermark template after the intersection, and then use the inpaint function that comes with opencv to combine the expansion After the watermark template is removed, the principle is to fill the pixels of the background area into the watermark area. The idea of expanding the watermark template here is to minimize the case of missing edges. Even if the background around the edge is added to the edge by mistake, it can be compensated by the next step of background filling.

In order to further illustrate this embodiment, after the watermark is removed, the method also includes image post-processing, using the alpha fusion method to perform alpha fusion on the watermark area initially positioned before the watermark removal and the corresponding watermark area after the watermark removal, and superimpose on the on the image in the background area. Wherein, the corresponding watermark area after the watermark removal is an area that is completely consistent with the watermark area initially positioned before the watermark removal after the watermark removal. This is to take into account the possible side effects of removing the watermark from the watermark area: some parts of the background area will be mistakenly detected as the watermark area, which will cause the area to be separated from the background after filling. The image post-processing uses alpha fusion, and alpha fusion is performed on the watermark area initially positioned before watermark removal and the corresponding watermark area after watermark removal. Gaussian coefficients are used for the fusion weight. The alpha fusion template is shown in Figure 4. A schematic diagram on the image is shown in Figure 5.

The working principle is: the existing traditional watermark removal technology is not suitable for the removal of grayscale video text watermarks that change with the background, and cannot solve the problem of real-time removal of watermarks in specific locations for grayscale imaging videos; and most algorithms based on machine learning require Use training data. Therefore, the present invention designs a grayscale video text watermark removal method that changes with the background, including: first, acquiring a watermark image to be processed, and preprocessing the watermark image to be processed to obtain a preliminary positioned watermark area; wherein Image preprocessing includes preliminary positioning of the watermark area and noise filtering; the primary function of the preliminary positioning of the watermark area is to reduce the amount of calculation, and secondly, it can reduce the interference of the background to the threshold segmentation; the noise filtering is to reduce the interference to the edge detection. Secondly, the watermark is precisely positioned in the preliminary positioned watermark area, and a watermark template is generated; then, the watermark area is filled according to the watermark template to remove the watermark; finally, the image post-processing is performed to improve the watermark appearing at the boundary of the watermark area after the watermark is removed. Fragmentation.

The method of the invention has a reasonable process, solves the problem of removing the grayscale video text watermark that changes with the background, removes the watermark in a specific position in real time for the grayscale imaging video, and overcomes the change of the watermark according to the light and dark changes of the image; in addition, it can be batched Get clean data with watermark removed.

Grayscale imaging is usually used in many fields, and artificial intelligence technology is used more and more frequently in these fields. Data collection and labeling are crucial to artificial intelligence. The present invention can provide clean data for data labeling and training.

Example 2

As shown in FIG. 7 , the difference between this embodiment and Embodiment 1 is that this embodiment provides a grayscale video text watermark removal system that changes with the background, and the system supports the background-dependent watermark described in Embodiment 1 A modified grayscale video text watermark removal method; the system includes an acquisition unit, a preprocessing unit, a watermark detection unit, a watermark removal unit and an image post-processing unit;

The obtaining unit is used to obtain the grayscale video to be processed;

The preprocessing unit is used to preprocess the original image in the grayscale video to be processed to obtain a watermark area that is initially positioned;

The watermark detection unit is configured to perform watermark detection processing (that is, precise positioning) on the preliminary positioned watermark region by using a two-way watermark detection processing method according to the preliminary positioned watermark region, and obtain two-way intermediate results; The above-mentioned two-way intermediate results are operated (for intersection) to generate a watermark template; the pixels with the largest number of identical pixel values in the watermark template are extracted as the watermark area of precise positioning, and the background area is obtained simultaneously; wherein, outside the watermark area of precise positioning The area is the background area;

The watermark removing unit is used to fill the precisely positioned watermark area by using the background area to obtain an image after watermark removal;

The image post-processing unit is used to perform alpha fusion between the watermark area preliminarily positioned before watermark removal and the corresponding watermark area after watermark removal by using the alpha fusion method, and superimpose it on the image in the background area.

In this embodiment, the watermark detection unit includes a first watermark detection processing subunit, a second watermark detection processing subunit and a calculation subunit;

Described first road watermark detection processing subunit, is used to adopt sobel edge detection method to carry out edge detection to horizontal and vertical directions respectively to the watermark area of described preliminary positioning, obtains horizontal edge and vertical edge; superimpose to obtain the superimposed edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and compare the expansion result with the first binary image. Perform operations on the intermediate results of the two channels to obtain the intermediate results of the first channel, that is, the first channel watermark template;

The second watermark detection processing sub-unit is used to detect the grayscale value of the watermark in the preliminary positioned watermark area, and use the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; and The obtained second binary image is morphologically expanded to obtain the second intermediate result, that is, the second watermark template;

The calculation subunit is configured to perform an operation on the intermediate result of the first path and the intermediate result of the second path (seek an intersection) to generate a watermark template.

The execution process of the other units may be executed according to the flow and steps of the method for removing a grayscale video text watermark that changes with the background described in Embodiment 1, which will not be repeated in this embodiment.

As will be appreciated by those skilled in the art, the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

The specific embodiments described above further describe the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. a grayscale video text class watermark removal method that changes with background, is characterized in that, this method comprises: Obtain the grayscale video to be processed, and preprocess the original image in the grayscale video to be processed to obtain the watermark area for preliminary positioning; According to the preliminarily positioned watermark area, a two-way watermark detection processing method is used to perform watermark detection processing on the preliminarily positioned watermark area to obtain two-way intermediate results; perform operations on the two-way intermediate results to generate a watermark template; extract The pixel with the largest number of identical pixel values in the watermark template is used as the watermark area for precise positioning, and the background area is obtained simultaneously; The precisely positioned watermark area is filled with the background area to obtain an image after watermark removal. 2. a kind of gray-scale video text watermark removal method that changes with background according to claim 1, is characterized in that, described preprocessing to described to-be-processed watermark image comprises watermark area preliminary location and noise filtering deal with. 3. a kind of grayscale video text class watermark removal method that changes with background according to claim 2, is characterized in that, described watermark area preliminary location is to carry out preliminary location to watermark, and the watermark type of preliminary location is included in the watermark of image. Corner type. 4. a kind of grayscale video text class watermark removal method that changes with background according to claim 1, is characterized in that, described adopting two-way watermark detection processing method carries out watermark detection processing to the watermark area of described preliminary positioning , including: The first watermark detection processing: the sobel edge detection method is used to detect the edges in the horizontal and vertical directions respectively for the watermark area of the preliminary positioning, so as to obtain the horizontal edge and the vertical edge; and superimpose the horizontal edge and the vertical edge to obtain the superimposed edge. edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and perform an operation on the expansion result and the second intermediate result , get the first intermediate result; The second watermark detection process: detecting the grayscale value of the watermark in the initially positioned watermark area, and using the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; The image is morphologically inflated to obtain a second-pass intermediate result. 5 . The method for removing grayscale video text watermark that changes with background according to claim 4 , wherein the watermark area of the first intermediate result is 1 or 255, and the background area is 0. 6 . 6 . The method for removing a text watermark in grayscale video that changes with the background according to claim 4 , wherein the detection of the grayscale value of the watermark adopts a sliding window method to determine. 7 . 7. a kind of grayscale video text class watermark removal method that changes with background according to claim 1, is characterized in that, after this method watermark is removed, also comprises image post-processing, adopts alpha fusion method to be preliminary positioned before watermark is removed The watermark area and the corresponding watermark area after watermark removal are alpha-fused and superimposed on the image in the background area. 8 . The method for removing text-like watermarks from grayscale video according to claim 7 , wherein the alpha fusion method uses Gaussian coefficients as the weight for fusion. 9 . 9. a grayscale video text class watermark removal system that changes with background, it is characterized in that, this system supports a kind of grayscale video text class watermark removal method that changes with background as described in any one of claim 1 to 8 ; The system includes an acquisition unit, a preprocessing unit, a watermark detection unit, a watermark removal unit and an image post-processing unit; The obtaining unit is used to obtain the grayscale video to be processed; The preprocessing unit is used to preprocess the original image in the grayscale video to be processed to obtain a watermark area that is initially positioned; The watermark detection unit is configured to perform watermark detection processing on the preliminary positioned watermark area by using a two-way watermark detection processing method according to the preliminary positioned watermark area, and obtain two-way intermediate results; Perform operations to generate a watermark template; extract the pixel with the largest number of identical pixel values in the watermark template as a precisely positioned watermark area, and simultaneously obtain a background area; The watermark removing unit is used to fill the precisely positioned watermark area by using the background area to obtain an image after watermark removal; The image post-processing unit is used to perform alpha fusion between the watermark area preliminarily positioned before watermark removal and the corresponding watermark area after watermark removal by using the alpha fusion method, and superimpose it on the image in the background area. 10. The grayscale video text watermark removal system that changes with the background according to claim 9, wherein the watermark detection unit comprises a first watermark detection processing subunit, a second watermark detection processing subunit units and computational subunits; Described first road watermark detection processing subunit, is used to adopt sobel edge detection method to carry out edge detection to horizontal and vertical directions respectively to the watermark area of described preliminary positioning, obtains horizontal edge and vertical edge; superimpose to obtain the superimposed edge image; perform adaptive threshold segmentation on the superimposed edge image to obtain the first binary image; perform morphological expansion on the obtained first binary image to obtain the expansion result, and compare the expansion result with the first binary image. Perform operations on the intermediate results of the second channel to obtain the intermediate results of the first channel; The second watermark detection processing sub-unit is used to detect the grayscale value of the watermark in the preliminary positioned watermark area, and use the grayscale value of the watermark as a threshold to perform threshold segmentation to obtain a second binary image; and The obtained second binary image is morphologically expanded to obtain the second intermediate result; The calculation subunit is configured to perform operations on the first intermediate result and the second intermediate result to generate a watermark template.

Images