WO2021035643A1

WO2021035643A1 - Monitoring image generation method, apparatus, device, system, and image processing device

Info

Publication number: WO2021035643A1
Application number: PCT/CN2019/103433
Authority: WO
Inventors: 杨磊; 张青涛; 麻军平
Original assignee: 深圳市大疆创新科技有限公司
Priority date: 2019-08-29
Filing date: 2019-08-29
Publication date: 2021-03-04
Also published as: CN112119631A

Abstract

The embodiments of the present description provide a monitoring image generation method, an apparatus, a device, a system, and an image processing device. Frames are read row by row in a manner of image column blocks, and one image column block comprises several columns of pixels, ensuring data read efficiency. Each image column block read respectively generates a sub-monitoring image, the generated sub-monitoring image is then written into a corresponding column block of a monitoring image storage space, and when the whole image is divided into N image column blocks, consumption of an on-chip memory RAM only occupies 1/N of the storage of the whole monitoring image, greatly saving storage resources. By means of the method, monitoring image processing of a high-resolution image can be realized, and a monitoring image with a high resolution and a high frame rate can be generated.

Description

Surveillance image generation method, device, equipment and system, and image processing equipment

Technical field

The present disclosure relates to the field of image processing technology, and in particular to monitoring image generation methods, devices, equipment and systems, and image processing equipment.

Background technique

The monitor image (waveform image) is a waveform diagram used to observe the brightness and color distribution of the scene on the monitor. It describes the concentrated distribution of a certain component of Y/U/V/R/G/B in each column of the image . The main application on the monitor includes RGB or YUV three components displayed on the monitor at the same time or a certain color component is displayed on the monitor alone.

Since the video images are cached in DDR (Double Data Rate SDRAM, Double Data Rate Synchronous Dynamic Random Access Memory) storage on a line-by-line basis, and images are retrieved from the DDR column by column for waveform calculation, there is a problem of low data reading efficiency. Another achievable structure is to develop a waveform to be displayed on the software architecture cache or FPGA (Field-Programmable Gate Array, Field-Programmable Gate Array) on-chip RAM (Random Access Memory) In the same space of the graph, the image is output in rows, and each column is independently counted during output. However, it is limited by the size of the cache or on-chip RAM, resulting in low accuracy and accuracy of the waveform graph, and will also occupy a large amount of on-chip storage. Resources.

Summary of the invention

In view of this, the embodiments of the present disclosure propose monitoring image generation methods, devices, equipment and systems, and image processing equipment to solve the problem of low data reading efficiency in related technologies, low waveform map accuracy and accuracy, and occupying on-chip Technical problems with large storage resources.

According to a first aspect of the embodiments of the present disclosure, a monitoring image generation method is provided, the method includes:

Write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame, wherein the image frame is divided into a plurality of image column blocks in advance Each image column block includes multiple columns of pixels of the image frame, and the sub-monitoring image corresponding to each image column block is generated in the following manner:

Reading the pixels in the image column block by row;

The sub-monitoring image corresponding to the image column block is generated according to the pixel value of the pixel point.

According to a second aspect of the embodiments of the present disclosure, a monitoring image generation device is proposed, which is characterized by comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor executes the The following methods are implemented in the program:

Reading the pixels in the image column block by row;

According to a third aspect of the embodiments of the present disclosure, a monitoring image generation device is provided, which is characterized in that it further includes:

Histogram statistical unit, data processing unit and data output unit;

For each image column block in the image frame, the histogram statistical unit is used to generate a statistical histogram of each column of the image column block, and send the statistical histogram to the data processing unit; wherein, the The image frame is pre-divided into multiple image column blocks;

The data processing unit is configured to generate a sub-monitoring image corresponding to the image column block according to the statistical histogram of each column of the image column block, and send the sub-monitoring image to a data output unit;

The data output unit is used to write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame.

According to a fourth aspect of the embodiments of the present disclosure, a surveillance image generation system is provided, which is characterized in that it includes:

The surveillance image generating device according to any embodiment; and

A storage unit for storing the image frame.

According to a fifth aspect of the embodiments of the present disclosure, an image processing device is provided, which is characterized by including the surveillance image generation system described in any of the embodiments.

Applying the solution of the embodiment of this specification, the image frame is read out in rows in the form of image column blocks. One image column block includes several columns of pixels, which ensures the efficiency of data reading. Each read-out image column generates a sub-monitoring image, and the generated sub-monitoring image is written to the corresponding column block in the monitoring image storage space. When the full image is divided into N image column blocks, the on-chip storage RAM consumes only It requires 1/N of the entire surveillance image storage, which greatly reduces storage resources. The method can realize the monitoring image processing of high-resolution images, and realize the generation of high-resolution, high-frame-rate monitoring images.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present disclosure, the following will briefly introduce the accompanying drawings used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained from these drawings without creative labor.

Fig. 1 is a schematic flowchart of a method for generating a surveillance image according to an embodiment of the present disclosure.

Fig. 2 is a schematic diagram showing an image column block according to an embodiment of the present disclosure.

Fig. 3 is a schematic diagram showing an interpolation processing process of a surveillance image according to an embodiment of the present disclosure.

Fig. 4 is a schematic diagram showing the structure of a surveillance image jump address according to an embodiment of the present disclosure.

Fig. 5 is a schematic diagram showing a monitoring image generating device according to an embodiment of the present disclosure.

Fig. 6 is a schematic diagram showing another monitoring image generating device according to an embodiment of the present disclosure.

Fig. 7 is a schematic diagram showing a monitoring image generation system according to an embodiment of the present disclosure.

Fig. 8 is a schematic diagram showing another surveillance image generation system according to an embodiment of the present disclosure.

Fig. 9 is a schematic diagram showing a monitoring image generating device for implementing the method of this specification according to an embodiment of the present disclosure.

detailed description

The exemplary embodiments will be described in detail here, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with this specification. On the contrary, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.

The terms used in this specification are only for the purpose of describing specific embodiments, and are not intended to limit the specification. The singular forms of "a", "said" and "the" used in this specification and appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and/or" as used herein refers to and includes any or all possible combinations of one or more associated listed items.

It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, the information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this specification, the first information may also be referred to as second information, and similarly, the second information may also be referred to as first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "when" or "in response to determination".

As shown in Fig. 1, it is a flowchart of a monitoring image generation method according to an embodiment of this specification. The method may include:

Step S101: Write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame, wherein the image frame is divided into a plurality of Image column blocks, each image column block includes multiple columns of pixels of the image frame, and the sub-monitoring image corresponding to each image column block is generated in the following manner:

Step S102: Read pixels in the image column block by row;

Step S103: Generate the sub-monitoring image corresponding to the image column block according to the pixel value of the pixel point.

Image frames are generally read and written in rows, while monitoring images require that the concentration of the same pixels of a certain color component of the entire image be processed according to the image columns, and histogram statistics are performed on each column of the image frame. If the image frame is large, the calculation time for processing all image columns by software is slow, and it is difficult to achieve high frame rate and large frame monitoring images or simultaneous display of three color components; if hardware such as FPGA (Field-Programmable Gate Array) is used, the field can be displayed. Program gate array) or ASIC (Application Specific Integrated Circuit, special application integrated circuit) for processing, because the image frame is output in rows and columns, it is difficult to implement pipelines in the FPGA architecture, and it is difficult to achieve all columns due to the limitation of on-chip RAM resources. Most of the monitoring images can only be displayed in small images, and it is difficult to realize the display of high frame rate and large-format fine monitoring images.

Based on this, in the method of the embodiment of this specification, the image frame in the image storage space is divided into a plurality of image tiles in advance, and the image frames are read block by block according to the pre-divided image column blocks. Among them, each image column block is read in a row-by-row reading manner, and the monitoring image calculation of the entire image frame is realized by calculating the monitoring image of each image column block. The embodiments of this specification can be implemented based on FPGA, the monitoring image generation process can be executed by on-chip RAM, the original image frame can be stored in DDR, and the generated monitoring image can also be stored in DDR.

A schematic diagram of image column blocks in an embodiment is shown in FIG. 2. The image frame in the figure includes 9 columns, and every 3 columns is divided into an image column block, which is divided into 3 image column blocks in total. When reading the image, you can read each pixel in the image column block 1 row by row; after all the pixels in the image column block 1 have been read, then read the image column block 2 row by row. Similarly, after each pixel in the image column block 2 has been read, read each pixel in the image column block 3 by row.

When accessing the pixel points of each pixel in an image column block, the histogram statistics can be performed according to the pixel value of each pixel point in the image column block to generate the sub-monitoring image corresponding to the image column block. Still taking the image shown in Figure 2 as an example, when accessing each pixel in the image column block 1, the sub-monitoring image of the image column block 1 can be generated according to the pixel value of each pixel in the image column block 1; For each pixel in column block 2, the sub-monitoring image of image column block 2 can be generated according to the pixel value of each pixel in image column block 2. Similarly, when accessing each pixel in image column block 3, you can The sub-monitoring image of the image column 3 is generated according to the pixel value of each pixel in the image column 3.

After the sub-monitoring image is acquired, the sub-monitoring image can be written to the corresponding column of the surveillance image storage space. When writing, one sub-monitoring image can be written every time it is acquired.

In the above embodiment, the image frame is read row by row from the image storage space in the manner of image column blocks. Since several columns can be read at a time, the image reading efficiency is ensured. Divide the entire image frame into N image column blocks, and the processing unit only needs to open up a storage space that can store the histogram of each image column block, so the resource consumption only needs 1/N of the entire monitoring image storage, which is greatly reduced Storage resources, the use of smaller on-chip RAM can realize the generation of large-scale surveillance images. Therefore, by adopting the method of the embodiments of the present specification, the monitoring image processing of high-resolution images can be realized, and the generation of monitoring images with high resolution and high frame rate can be realized.

In an embodiment, the number of image column blocks to be divided for an image may be determined first. In order to ensure the efficiency of data reading and writing, several data bytes will be continuously read and written when reading and writing data. Therefore, the amount of read and write data when reading and writing the image frame once can be determined according to the preset data read and write efficiency, and then the number of image column blocks divided by the image frame can be determined according to the amount of read and write data.

Assuming that the image has a total of k columns, the depth of the color components needs to be counted as q bits, and a minimum burst amount that meets the efficiency of data reading and writing is selected, and the number of image columns in each image column block of the monochrome surveillance image is determined to be M, three colors The number of image columns for each image column block of the monitoring image is 3M. When the three-color monitoring images need to be displayed at the same time, three times the RAM can be instantiated, and the corresponding color component monitoring images can be generated at the same time. Assuming that the image is a 1080p RGB 8bit image, M is 64, the size of the monochrome and three-color monitoring images to be displayed is 1920*512, and the three-color waveform image is the stitching result of the three waveform images of 640*512 R, G, and B . The single-color display is divided into 30 image column blocks, and the three-color display is divided into 10 image column blocks. The RAM space size of each image column block storing each color component is 64*256*12bit. Specifically, the monochrome surveillance image M satisfies:

M*q=burst amount*burst width;

The three-color surveillance image M satisfies:

3M*q=burst amount*burst width;

Among them, in order to ensure the efficiency of DDR reading and writing, the burst width is generally greater than 32 bytes. In some cases, it is also necessary to ensure that a minimum burst of data is continuously read exactly as the M pixel data to be read (for monochrome surveillance images). ) Or the size of 3M pixel data (for three-color monitoring images).

In one embodiment, monitoring image generation includes two steps: histogram statistics and monitoring image mapping output. The step of generating the sub-monitoring image corresponding to the image column block according to the pixel value of the pixel point includes: respectively generating a statistical histogram of each column in the image column block, and according to the statistics of each column in the image column block. The histogram generates the sub-monitoring image corresponding to the image column block, wherein the statistical histogram of each column in the image column block is generated according to the following method: the number of pixels of each pixel value in the column is counted respectively, Obtain the statistical value of the corresponding pixel value in the column; generate a statistical histogram of the column according to the statistical value of each pixel value in the column.

Now take the image column block 1 in the image frame shown in FIG. 2 as an example to describe the process of generating the sub-monitoring images of the image column block 1. The sub-monitoring images of the image column block 2 and the image column block 3 are generated in a similar manner. I won't repeat it hereafter. The statistical histogram T1 in the first column, the statistical histogram T2 in the second column, and the statistical histogram T3 in the third column can be respectively generated, and then the sub-monitoring images of the image column block 1 are generated according to T1, T2, and T3. Among them, the statistical histogram T1 can be generated according to the following method: the pixel value of each pixel in the first column of the image column block 1 is counted to obtain the number of pixels of each pixel value in the column, for example, the pixel value is 0 The number of pixels is N0, the number of pixels with a pixel value of 1 is N1, the number of pixels with a pixel value of 2 is N2, and so on. N0, N1, N2, ... are the statistical values of pixel value 0, pixel value 1, pixel value 2, ... respectively. Then, the statistical histogram of the first column of the image column block 1 can be generated according to N0, N1, N2,... The statistical histograms in the second and third columns are generated in a similar way.

Execute the statistical method of image column by column during histogram statistics. For the statistics of monochrome surveillance images, statistics are only performed in the RAM in the corresponding color space. When an image column block includes 64 columns, each RAM is divided into 64 histogram storage spaces. For an image with a resolution of 1080p, the image read from DDR is 64 pixels per line, a total of 1080 lines, output in line, so the corresponding component of the input first pixel is counted in the first histogram, and the second The corresponding component of each pixel is counted in the second histogram, the corresponding component of the 64th pixel is counted in the 64th histogram, the corresponding component of the 65th pixel is counted again in the first histogram, and so on After counting all 64 columns of pixels.

For the statistics of the three-color monitoring image, the statistics are carried out in three RAMs at the same time. For an image with a resolution of 1080p, the input image is 192 pixels per line, a total of 1080 lines, and the

input pixels

1, 2, and 3 correspond to Each component of is counted in the corresponding first histogram, the 4th, 5th, and 6th pixels are counted in the second histogram, the 190th, 191, and 192th pixels are counted in the 64th histogram, the 193th , 194, and 195 pixels are counted again in the first histogram, and so on, all 192 columns of pixels are counted.

The sub-monitoring image can be generated directly based on the statistical histogram, or based on the brightness of the image. Specifically, the step of generating the sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block includes: splicing the statistical histogram of each column in the image column block to obtain all The sub-monitoring image corresponding to the image column block; or each statistical value is mapped to the corresponding brightness value, and the brightness distribution map of the corresponding column is generated according to the brightness value of each column, and the brightness distribution map of each column of the image column block Splicing is performed to obtain sub-monitoring images corresponding to the image column blocks.

For the method of generating sub-monitoring images based on statistical histograms, the statistical histograms of each column may be spliced horizontally or vertically to obtain sub-monitoring images corresponding to the image column blocks. For the method of generating sub-monitoring images based on brightness values, taking the first column in image column block 1 in Figure 2 as an example, each statistical value in the statistical histogram T1 can be mapped to a brightness value, for example, N0 is mapped Is Y0, N1 is mapped to Y1, N2 is mapped to Y2,..., and so on. When the statistical value is mapped to the brightness value, the following principle can be used: a larger statistical value is mapped to a brighter brightness value, and a smaller statistical value is mapped to a darker brightness value. According to Y0, Y1, Y2, ... the brightness distribution diagram of the first column can be obtained. In the same way, for the second column and the third column in the image column block 1, the brightness distribution map can also be generated in a similar manner. Then, horizontal splicing or vertical splicing is performed on the brightness distribution map of the first column, the brightness distribution map of the second column, and the brightness distribution map of the third column to obtain the sub-monitoring image of the image column block 1. In a similar manner, the sub-monitoring images of the image column block 2 and the image column block 3 can be obtained, which will not be repeated here.

In one embodiment, each statistical value may be mapped separately, and the mapping methods include linear mapping method, nonlinear mapping method, table look-up mapping method, etc., after mapping, the brightness value corresponding to the statistical value is obtained. Taking the linear mapping method as an example, in one embodiment, the step of linearly mapping each statistical value includes:

Perform linear mapping on the statistical values in the following way:

Y=histo(x)*(2 ^k -1)/T;

In the formula, histo(x) is the statistical value of the number of pixels with the pixel value x, T is the preset upper limit of the statistical value, k is the number of bits representing the brightness value, and Y is the brightness value.

The mapping method for linearly mapping the statistical value to the brightness value in this specification can also use other formulas. In addition, other mapping methods other than linear mapping can also be used to map the statistical value to the brightness value, which will not be repeated here.

In order to ensure the high frame rate of the surveillance image, the two surveillance image generation modules can be used for ping-pong operation. Through the "input data selection unit" and "output data selection unit", according to the rhythm and mutual cooperation, the buffered data stream is continuously sent to the "image data processing module" for calculation and processing. The use of ping-pong operation can complete the seamless buffering and processing of data, and can also save buffer space.

After a monitoring image generation module completes histogram statistics, it needs to map and output the histogram data of the image block. At this time, another monitoring image generation module can perform the histogram statistics of the next image block at this time. The two modules ping pong Interleaving ensures that the monitoring image generation module can process continuous high frame rate video source data. Ping-pong processing is not required when the frame rate of the monitoring image is not high. After reading the image of an image block from the image storage unit, wait for the histogram data statistics and mapping output to complete before reading the new data. The logic resources and on-chip RAM occupied by the monitoring image generation process are nearly doubled.

The surveillance image needs to be displayed on the display screen. The surveillance image obtained from the original large image may be too large to display. In order to solve the above problem, at least one of the following three methods can be adopted:

method one:

If the pixel width of the display screen is less than the number of columns required for the display of each image column block, the combined statistics of adjacent columns can be performed. Among them, the number of columns required for display refers to the number of column pixels occupied by the monitoring image of this image column when it is actually displayed on the display screen. Consolidated statistics refers to the use of one histogram for multiple columns. Assuming that the display requirements need to reduce the width of the monitored image by 3 times, then the three columns are combined for statistics. In this way, it is directly merged during statistics, and a small monitor image suitable for display is directly generated. There is no need to generate a large monitor image according to requirements and perform a proportional reduction when the monitoring image is displayed.

Way two:

If the pixel height of the display screen is less than 2q, the combined statistics of adjacent pixel values can be performed. For example, the 8-bit pixel statistical height is 256. When the height of the monitor image to be displayed is 128, the pixel value 0 and the pixel value 1 can be regarded as a pixel value statistics, ..., the pixel value 254 and the pixel value 255 is regarded as a pixel value statistics, and the size of the statistical histogram is 128. In this way, it is directly merged during statistics, and a small monitor image suitable for display is directly generated. There is no need to generate a large monitor image and scale it down according to the requirements when the monitor image is displayed.

Way three:

Before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space, each sub-monitoring image is respectively scaled. This solution can also be used for the display requirements of surveillance images of any resolution. You can first generate a surveillance image of a block of images with a fixed resolution, then zoom the surveillance image and then write it into the surveillance image storage space. The splicing of surveillance images can form surveillance images with the required resolution. It is also possible to generate a surveillance image with a fixed resolution, and the surveillance image of one frame of image can be scaled to obtain the surveillance image with the required resolution after the calculation is completed. For example, 1080p, 8-bit images all generate 1920*256 size surveillance images, and subsequent use of surveillance image scaling and splicing to display the surveillance images to the required resolution. In the case where an image frame is divided into multiple image column blocks, each image column block undergoes corresponding scaling.

In one embodiment, if the sub-monitoring image is a single-color monitoring image, the statistical step size for counting the number of pixels of each pixel value in the column is 3; if the sub-monitoring image is a three-color monitoring For the image, the statistical step size for the number of the pixel points of each pixel value in the column is 1 respectively. The statistical step size is 3, that is, statistical histogram statistics are calculated according to the method of adding 3 increments. Each time the number of pixels corresponding to a pixel value increases by 1, the statistical value corresponding to the pixel value is increased by 3. Since each statistical histogram counts image height pixels in monochrome statistics, and each statistical histogram counts image height*3 pixels in three-color statistics, the intention is to make each statistical histogram of monochrome statistics The value of is finally multiplied by 3, which is consistent with the statistical histogram of the three-color statistics, and the same brightness mapping can be performed.

When the statistical value under a certain statistical address reaches the highest upper limit T defined by the color component, the statistical value of the statistical histogram remains unchanged at T. Among them, the statistical address is the statistical pixel value of the histogram statistics. For example, suppose in the statistical histogram, the statistical value of the number of pixels with a pixel value of 0 is 25, and the statistical value of the number of pixels with a pixel value of 6 is 2. The statistical value of the number of pixels with a pixel value of 255 is 10, then 0, 6 and 255 are three statistical addresses respectively, and 25, 2 and 10 are the statistical values under these three statistical addresses respectively. This embodiment can limit the maximum statistical value of each pixel value to achieve the effect of eliminating the top of the histogram.

When the three colors are displayed at the same time, you can first output the first row of r histogram data (the maximum address of each histogram), then output the first row of g histogram data, the first row of b histogram data, and finally Output the last line of b histogram data. Other output sequences are also possible, just ensure that the three-color histogram data of an image column is written into the monitoring image storage space of the corresponding color in the DDR during output.

In an embodiment, before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space, the pixel points of the sub-monitoring image may also be subjected to row-by-row interpolation processing. A feasible interpolation method is: each row of data is stretched into two rows, except for the last row, the first row of expansion is the original data, the second row of expansion is the mean value of the original data of the current row and the next row; the last row is expanded The two rows of data are copies of the original data. When outputting, assuming that the sub-monitoring image includes K lines of data before interpolation, the sub-monitoring image can be expanded into 2K lines of data through interpolation. In practical applications, it can be determined whether to extend the sub-monitoring image according to the display requirements. The size of a normal surveillance image is the image width*2m, and m is the width of statistical pixel data. If the number of bits of the surveillance image is 8 bits, its height is 256. If it is displayed on a screen with a resolution of 1080p, it may be too small, so , The height of the surveillance image can be enlarged by interpolation. For example, when the height of the original surveillance image is 256, the height of the surveillance image can be enlarged to 512 by the above method.

The monitoring image mapping output process is shown in Figure 3. The first line of the original image stores a, e, ..., h, and the second line of the original image stores b, f, ..., j, then after expansion The first row of is the same as the first row of the original image, which is also a, e,..., h. The expanded second row is the average value of the first row and the second row of the original image, that is (a+b)/ 2, (e+f)/2, (h+j)/2, the third row after expansion is the same as the second row of the original image, also b, f,..., j, the fourth row after expansion is the original The average value of the 2nd and 3rd rows in the figure, and so on.

In addition to the above interpolation methods, other linear interpolation methods or nonlinear interpolation methods can also be used for interpolation, or image pixel replication can be used, which will not be repeated here.

Each of the generated sub-monitoring images can be written into the surveillance image storage space according to row jump addresses. When reading the image storage space, the data in each burst is continuously read, but the address is jumped between the burst queues, and the jump address is the corresponding position of the next line of the image. Take the burst length as 64 pixels, the monochrome display is jumped after each burst, and the three-color display jumps after the three bursts. The jump address ensures that when reading image data from the image storage space, each image column block is output in line from left to right and from top to bottom. When the surveillance image is written into the surveillance image storage space, it is also jumped to write to ensure that the surveillance image written into the surveillance image storage space is the continuous surveillance image to be displayed.

The structure of jump address reading and writing is shown in Figure 4. The images stored in DDR are read out from DMA (Direct Memory Access) by jump address according to the label from small to large. Assuming that each image column block includes 64 columns, the image has a total of 1080 lines, for a monochrome surveillance image, 64 columns are read at a time, and the 64 columns are read in rows from the 1st line to the 1080th line; for the three-color surveillance image , Read 192 columns each time, that is, read the first column of the r component first, then read the first column of the g component, and then read the first column of the b component, ..., until the 64th of the r component Column, the 64th column of the g component, and the 64th column of the b component. These 192 columns are read in rows from the 1st row to the 1080th row. The read image column blocks are counted by column histogram, and after the statistics are completed, the DMA is written into the DMA according to the write address from small to large. Assuming that 256 rows of data are interpolated and expanded into 512 rows, when writing, for a monochrome surveillance image, first write the first to 64th columns, and the 64 columns of data are written in rows from the 1st to the 512th row; For the three-color monitoring image, write the first to 192th columns first, and the 192 columns of data are written row by row from the first row to the 512th row.

Reading the image from the image storage space and writing the surveillance image into the surveillance image storage space are all subjected to addressing processing. The image stored in the image storage space is an entire frame image written line by line, and the read image is a line-by-line image outputted in image column blocks. The surveillance image written in the surveillance image storage space is an image block area of the entire surveillance image. After all the writing is completed, the surveillance image storage space is exactly assembled into a surveillance image stored row by row to be displayed. Both read and write jump addresses ensure the efficiency of data read and write.

The jump address sequence of Figure 4 can be arbitrarily specified, only the data of one image column block and the monitoring image of the corresponding image column block are written to the corresponding position. Changing the monitoring image output sequence in Figure 3, the jump location method in Figure 4, etc. can easily realize the upside down, left and right flipping of the monitoring image, and the monitoring image arrangement sequence of the three-color monitoring image display, etc.

In an embodiment, the image frame and its monitoring image can also be superimposed and displayed. The surveillance image stored in the surveillance image storage space is a monochrome brightness map. When the surveillance image is displayed, it is necessary to read the corresponding brightness data from the surveillance image storage space and map the surveillance image representation for each color component to the corresponding color component. The mapped surveillance image needs to be displayed superimposed with the original video. The superposition can use the blend algorithm, which is generally to superimpose the surveillance image on the original image after being color-mapped.

For a 4k high-resolution image that processes multiple pixels in one clock, the monitoring image statistics RAM used for the storage of one image column can be divided into several interleaved copies, so that all RAMs can perform histogram statistics at the same time within one clock , It can also output multiple pixels of the monitoring image in one clock to ensure real-time processing. For example, suppose that a clock processes 4 pixels and 4 statistics are in parallel. The original image column block uses a large RAM to realize statistics. Now 4 interleaved small RAMs are used for statistics. The first RAM counts the first column and the first RAM. 5 columns, 9th column, second RAM statistics, 2nd column, 6th column, 10th column, etc. It does not occupy much RAM resources, but the statistical logic is multiplexed 4 times, and other logic resources are not multiplexed. Not much resources.

The embodiments of this specification have the following advantages:

(1) It can realize the generation of surveillance images of 4k ultra-large resolution video, and the processing frequency of surveillance images can be consistent with the input frequency.

(2) Fine histogram statistics can be performed for each column of the image to generate large-resolution, high-precision and delicate monitoring images, which can not only realize monochrome monitoring images, but also realize simultaneous calculation and simultaneous display of monitoring images with multiple color components .

(3) The generated surveillance image is the surveillance image organized according to display requirements, and no additional stitching or data reformation is required.

As shown in FIG. 5, it is a schematic diagram of a surveillance image generating apparatus 500 according to an embodiment of this specification. The monitoring image generating device 500 includes:

Histogram statistics unit 501, data processing unit 502 and data output unit 503;

For each image column block in the image frame, the histogram statistics unit 501 is configured to generate a statistical histogram of each column of the image column block, and send the statistical histogram to the data processing unit 502; wherein, The image frame is divided into a plurality of image column blocks in advance;

The data processing unit 502 is configured to generate a sub-monitoring image corresponding to the image column block according to the statistical histogram of each column of the image column block, and send the sub-monitoring image to the data output unit 503;

The data output unit 503 is configured to write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame.

In the embodiment of this specification, the image frame in the image storage space is pre-divided into a plurality of image column blocks (tile), and the image frame is read block by block according to the pre-divided image column blocks, where each image The column blocks are read in a row-by-row reading manner, and the monitoring image calculation of the entire image frame is realized by calculating the monitoring images of the image column blocks one by one. The embodiments of this specification can be implemented based on FPGA, and both the original image frame and the generated surveillance image can be stored in the DDR.

In the above embodiment, the image frame is read row by row from the image storage space in the manner of image column blocks. Since several columns can be read at a time, the image reading efficiency is ensured. Divide the entire image frame into N image column blocks, and the processing unit only needs to open up a storage space that can store the histogram of each image column block, so the resource consumption only needs 1/N of the entire monitoring image storage, which is greatly reduced Storage resources, the use of smaller on-chip RAM can realize the generation of large-scale surveillance images. Therefore, the device according to the embodiments of the present specification can realize the monitoring image processing of high-resolution images, and realize the generation of high-resolution, high-frame-rate monitoring images.

In an embodiment, the amount of read and write data when reading and writing the image frame once can be determined according to the preset data read and write efficiency, and then the image sequence divided by the image frame can be determined according to the amount of read and write data. The number of blocks. Specifically, the number of image columns M of each image column block in the monochrome surveillance image satisfies:

M*q=burst amount*burst width;

The number of image columns M of each image column block in the three-color surveillance image satisfies:

3M*q=burst amount*burst width.

In one embodiment, the step of generating the sub-monitoring image corresponding to the image column block by the data processing unit 502 according to the pixel value of the pixel point includes: respectively generating a statistical histogram of each column in the image column block; And generate the sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block, wherein the statistical histogram of each column in the image column block is generated according to the following method: The number of pixels of each pixel value in the column is counted to obtain the statistical value of the corresponding pixel value in the column; and the statistical histogram of the column is generated according to the statistical value of each pixel value in the column.

The sub-monitoring image can be generated directly based on the statistical histogram, or based on the brightness of the image. Specifically, the step of generating the sub-monitoring image corresponding to the image column block by the data processing unit 502 according to the statistical histogram of each column in the image column block includes: calculating the statistical histogram of each column in the image column block. The images are spliced to obtain the sub-monitoring image corresponding to the image column block; or each statistical value is mapped to the corresponding brightness value, and the brightness distribution map of the corresponding column is generated according to the brightness value of each column, and the image column block The brightness distribution maps of each column are spliced to obtain sub-monitoring images corresponding to the image column blocks.

In one embodiment, each statistical value may be mapped separately, and the mapping methods include linear mapping method, nonlinear mapping method, table look-up mapping method, etc., after mapping, the brightness value corresponding to the statistical value is obtained. Taking the linear mapping method as an example, in one embodiment, the step of the data processing unit 502 linearly mapping each statistical value includes:

Perform linear mapping on the statistical values in the following way:

Y=histo(x)*(2 ^k -1)/T;

In one embodiment, when the data processing unit 502 obtains the statistical value, it may perform combined statistics of adjacent columns. Among them, the number of columns required for display refers to the number of column pixels occupied by the monitoring image of this image column when it is actually displayed on the display screen. Consolidated statistics refers to the use of one histogram for multiple columns. Assuming that the display requirements need to reduce the width of the monitored image by 3 times, the three columns are combined for statistics. In this way, it is directly merged during statistics, and a small monitor image suitable for display is directly generated. There is no need to generate a large monitor image and scale it down according to the requirements when the monitor image is displayed.

In one embodiment, when the data processing unit 502 obtains the statistical value, it may perform the combined statistics of the adjacent pixel values. For example, the 8-bit pixel statistical height is 256. When the height of the monitor image to be displayed is 128, the pixel value 0 and the pixel value 1 can be regarded as a pixel value statistics, ..., the pixel value 254 and the pixel value 255 is regarded as a pixel value statistics, and the size of the statistical histogram is 128. In this way, it is directly merged during statistics, and a small monitor image suitable for display is directly generated. There is no need to generate a large monitor image and scale it down according to the requirements when the monitor image is displayed.

In an embodiment, the data processing unit 502 may also perform scaling processing on each sub-monitoring image, and then send the scaled sub-monitoring image to the data output unit 503. This solution can also be used for the display requirements of surveillance images of any resolution. You can first generate a surveillance image of a block of images with a fixed resolution, then zoom the surveillance image and then write it into the surveillance image storage space. The splicing of surveillance images can form surveillance images with the required resolution. It is also possible to generate a surveillance image with a fixed resolution, and the surveillance image of one frame of image can be scaled to obtain the surveillance image with the required resolution after the calculation is completed. For example, 1080p, 8-bit images all generate 1920*256 size surveillance images, and subsequent use of surveillance image scaling and splicing to display the surveillance images to the required resolution. In the case where an image frame is divided into multiple image column blocks, each image column block undergoes corresponding scaling.

In one embodiment, if the sub-monitoring image is a single-color monitoring image, the statistical step size for counting the number of pixels of each pixel value in the column is 3; if the sub-monitoring image is a three-color monitoring For the image, the statistical step size for the number of the pixel points of each pixel value in the column is 1 respectively. The statistical step size is 3, that is, the statistical histogram statistics are calculated according to the method of adding 3 increments. Each time the number of pixels corresponding to a pixel value increases by 1, the statistical value corresponding to the pixel value is increased by 3. Since each statistical histogram counts image height pixels in monochrome statistics, and each statistical histogram counts image height*3 pixels in three-color statistics, the intention is to make each statistical histogram of monochrome statistics The value of is finally multiplied by 3, which is consistent with the statistical histogram of the three-color statistics, and the same brightness mapping can be performed.

In an embodiment, the data processing unit 502 may also perform interpolation processing on pixel clicks of the sub-monitoring image, and then send the scaled sub-monitoring image to the data output unit 503.

The data output unit 503 may write each of the generated sub-monitoring images into the surveillance image storage space by line jump. When reading the image storage space, the data in each burst is continuously read, but the address is jumped between the burst queues, and the jump address is the corresponding position of the next line of the image. Take the burst length as 64 pixels, the monochrome display is jumped after each burst, and the three-color display jumps after the three bursts. The jump address ensures that when reading image data from the image storage space, each image column block is output in line from left to right and from top to bottom. When the surveillance image is written into the surveillance image storage space, it is also jumped to write to ensure that the surveillance image written into the surveillance image storage space is the continuous surveillance image to be displayed.

In an embodiment, the number of the data processing unit 502 is multiple; each data processing unit 502 is respectively configured to generate the image column according to the statistical histogram of each column of the image column block on one of the color components The sub-monitoring image of the block on the corresponding color component. For monochrome surveillance image generation, one data processing unit can be used to generate statistical histograms and obtain sub-monitoring images; for three-color surveillance image generation, three data processing units can be used to generate statistical histograms and obtain sub-monitoring images. Each data processing unit generates a sub-monitoring image on the color component. An embodiment of the three-color surveillance image generating device is shown in FIG. 6.

In one embodiment, the data processing unit is RAM. For a 4k high-resolution image that processes multiple pixels in one clock, the monitoring image statistics RAM used for the storage of one image column can be divided into several interleaved copies, so that all RAMs can perform histogram statistics at the same time within one clock , It can also output multiple pixels of the monitoring image in one clock to ensure real-time processing. For example, suppose that a clock processes 4 pixels and 4 statistics are in parallel. The original image column block uses a large RAM to realize statistics. Now 4 interleaved small RAMs are used for statistics. The first RAM counts the first column and the first RAM. 5 columns, 9th column, second RAM statistics, 2nd column, 6th column, 10th column, etc. It does not occupy much RAM resources, but the statistical logic is multiplexed 4 times, and other logic resources are not multiplexed. Not much resources.

Other embodiments of the monitoring image generation device are the same as the above-mentioned embodiments of the monitoring image generation method, and will not be repeated here.

As shown in FIG. 7, it is a schematic diagram of a surveillance image generation system according to an embodiment of this specification. The surveillance image generation system 700 may include: the surveillance image generation device 500 described in any of the above embodiments; and a storage unit 701 for storing the image frame. Further, the storage unit 701 may also be used to store the generated surveillance image.

In one embodiment, the monitoring image generating device 500 includes a first monitoring image generating device 500a and a second monitoring image generating device 500b; the second monitoring image generating device 500b is used for the first monitoring image generating device When 500a outputs the sub-monitoring image corresponding to one of the image column blocks, the sub-monitoring image corresponding to the next image column block is generated.

This embodiment exemplifies two monitoring image generation modules to perform ping-pong operation processing. After a monitoring image generation module completes histogram statistics, it needs to map and output the histogram data of the image block. At this time, another monitoring image generation module can perform the histogram statistics of the next image block at this time. The two modules ping pong Interleaving ensures that the monitoring image generation module can process continuous high frame rate video source data. Ping-pong processing is not required when the frame rate of the monitoring image is not high. After reading the image of an image block from the image storage unit, wait for the histogram data statistics and mapping output to complete before reading the new data. The logic resources and on-chip RAM occupied by the monitoring image generation process are nearly doubled.

In an embodiment, the surveillance image generation system 700 further includes a data distribution unit 702; the data distribution unit 702 is configured to read the image column blocks of the image frame from the storage unit 701, and combine the The image sequence block is distributed to the first surveillance image generating device 500a or the second surveillance image generating device 500b. The monitoring image generation system of an embodiment is shown in FIG. 8. The system in FIG. 8 includes two monitoring module generating devices and a distribution unit. In practical applications, the number of monitoring module generating devices can also be more than two.

In an embodiment, the surveillance image generation system 700 further includes a display unit 703 for displaying the surveillance image. In the display unit 703, the image frame and its monitoring image can be superimposed and displayed. The surveillance image stored in the surveillance image storage space is a monochrome brightness map. When the surveillance image is displayed, it is necessary to read the corresponding brightness data from the surveillance image storage space and map the surveillance image representation for each color component to the corresponding color component. The mapped surveillance image needs to be displayed superimposed with the original video. The superposition can use the blend algorithm, which is generally to superimpose the surveillance image on the original image after being color-mapped.

In one embodiment, the storage unit 701 is a DDR.

In one embodiment, the monitoring image generation system 700 is implemented based on FPGA.

The embodiments of this specification have the following advantages:

The embodiments of this specification also provide an image processing device, including the surveillance image generation system described in any of the above embodiments.

In one embodiment, the image processing device is an image monitor or a camera.

The embodiment of this specification also provides a monitoring image generating device, and the embodiment of the method of this specification can be implemented by the monitoring image generating device.

In one embodiment, the monitoring image generation device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the following method when the program is executed:

Reading the pixels in the image column block by row;

In one embodiment, the processor further implements the following method when executing the program: determining the amount of read and write data when reading and writing the image frame once according to a preset data read and write efficiency; The amount of data determines the number of image column blocks divided by the image frame.

In one embodiment, the step of generating the sub-monitoring image corresponding to the image column block by the processor according to the pixel value of the pixel point includes: respectively generating a statistical histogram of each column in the image column block, and according to The statistical histogram of each column in the image column block generates the sub-monitoring image corresponding to the image column block, wherein the statistical histogram of each column in the image column block is generated according to the following method: The number of pixel points of each pixel value is counted to obtain the statistical value of the corresponding pixel value in the column; and the statistical histogram of the column is generated according to the statistical value of each pixel value in the column.

In an embodiment, the step of the processor generating the sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block includes: calculating the statistical histogram of each column in the image column block The images are spliced to obtain the sub-monitoring image corresponding to the image column block; or each statistical value is respectively mapped to the corresponding brightness value, and the brightness distribution map of the corresponding column is generated according to the brightness value of each column, and the image column block The brightness distribution maps of each column are spliced to obtain sub-monitoring images corresponding to the image column blocks.

In one embodiment, the processor further implements the following method when executing the program: merging and counting the pixels corresponding to the same pixel value in several columns; and/or corresponding to several adjacent pixel values in the same column Pixels are combined and counted.

In one embodiment, if the sub-monitoring image is a single-color monitoring image, the statistical step size for counting the number of pixels of each pixel value in the column is 3; if the sub-monitoring image is a three-color monitoring For the image, the statistical step size for the number of the pixel points of each pixel value in the column is 1 respectively.

In an embodiment, the step of the processor respectively mapping each statistical value to a corresponding brightness value includes: respectively mapping each statistical value to obtain the brightness value corresponding to the statistical value.

In an embodiment, the step of the processor respectively mapping each statistical value includes: linearly mapping the statistical value in the following manner:

Y=histo(x)*(2 ^k -1)/T.

In one embodiment, before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block in the monitoring image storage space, the processor further implements the following method when executing the program: The pixel points of the sub-monitoring image are interpolated line by line.

In an embodiment, each sub-monitoring image is written into the surveillance image storage space by row jump address.

In an embodiment, the processor further implements the following method when executing the program: superimposing and displaying the image frame and its monitoring image.

In one embodiment, before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block in the monitoring image storage space, the processor further implements the following method when executing the program: The sub-monitoring image is scaled.

Other embodiments of the method executed by the processor are the same as the foregoing embodiments of the monitoring image generation method, and will not be repeated here.

The monitoring image generation device of the embodiment of this specification may be, for example, a server or a terminal device. The method embodiments can be implemented by software, or can be implemented by hardware or a combination of software and hardware. Taking software implementation as an example, as a logical device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory by the processor that processes the file where it is located. From a hardware perspective, as shown in FIG. 9, it is a hardware structure diagram of a surveillance image generating device that implements the method of this specification, except for the processor 901, memory 902, network interface 903, and non-volatile memory shown in FIG. In addition to the sexual memory 904, the monitoring image generating device used to implement the method of this specification in the embodiment may also include other hardware according to the actual function of the monitoring image generating device, which will not be repeated here.

The various technical features in the above embodiments can be combined arbitrarily, as long as there is no conflict or contradiction between the combinations of features, but due to space limitations, they are not described one by one. Therefore, the various technical features in the above embodiments can be combined arbitrarily. It also belongs to the scope of the disclosure of this specification.

The embodiments of this specification may adopt the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing program codes. Computer usable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices.

Those skilled in the art will easily think of other embodiments of the present disclosure after considering the specification and practicing the specification disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptive changes of the present disclosure. These variations, uses, or adaptive changes follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field that are not disclosed in the present disclosure. . The description and the embodiments are to be regarded as exemplary only, and the true scope and spirit of the present disclosure are pointed out by the following claims.

It should be understood that the present disclosure is not limited to the precise structure that has been described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

The above are only the preferred embodiments of the present disclosure and are not intended to limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the present disclosure. Within the scope of protection.

Claims

A method for generating surveillance images, characterized in that the method includes:

Write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame, wherein the image frame is divided into a plurality of image column blocks in advance Each image column block includes multiple columns of pixels of the image frame, and the sub-monitoring image corresponding to each image column block is generated in the following manner:

Reading the pixels in the image column block by row;

The sub-monitoring image corresponding to the image column block is generated according to the pixel value of the pixel point.
The method according to claim 1, wherein the method further comprises:

Determine the amount of read and write data when reading and writing the image frame once according to the preset data read and write efficiency;

The number of image column blocks divided by the image frame is determined according to the amount of read and write data.
The method according to claim 1, wherein the step of generating the sub-monitoring image corresponding to the image column block according to the pixel value of the pixel point comprises:

Generate a statistical histogram of each column in the image column block respectively, and generate a sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block, wherein each of the image column blocks The statistical histograms of the columns are all generated according to the following methods:

Performing quantitative statistics on the pixel points of each pixel value in the column respectively to obtain the statistical value of the corresponding pixel value in the column;

A statistical histogram of the column is generated according to the statistical value of each pixel value in the column.
The method according to claim 3, wherein the step of generating the sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block comprises:

Splicing the statistical histograms of each column in the image column block to obtain the sub-monitoring image corresponding to the image column block; or

Each statistical value is mapped to the corresponding brightness value, the brightness distribution map of the corresponding column is generated according to the brightness value of each column, and the brightness distribution map of each column of the image column block is spliced to obtain the corresponding image column block. Sub-monitor image.
The method according to claim 3, wherein the method further comprises:

Merging and counting the pixels corresponding to the same pixel value in several columns; and/or

The pixel points corresponding to several adjacent pixel values in the same column are combined and counted.
The method according to claim 3, wherein if the sub-monitoring image is a monochromatic monitoring image, the statistical step size for performing quantitative statistics on the pixel points of each pixel value in the column is 3; if the sub-monitoring image is a monochromatic monitoring image, The sub-monitoring image is a three-color monitoring image, and the statistical step size of the pixel points of each pixel value in the column is 1 respectively.
The method according to claim 4, wherein the step of respectively mapping each statistical value to a corresponding brightness value comprises:

Each statistical value is respectively mapped to obtain the brightness value corresponding to the statistical value.
8. The method according to claim 7, wherein the step of respectively mapping each statistical value comprises:

Perform linear mapping on the statistical values in the following way:

Y=histo(x)*(2 k -1)/T;

In the formula, histo(x) is the statistical value of the number of pixels with the pixel value x, T is the preset upper limit of the statistical value, k is the number of bits representing the brightness value, and Y is the brightness value.
The method according to claim 1, characterized in that, before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space, the method further comprises:

Perform interpolation processing on pixel points of the sub-monitoring image by line.
The method according to claim 1, wherein each sub-monitoring image is written into the surveillance image storage space by row jump address.
The method according to claim 1, wherein the method further comprises:

The image frame and its monitoring image are superimposed and displayed.
The method according to claim 1, characterized in that, before writing the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space, the method further comprises:

Respectively perform zoom processing on each sub-monitoring image.
A monitoring image generation device is characterized by comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the following method when the program is executed:

Write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame, wherein the image frame is divided into a plurality of image column blocks in advance Each image column block includes multiple columns of pixels of the image frame, and the sub-monitoring image corresponding to each image column block is generated in the following manner:

Reading the pixels in the image column block by row;

The sub-monitoring image corresponding to the image column block is generated according to the pixel value of the pixel point.
The monitoring image generation device according to claim 13, wherein the processor further implements the following method when executing the program:

Determine the amount of read and write data when reading and writing the image frame once according to the preset data read and write efficiency;

The number of image column blocks divided by the image frame is determined according to the amount of read and write data.
The monitoring image generation device according to claim 13, wherein the step of the processor generating the sub-monitoring image corresponding to the image column block according to the pixel value of the pixel point comprises:

Generate a statistical histogram of each column in the image column block respectively, and generate a sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block, wherein each of the image column blocks The statistical histograms of the columns are all generated according to the following methods:

Performing quantitative statistics on the pixel points of each pixel value in the column respectively to obtain the statistical value of the corresponding pixel value in the column;

A statistical histogram of the column is generated according to the statistical value of each pixel value in the column.
The monitoring image generating device according to claim 15, wherein the step of the processor generating the sub-monitoring image corresponding to the image column block according to the statistical histogram of each column in the image column block comprises:

Splicing the statistical histograms of each column in the image column block to obtain the sub-monitoring image corresponding to the image column block; or

Each statistical value is mapped to the corresponding brightness value, the brightness distribution map of the corresponding column is generated according to the brightness value of each column, and the brightness distribution map of each column of the image column block is spliced to obtain the corresponding image column block. Sub-monitor image.
The monitoring image generation device according to claim 15, wherein the processor further implements the following method when executing the program:

Merging and counting the pixels corresponding to the same pixel value in several columns; and/or

The pixel points corresponding to several adjacent pixel values in the same column are combined and counted.
The monitoring image generating device according to claim 15, wherein if the sub-monitoring image is a monochromatic monitoring image, the statistical step size for respectively counting the number of pixels of each pixel value in the column is 3; If the sub-monitoring image is a three-color monitoring image, the statistical step size for counting the number of pixels of each pixel value in the column is 1.
The monitoring image generation device according to claim 15, wherein the step of the processor respectively mapping each statistical value to a corresponding brightness value comprises:

Each statistical value is respectively mapped to obtain the brightness value corresponding to the statistical value.
The monitoring image generation device according to claim 19, wherein the step of the processor respectively mapping each statistical value comprises:

Perform linear mapping on the statistical values in the following way:

Y=histo(x)*(2 k -1)/T;

In the formula, histo(x) is the statistical value of the number of pixels with the pixel value x, T is the preset upper limit of the statistical value, k is the number of bits representing the brightness value, and Y is the brightness value.
The monitoring image generation device according to claim 13, wherein the processor executes the sub-monitoring image corresponding to each image column block in the image frame on the corresponding column block of the monitoring image storage space. The program also implements the following methods:

Perform interpolation processing on pixel points of the sub-monitoring image by line.
The monitoring image generating device according to claim 13, wherein each sub-monitoring image is written into the monitoring image storage space by row jump address.
The monitoring image generation device according to claim 13, wherein the processor further implements the following method when executing the program:

The image frame and its monitoring image are superimposed and displayed.
The monitoring image generation device according to claim 13, wherein the processor executes the sub-monitoring image corresponding to each image column block in the image frame on the corresponding column block of the monitoring image storage space. The program also implements the following methods:

Respectively perform zoom processing on each sub-monitoring image.
A monitoring image generation device, characterized in that it further comprises:

Histogram statistical unit, data processing unit and data output unit;

For each image column block in the image frame, the histogram statistical unit is used to generate a statistical histogram of each column of the image column block, and send the statistical histogram to the data processing unit; wherein, the The image frame is pre-divided into multiple image column blocks;

The data processing unit is configured to generate a sub-monitoring image corresponding to the image column block according to the statistical histogram of each column of the image column block, and send the sub-monitoring image to a data output unit;

The data output unit is used to write the sub-monitoring image corresponding to each image column block in the image frame to the corresponding column block of the monitoring image storage space to obtain the monitoring image of the image frame.
The surveillance image generating device according to claim 25, wherein the number of the data processing unit is multiple;

Each data processing unit is respectively configured to generate a sub-monitoring image of the image column block on the corresponding color component according to the statistical histogram of each column of the image column block on one of the color components.
The surveillance image generation device according to claim 26, wherein the data processing unit is a RAM.
A monitoring image generation system is characterized in that it comprises:

The monitoring image generating device according to any one of claims 25 to 27; and

A storage unit for storing the image frame.
The surveillance image generation system according to claim 28, wherein the surveillance image generation device comprises a first surveillance image generation device and a second surveillance image generation device;

The second monitoring image generating device is used for generating the sub-monitoring image corresponding to the next image column block when the first monitoring image generating device outputs the sub-monitoring image corresponding to one of the image column blocks.
The surveillance image generation system according to claim 29, further comprising:

Data distribution unit;

The data distribution unit is configured to read the image sequence block of the image frame from the storage unit, and distribute the image sequence block to the first surveillance image generation device or the second surveillance image generation device .
The surveillance image generation system according to claim 30, further comprising:

A display unit for displaying the monitoring image.
The surveillance image generation system according to claim 28, wherein the storage unit is a DDR.
The surveillance image generation system according to claim 28, wherein the surveillance image generation system is implemented based on FPGA.
An image processing device, characterized by comprising the surveillance image generation system according to any one of claims 28 to 33.
The image processing device according to claim 34, wherein the image processing device is an image monitor or a camera.