Image data analysis system based on algorithm analysis

Abstract

The invention relates to an image data analysis system based on algorithm analysis, and the system comprises a block monitoring device which is disposed right above the central position of a block, and is used for carrying out the image monitoring of a real-time scene of the block in a high-angle shooting mode, so as to obtain and output a current block image; and the amplitude comparison mechanism is used for outputting the filtering algorithm corresponding to the minimum value in the maximum noise amplitudes output by the multiple filtering algorithms through the AI prediction model as a reference filtering algorithm. The image data analysis system based on algorithm analysis is compact in logic and wide in application. The intelligent prediction of the noise amplitude of the filtered image corresponding to each filtering algorithm based on the AI prediction model can be performed on the image data after the block monitoring image is preprocessed. And the filtering algorithm corresponding to the minimum value in the noise amplitudes corresponding to the various filtering algorithms is used as the reference filtering algorithm, so that the filtering effect of each frame of monitoring image is ensured.

Description

Technical Field

[0001] The present invention relates to the field of image monitoring, and in particular to an image data analysis system based on algorithm analysis. Background Art

[0002] Blocks are an important factor in urban planning and urban design. The blocks that people usually refer to are urban blocks. The smallest granularity of the population census is also a block. Each census content can be counted according to the block. Therefore, in demography, the block can be called a demographic block. Large real estate projects are divided into several blocks. They are the result of the first and second phases of a real estate brand. In real estate projects, it is usually a commercial concept.

[0003] In cities, blocks are usually areas surrounded by roads and are the basic building blocks of the urban structure. However, in suburbs and towns, roads are not as dense as in cities. Therefore, demographic blocks are divided and segmented with the help of other natural or cultural features, such as administrative boundaries, rivers, lakes, railways, mountains, cliffs, etc.

[0004] Due to the complexity of street scenes and the diversity of content over time, the filtering algorithms selected for different time-sharing acquisition images of the same street scene may be different. How to select a suitable filtering algorithm for the acquisition images of the same type of street scenes to ensure the best filtering effect for different acquisition images is one of the difficult problems that need to be solved in the existing technology. Summary of the invention

[0005] In order to solve the technical problems in the prior art, the present invention provides an image data analysis system based on algorithm analysis, which can perform intelligent prediction of the noise amplitude of the filtered image corresponding to each filtering algorithm based on the AI ​​prediction model on the preprocessed image data of the block monitoring image, and put the filtering algorithm corresponding to the minimum value of the noise amplitudes corresponding to the various filtering algorithms into use as the reference filtering algorithm, thereby ensuring the filtering effect of each frame of the monitoring image.

[0006] The present invention provides an image data analysis system based on algorithm analysis, comprising: The block monitoring device is arranged just above the center of the block and is used to monitor the real-time scene of the block in a bird's-eye view mode to obtain and output the current block image; A smoothing processing device, connected to the block monitoring device, is used to perform a non-scaling transformation blur process on the received current block image to obtain and output a corresponding smoothed image; A data sharpening device, connected to the smoothing device, for performing a spatial differential sharpening process on the received smoothed image to obtain and output a corresponding data sharpened image; A real-time enhancement device, connected to the data sharpening device, for performing contrast enhancement processing on the received data sharpened image to obtain and output a corresponding real-time enhanced image; A content recognition mechanism is arranged in a monitoring room of a block and connected to the real-time enhancement device, and is used to identify each edge pixel point in the received real-time enhancement image, and to combine each edge pixel point to obtain each distribution area in the real-time enhancement image; A region processing mechanism connected to the content identification mechanism, used to detect multiple information parameters of each distribution region to obtain multiple information parameters of each distribution region, wherein the multiple information parameters of each distribution region are color component values ​​corresponding to each pixel point in the partitioned region, and the color component value corresponding to each pixel point includes a red and green component value, a black and white component value, and a yellow and blue component value corresponding to the pixel point; A directional prediction mechanism, connected to the regional processing mechanism, for inputting a plurality of information parameters of each distribution area, the clarity of the real-time enhanced image, the number of edge pixels, and the algorithm identifier corresponding to the current filtering algorithm into the AI ​​prediction model and running the AI ​​prediction model to obtain the maximum noise amplitude of the picture content obtained after the current filtering algorithm is executed on the real-time enhanced image output by the AI ​​prediction model, wherein the AI ​​prediction model is a feedforward deep network after completing each training; The amplitude comparison mechanism is connected to the directional prediction mechanism and is used to output the filtering algorithm corresponding to the minimum value of each maximum noise amplitude output by multiple filtering algorithms through the AI ​​prediction model as a reference filtering algorithm.

[0007] The algorithm-based image data analysis system of the present invention has compact logic and wide application. By performing intelligent prediction of the noise amplitude of the filtered image corresponding to each filtering algorithm based on the AI ​​prediction model on the pre-processed image data of the block monitoring image, the filtering algorithm corresponding to the minimum value of each noise amplitude corresponding to each filtering algorithm is used as the reference filtering algorithm, thereby ensuring the filtering effect of each frame of the monitoring image. DETAILED DESCRIPTION

[0008] The following is a detailed description of an embodiment of the image data analysis system based on algorithm analysis of the present invention.

[0009] A. Embodiment

[0010] According to embodiment A of the present invention, an image data analysis system based on algorithm analysis is shown, and the system comprises: The block monitoring device is arranged just above the center of the block and is used to monitor the real-time scene of the block in a bird's-eye view mode to obtain and output the current block image; For example, a block monitoring device is arranged just above the center of a block and is used to monitor the real-time scene of the block in a bird's-eye view mode to obtain and output the current block image, including: the block monitoring device includes a field of view adjustment unit and an imaging unit, the field of view adjustment unit is connected to the imaging unit, and is used to adjust the imaging field of view of the imaging unit, so that the real-time scene of the entire block can be obtained; A smoothing processing device, connected to the block monitoring device, is used to perform a non-scaling transformation blur process on the received current block image to obtain and output a corresponding smoothed image; A data sharpening device, connected to the smoothing device, for performing a spatial differential sharpening process on the received smoothed image to obtain and output a corresponding data sharpened image; A real-time enhancement device, connected to the data sharpening device, for performing contrast enhancement processing on the received data sharpened image to obtain and output a corresponding real-time enhanced image; A content recognition mechanism is arranged in a monitoring room of a block and connected to the real-time enhancement device, and is used to identify each edge pixel point in the received real-time enhancement image, and to combine each edge pixel point to obtain each distribution area in the real-time enhancement image; A region processing mechanism connected to the content identification mechanism, used to detect multiple information parameters of each distribution region to obtain multiple information parameters of each distribution region, wherein the multiple information parameters of each distribution region are color component values ​​corresponding to each pixel point in the partitioned region, and the color component value corresponding to each pixel point includes a red and green component value, a black and white component value, and a yellow and blue component value corresponding to the pixel point; A directional prediction mechanism, connected to the regional processing mechanism, for inputting a plurality of information parameters of each distribution area, the clarity of the real-time enhanced image, the number of edge pixels, and the algorithm identifier corresponding to the current filtering algorithm into the AI ​​prediction model and running the AI ​​prediction model to obtain the maximum noise amplitude of the picture content obtained after the current filtering algorithm is executed on the real-time enhanced image output by the AI ​​prediction model, wherein the AI ​​prediction model is a feedforward deep network after completing each training; An amplitude comparison mechanism, connected to the directional prediction mechanism, is used to output the filtering algorithm corresponding to the minimum value of each maximum noise amplitude output by the multiple filtering algorithms through the AI ​​prediction model as a reference filtering algorithm; Among them, various filtering algorithms include statistical sorting filtering algorithm, maximum filtering algorithm, minimum filtering algorithm, midpoint filtering algorithm, alpha mean filtering algorithm, adaptive filtering algorithm, bilateral filtering algorithm, guided filtering algorithm and box filtering algorithm.

[0011] B. Embodiment

[0012] Embodiment B according to the present invention shows an image data analysis system based on algorithm analysis.

[0013] Different from the A embodiment, the algorithm analysis-based image data analysis system in the B embodiment may further include the following components: a temperature detection device, arranged near the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and connected to the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism respectively; Among them, the temperature detection device is arranged near the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and is respectively connected to the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism, including: the temperature detection device is used to respectively realize on-site measurement of the current temperature of the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism.

[0014] C Example

[0015] Embodiment C according to the present invention shows an image data analysis system based on algorithm analysis.

[0016] Different from the A embodiment, the algorithm analysis-based image data analysis system in the C embodiment may further include the following components: a voltage detection device, arranged near the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and connected to the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism respectively; Among them, the voltage detection device is arranged near the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and is respectively connected to the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism, including: the voltage detection device is used to respectively realize on-site measurement of the current voltage of the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism.

[0017] Next, the specific structure of the image data analysis system based on algorithm analysis of the present invention will be further described.

[0018] In the image data analysis system based on algorithm analysis according to various embodiments of the present invention: A SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

[0019] In the image data analysis system based on algorithm analysis according to various embodiments of the present invention: Using a SOC chip to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively includes: performing nearest neighbor interpolation processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

[0020] In the image data analysis system based on algorithm analysis according to various embodiments of the present invention: Using a SOC chip to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively includes: performing bilinear interpolation processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

[0021] In the image data analysis system based on algorithm analysis according to various embodiments of the present invention: The SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively, including: performing bicubic interpolation processing of a 16×16 pixel neighborhood on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

[0022] And in the image data analysis system based on algorithm analysis according to various embodiments of the present invention: The SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively, including: performing Lanczos interpolation processing of a 4×4 pixel neighborhood on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

[0023] In addition, in the image data analysis system based on algorithm analysis, multiple information parameters of each distribution area, the clarity of the real-time enhanced image, the number of edge pixels, and the algorithm identifier corresponding to the current filtering algorithm are input into the AI ​​prediction model and the AI ​​prediction model is run to obtain the maximum noise amplitude of the picture content obtained after the current filtering algorithm is executed on the real-time enhanced image output by the AI ​​prediction model. The AI ​​prediction model is a feedforward deep network after completing each training, including: using a numerical simulation mode to implement modeling and simulation processing of the AI ​​prediction model.

[0024] The technical effects of the present invention are as follows: (1): Identify each edge pixel point in the received real-time enhanced image, combine the each edge pixel point to obtain each distribution area in the real-time enhanced image, detect multiple information parameters of each distribution area to obtain multiple information parameters of each distribution area, the multiple information parameters of each distribution area are color component values ​​corresponding to each pixel point in the partition area, and the color component value corresponding to each pixel point includes a red and green component value, a black and white component value, and a yellow and blue component value corresponding to the pixel point; (2): Input multiple information parameters of each distribution area, the clarity of the real-time enhanced image, the number of edge pixels, and the algorithm identifier corresponding to the current filtering algorithm into the AI ​​prediction model and run the AI ​​prediction model to obtain the maximum noise amplitude of the picture content obtained after the current filtering algorithm is executed on the real-time enhanced image output by the AI ​​prediction model, wherein the AI ​​prediction model is a feedforward deep network after completing each training; (3): After preprocessing the image data of the block surveillance image, intelligent prediction of the noise amplitude of the filtered image corresponding to each filtering algorithm based on the AI ​​prediction model is performed. The filtering algorithm corresponding to the minimum value of the noise amplitudes corresponding to various filtering algorithms is used as the reference filtering algorithm, thereby ensuring the filtering effect of each frame of the surveillance image.

[0025] Thus, the objects and advantages can be achieved through the present invention. Although preferred specific examples have been disclosed and described in detail herein, the scope and objects of the present invention are not limited thereto. The present invention is defined by the appended claims.

Claims

1. An image data analysis system based on algorithm analysis, characterized in that: The system comprises: The block monitoring device is arranged just above the center of the block and is used to monitor the real-time scene of the block in a bird's-eye view mode to obtain and output the current block image; A smoothing processing device, connected to the block monitoring device, is used to perform a non-scaling transformation blur process on the received current block image to obtain and output a corresponding smoothed image; A data sharpening device, connected to the smoothing device, for performing a spatial differential sharpening process on the received smoothed image to obtain and output a corresponding data sharpened image; A real-time enhancement device, connected to the data sharpening device, for performing contrast enhancement processing on the received data sharpened image to obtain and output a corresponding real-time enhanced image; A content recognition mechanism is arranged in a monitoring room of a block and connected to the real-time enhancement device, and is used to identify each edge pixel point in the received real-time enhancement image, and to combine each edge pixel point to obtain each distribution area in the real-time enhancement image; A region processing mechanism connected to the content identification mechanism, used to detect multiple information parameters of each distribution region to obtain multiple information parameters of each distribution region, wherein the multiple information parameters of each distribution region are color component values ​​corresponding to each pixel point in the partitioned region, and the color component value corresponding to each pixel point includes a red and green component value, a black and white component value, and a yellow and blue component value corresponding to the pixel point; A directional prediction mechanism, connected to the regional processing mechanism, for inputting a plurality of information parameters of each distribution area, the clarity of the real-time enhanced image, the number of edge pixels, and the algorithm identifier corresponding to the current filtering algorithm into the AI ​​prediction model and running the AI ​​prediction model to obtain the maximum noise amplitude of the picture content obtained after the current filtering algorithm is executed on the real-time enhanced image output by the AI ​​prediction model, wherein the AI ​​prediction model is a feedforward deep network after completing each training; The amplitude comparison mechanism is connected to the directional prediction mechanism and is used to output the filtering algorithm corresponding to the minimum value of each maximum noise amplitude output by multiple filtering algorithms through the AI ​​prediction model as a reference filtering algorithm.

2. The image data analysis system based on algorithm analysis as claimed in claim 1, characterized in that: Various filtering algorithms include statistical sorting filtering algorithm, maximum filtering algorithm, minimum filtering algorithm, midpoint filtering algorithm, alpha mean filtering algorithm, adaptive filtering algorithm, bilateral filtering algorithm, guided filtering algorithm and box filtering algorithm.

3. The image data analysis system based on algorithm analysis as claimed in claim 2, characterized in that: The system further comprises: a temperature detection device, arranged near the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and connected to the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism respectively; Among them, the temperature detection device is arranged near the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and is respectively connected to the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism, including: the temperature detection device is used to respectively realize on-site measurement of the current temperature of the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism.

4. The image data analysis system based on algorithm analysis as claimed in claim 2, characterized in that: The system further comprises: a voltage detection device, arranged near the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and connected to the smoothing device, the data sharpening device, the real-time enhancement device and the content identification mechanism respectively; Among them, the voltage detection device is arranged near the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism and is respectively connected to the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism, including: the voltage detection device is used to respectively realize on-site measurement of the current voltage of the smoothing processing device, the data sharpening device, the real-time enhancement device and the content identification mechanism.

5. The image data analysis system based on algorithm analysis according to any one of claims 2 to 4, characterized in that: A SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

6. The image data analysis system based on algorithm analysis as claimed in claim 5, characterized in that: Using a SOC chip to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively includes: performing nearest neighbor interpolation processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

7. The image data analysis system based on algorithm analysis as claimed in claim 5, characterized in that: Using a SOC chip to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively includes: performing bilinear interpolation processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

8. The image data analysis system based on algorithm analysis as claimed in claim 5, characterized in that: The SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively, including: performing bicubic interpolation processing of a 16×16 pixel neighborhood on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.

9. The image data analysis system based on algorithm analysis as claimed in claim 5, characterized in that: The SOC chip is used to perform image data processing on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively, including: performing Lanczos interpolation processing of a 4×4 pixel neighborhood on the output data of the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism to obtain the output processing data corresponding to the smoothing device, the data sharpening device, the real-time enhancement device and the content recognition mechanism respectively.