Colorization conversion method for multi-source heterogeneous data and multi-channel data

Abstract

A method for color conversion of multi-source heterogeneous data and multi-channel data, the specific process of which is as follows: first, digital signals obtained by multiple sensors are input into an n×3 matrix to form standard data; second, the standard data is input into the matrix and mapped to R, G, and B data after matrix operation to obtain n RGB color images; third, the n RGB color images obtained in step 2 are converted into multi-channel data fusion images through matrix multiplication. The present invention solves the complexity of expressing multiple sensor data and the practical limitations of RGB representation, and achieves optimal accuracy when performing color conversion of multi-channel data.

Description

Technical Field

[0001] The present invention relates to the technical field related to machine vision and image processing, and in particular to a method for color conversion of multi-source heterogeneous data and multi-channel data. Background Art

[0002] In response to the progress of social implementation of the Society 5.0 initiative, an increasing number of sensors and devices are expected to be connected to the internet. As the number of sensors increases, the amount of data they generate will also increase. Therefore, improved data analysis methods are needed to process the data continuously generated at any given time with high efficiency and accuracy. When sensing information in a specific environment, several different types of data are typically acquired, which are often processed and analyzed in a multidimensional manner by computer science to obtain useful knowledge. Regardless of the analysis field, when multiple sensing operations are performed on a single analysis target and different types of data are acquired, in many cases each element will change over time. Considering the correlation effects of multiple factors rather than processing each factor individually, analytical performance should be improved. However, in this case, how to effectively utilize dimensional correlation for feature extraction is a subject of research by those skilled in the art. Summary of the Invention

[0003] The present invention provides a method for analyzing data by color attributes. This method allows us to consider the relationship between dimensions and implement a high-performance feature extraction method for multi-source heterogeneous data and multi-channel data color conversion.

[0004] The specific process of the color conversion method for multi-source heterogeneous data and multi-channel data is as follows:

[0005] First, digital signals obtained by various types of sensors are input into an n×3 matrix to form standard data;

[0006] Second, input the standard data into the matrix operation and map it to R, G, and B data to obtain n RGB color images;

[0007] 3. Convert the n RGB color images obtained in step 2 into a multi-channel data fusion image through matrix multiplication;

[0008] The n×3 matrix is:

[0009]

[0010] The matrix operation logic is as follows: after the start, the sensor data input of multiple different models is converted into an n×3 matrix, and the initial number of rows of the matrix is ​​specified to be i=1. When i is less than or equal to n, the acquired window unit is slid to the sensor data X1, X2, and X3 of the i-th row, and then the sensor data of different models are subjected to n×3 matrix operation with R, G, and B components to obtain an RGB image. When the number of RGB images is less than the number of sensor data, the acquired window unit is returned to slide to the sensor data X1, X2, and X3 of the i-th row, and then the sensor data of different models are subjected to n×3 matrix operation with R, G, and B components to obtain an RGB image. The operation is repeated until the number of images is greater than or equal to the number of data: i=i+1. After that, the RGB color image is converted into a multi-channel data fusion image abbreviated as X′1X′2X′3 through matrix multiplication.

[0011] The matrix multiplication formula is:

[0012]

[0013] In the research and development of the present invention, the problem that a single data displayed by different types of sensors does not have the multi-dimensional fusion capability of multiple data was first solved. Secondly, the two-dimensional limitation of the simple RGB image representation was solved, including the computational challenges it brings when representing the stimulation of color components. In the RGB color space, sensor data are all interpreted as points, and there will be data points with the same distance to the red on the left, and there will be no difference in the image display. In addition, the simple RGB image representation cannot correctly handle the correlation between the three primary colors, which are the components of all colors, because each color component is processed independently when training CNN. In order to overcome these problems, a new color space was developed, called the multi-channel data fusion image, abbreviated as X′1X′2X′3. It can make up for the shortcomings of the RGB color space because it is a color representation that reflects human vision and can handle the correlation between the R, G and B color components.

[0014] In step 1, by converting sensor data of multiple different models into color representations, the three dimensions of the data can be converted into an element called color while retaining the relationship between the dimensions, that is, the color representation of human vision, and completing the processing of the correlation between the R, G and B color components.

[0015] In the second step, the sensor data of various types are integrated into an N×3 matrix, and each data is traversed through matrix operations to finally obtain N RGB color images to ensure the accuracy of color data extraction.

[0016] The multi-channel data fusion image in step 3 is a linear combination of the R, G, and B color components. In step 3, the RGB representation can also cause problems, such as different actual colors even if the distance from each color to the reference color is the same. Therefore, another color space, also known as the multi-channel data fusion image, is used. This space represents each color component as a simple linear combination of R / G and B, making it easier to interpret the relationship between the three colors.

[0017] The window unit is a set of data values ​​of sensor data of different models within the same time period.

[0018] The various types of sensors include displacement sensors, force sensors, speed sensors, temperature sensors, flow sensors, and gas composition sensors.

[0019] The present invention solves the complexity of expressing multiple sensor data and the practical limitations of RGB representation, and achieves optimal accuracy when performing color conversion of multi-channel data. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be further described below with reference to the accompanying drawings:

[0021] Figure 1 This is a flow chart of the method for color conversion of multi-source heterogeneous data and multi-channel data according to the present invention;

[0022] Figure 2 A data situation view of an embodiment used in the present invention;

[0023] Figure 3 This is a schematic diagram of a window unit in the present invention;

[0024] Figure 4 This is a schematic diagram of the multi-channel data fusion image conversion principle of the present invention. Specific implementation plan

[0025] The following is combined with Figure 1-4 The specific implementation method further describes the technical solution of this patent in detail.

[0026] The specific process of the color conversion method for multi-source heterogeneous data and multi-channel data is as follows:

[0027] First, digital signals obtained by various types of sensors are input into an n×3 matrix to form standard data;

[0028] Second, input the standard data into the matrix operation and map it to R, G, and B data to obtain n RGB color images;

[0029] 3. Convert the n RGB color images obtained in step 2 into a multi-channel data fusion image through matrix multiplication;

[0030] The n×3 matrix is:

[0031]

[0032] The matrix operation logic is as follows: after the start, the sensor data input of various models is converted into an n×3 matrix. The initial number of rows of the matrix is ​​specified to be i=1. When i is less than or equal to n, the acquired window unit slides to the sensor data X1, X2, and X3 in the i-th row. Then, the sensor data of different models are subjected to n×3 matrix operations with the R, G, and B components to obtain an RGB image. When the number of RGB images is less than the number of sensor data, the acquired window unit is returned to slide to the sensor data X1, X2, and X3 in the i-th row. Then, the sensor data of different models are subjected to n×3 matrix operations with the R, G, and B components to obtain an RGB image. The operation is repeated until the number of images is greater than or equal to the number of data: i=i+1. After that, the RGB color image is converted into a multi-channel data fusion image through matrix multiplication. The multi-channel data fusion image is abbreviated as X′1X′2X′3.

[0033] The matrix multiplication formula is:

[0034]

[0035] In step one, by converting sensor data from multiple different models into color representations, the three dimensions of the data can be converted into an element called color while preserving the relationship between the dimensions.

[0036] In step 2, the sensor data of various models are integrated into an N×3 matrix, and each data is traversed through matrix operations to finally obtain N RGB color images.

[0037] In step 3, the multi-channel data fusion image is a linear combination of R, G, and B color components.

[0038] As a preferred structure, multiple different types of sensors include displacement sensors, force sensors, speed sensors, temperature sensors, flow sensors, and gas composition sensors. The window unit is a collection of data values ​​of different types of sensors within the same time period.

[0039] The data used in the specific embodiment of the present invention is as follows Figure 2 As shown, different sensor data X1, X2, and X3 are used to sense different data conditions at different times, such as standing data, sitting data, going up / down stairs data, cycling data, etc., and then Figure 3 As shown, the data is cut and slid window by window, and finally Figure 4As shown, the window-by-window cropped data is converted into R, G, and B data channels through an N×3 matrix to form N RGB color images, and the N RGB color images are converted into multi-channel data fusion images through matrix operation colors.

[0040] The above describes in detail the preferred embodiments of this patent, but this patent is not limited to the above embodiments. Various changes can be made within the knowledge of ordinary technicians in this field without departing from the purpose of this patent.

Claims

1. A method for color conversion of multi-source heterogeneous data and multi-channel data, characterized by: The specific process of the conversion method is as follows: First, digital signals obtained by various types of sensors are input into an n×3 matrix to form standard data; Second, input the standard data into the matrix operation and map it to R, G, and B data to obtain n RGB color images; 3. Convert the n RGB color images obtained in step 2 into a multi-channel data fusion image through matrix multiplication; The n×3 matrix is: The matrix operation logic is as follows: after the start, the sensor data input of multiple different models is converted into an n×3 matrix, and the initial number of rows of the matrix is ​​specified to be i=1. When i is less than or equal to n, the acquired window unit is slid to the sensor data X1, X2, and X3 of the i-th row, and then the sensor data of different models are subjected to n×3 matrix operation with R, G, and B components to obtain an RGB image. When the number of RGB images is less than the number of sensor data, the acquired window unit is returned to slide to the sensor data X1, X2, and X3 of the i-th row, and then the sensor data of different models are subjected to n×3 matrix operation with R, G, and B components to obtain an RGB image. The operation is repeated until the number of images is greater than or equal to the number of data: i=i+1. After that, the RGB color image is converted into a multi-channel data fusion image through matrix multiplication. The multi-channel data fusion image is abbreviated as X′1X′2X′3. The matrix multiplication formula is:

2. The method for color conversion of multi-source heterogeneous data and multi-channel data according to claim 1, characterized in that: In step 1, by converting sensor data of multiple different models into color representations, the three dimensions of the data can be converted into an element called color while preserving the relationship between the dimensions.

3. The method for color conversion of multi-source heterogeneous data and multi-channel data according to claim 1, characterized in that: In the second step, the sensor data of various types are integrated into an N×3 matrix, and each data is traversed through matrix operations to finally obtain N RGB color images.

4. The method for color conversion of multi-source heterogeneous data and multi-channel data according to claim 1, characterized in that: The multi-channel data fusion image in step three is a linear combination of R, G, and B color components.

5. The method for color conversion of multi-source heterogeneous data and multi-channel data according to claim 1, characterized in that: The window unit is a set of data values ​​of sensor data of different models within the same time period.

6. The method for color conversion of multi-source heterogeneous data and multi-channel data according to claim 1, characterized in that: The various types of sensors include displacement sensors, force sensors, speed sensors, temperature sensors, flow sensors, and gas composition sensors.

Images