TWI609352B - Method of color identification for paper-based analytical devices - Google Patents

Description

Color identification method of paper detecting device

The present invention relates to a color identification method for a paper type detecting device, and more particularly to a method for identifying the color of a detection area of a paper type detecting device by image processing.

Microfluidic Paper-Based Analytical Devices (μPADs) can perform preliminary screening for specific diseases. When the operator takes a test solution from the patient, the test droplet can be placed on the microfluidic paper. In the detection area of the wafer detecting device, the detection area is changed in color due to a chemical reaction. At this time, the operator can judge the degree of the patient's disease according to the color change of the detection area.

After the operator performs the above-mentioned paper detecting device, the operator may directly observe the color of the detecting area with the naked eye. However, different people may have different interpretations of the same color, and the color of the detecting area cannot be the same. When the judgment standard is identified, the detection result presented by the color change will have no reference value.

In order to avoid adverse effects caused by the different criteria, it is also known to detect the color of the detection zone of the paper detecting device through a detecting device. However, the detection area of the paper detecting device may generate noise points (such as wax spots or other defects) during manufacture, and when the detection area reacts with the liquid to be tested to generate a chemical reaction, the color change at the noise point will be Different from the color change of the rest of the detection area, when the detecting instrument calculates an average color value from all the color points in the detection area, the average color value will be affected by the color at the noise point and is distorted. , which in turn causes the color judgment result in the detection area to be out of alignment.

In view of this, the present invention provides a color identification method for a paper type detecting device to solve the problem of poor color recognition of a conventional paper detecting device.

The object of the present invention is to provide a color identification method for a paper type detecting device, and the color identification method of the paper detecting device can improve the color identification correctness.

In order to achieve the foregoing object, the technical means used in the present invention is as follows: a color identification method for a paper detecting device comprises the following steps: an image processing component having an image capturing device and a processor, the steps comprising: An image capturing step, the image capturing device captures a detection area of a paper detecting device to generate an image data, and transmits the image data to the processor; an image pre-processing step by the processor Performing a gray scale conversion on the pixels of the image data to obtain gray scale values of the plurality of pixel points, and setting a threshold value, and determining, by the plurality of pixel points, which pixel points have a gray scale value smaller than The threshold value is further regarded as a plurality of operation points for each pixel point satisfying the above judgment result; and a color judging step, the processor performs a color quantization process on the plurality of operation points to perform the plurality of operations The point is restored to the original color represented by the image data, and the color values of the plurality of operation points are obtained, and the plurality of operation points are grouped according to the color value by a split algorithm. a color region, and outputting a judgment color, and the judgment color is a color represented by the color region having the most operation point; after performing the color determination step, performing another comparison step, wherein the comparison step is performed by The processor reads a plurality of sample colors of a database, and compares the determined color and the plurality of sample colors by a comparison formula, and then uses the sample color having the smallest difference value from the determined color as one The color is compared to the result, and the comparison result is output; wherein the comparison formula is: Wherein, d for the difference value; red color of the sample values for q _r; q _g of sample color values for green; blue values for color samples of the q _b; p for _{R & lt} Analyzing the red color value; p _{g is} the green value of the judgment color; p _{b is} the blue value of the judgment color.

Wherein, in the color determining step, the splitting algorithm is a median cutting algorithm.

Accordingly, the color identification method of the paper type detecting device of the present invention can accurately distinguish the main representative color of the detecting flow path of the paper detecting device, and has the effect of improving the correctness of color recognition.

〔this invention〕

1‧‧‧Image capture device

2‧‧‧ Processor

3‧‧‧Database

4‧‧‧Loading station

5‧‧‧paper detection device

S1‧‧‧Image capture steps

S2‧‧‧Image pre-processing steps

S3‧‧‧Color judgment steps

S4‧‧‧ alignment steps

Fig. 1 is a view showing an apparatus for implementing a color identification method of a paper type detecting device of the present invention.

Fig. 2 is a flow chart showing the steps of the color identification method of the paper type detecting device of the present invention.

The above and other objects, features, and advantages of the present invention will become more apparent from the aspects of the appended claims. It is a reference implementation device for the color identification method of the paper detection device of the present invention. The implementation device can be an image processing component, and the image processing component includes an image capture device 1 and a processor 2, and the processor The 2 series is electrically connected to a database 3.

The image capturing device 1 can be any imaging component and can be used to capture a paper detecting device 5 on a loading platform 4 and generate an image data.

The processor 2 can be a computer or any computing processor, and can execute a software or a program for performing operations and the like. In this embodiment, the processor 2 is electrically connected to the image capturing device 1 to receive the image data, and can perform related image processing on the image data.

The database 3 is used to store a pair of data, which can be a plurality of sample colors. In this embodiment, the database 3 is electrically connected to the processor 2, so that the processor 2 reads the plurality of sample colors in the database 3, and enables the processor 2 to perform subsequent ratios. And the comparison data stored in the database 3 may be the image data captured by the image capturing device 1 or the image data processed by the processor 2 The plurality of sample colors, and the above data can be transferred to the database 3 through the processor 2 for use by the processor 2 as a reference for subsequent comparison and judgment.

Referring to FIG. 2, the color identification method of the paper type detecting device of the present invention comprises an image capturing step S1, an image pre-processing step S2 and a color determining step S3.

In the image capturing step S1, the image capturing device 1 captures a detection area of the paper detecting device 5 to generate an image data, and transmits the image data to a processor. The color of the paper detecting device 5 should preferably be separated from the color of the loading platform 4 to prevent the image data captured by the image capturing device 1 from simultaneously including the paper detecting device 5 and the carrying platform. At 4 o'clock, the paper detecting device 5 and the carrying table 4 are confused due to the similar colors.

In the image pre-processing step S2, the processor 2 performs a gray-scale conversion on a plurality of pixel points of the image data to obtain gray scale values of the plurality of pixel points, and sets a threshold value, and the plurality of pixels are set. In the pixel, it is determined which pixel points have a grayscale value smaller than the threshold, and each pixel corresponding to the above judgment result is collectively regarded as a plurality of operation points.

More specifically, the paper detecting device 5 usually has the detecting area and a non-detecting area (for example, a substrate of the paper detecting device 5 or a separation wax line surrounding the detecting area) when the detecting droplet After the detection zone of the paper detecting device 5, the detection zone may undergo a color change due to a chemical reaction (for example, darkening from white to purple), and the color of the detection zone may be different from the color of the non-detection zone; or If the detection zone produces noise (such as wax spots or other defects) during the manufacturing process, the color change at the noise point (eg, still maintaining white) will be different from the color change of the rest of the detection zone. Therefore, when the processor 2 performs gray scale conversion on the pixels of the image data, the color of the detection zone of the normal reaction can be clearly distinguished from the color of the noise or the non-detection zone (for example, at a lower level). The grayscale value represents a dark purple pixel, and the white pixel point is represented by a higher grayscale value. At this time, the processor 2 can set the grayscale value of the plurality of pixels with The threshold is compared, and only the pixels whose grayscale value is less than the threshold are regarded as a plurality of operation points, thereby removing unnecessary non-detection regions or noises in the image data, and only Subsequent color comparisons are performed on the pixels of the detection zone of the normal reaction, which has the effect of improving the correctness of the subsequent color determination. The threshold value can be freely set by the user according to the color distribution of the paper detecting device 5, and can be set between 0 and 255, which is well known to those skilled in the art, and is not limited or described herein. .

In the color determining step S3, the processor 2 performs a color quantization process on the plurality of operation points to obtain color values of the plurality of operation points, and then groups the plurality of operation points by a split algorithm according to the color value. It is a plurality of color zones, and the color represented by the color zone having the most operation point is used as a judgment color, and the judgment color is output by the processor 2.

More specifically, since the plurality of operation points generated in the image pre-processing step S2 are obtained by gray scale conversion, in order to present the original colors of the plurality of operation points, in the color determination step S3, The color quantization process first restores the plurality of operation points to the original color represented by the image data, and then obtains the color values of the plurality of operation points according to the original colors of the plurality of operation points, and the color value can be obtained from the three primary colors. (RGB) representation.

Moreover, the split algorithm can be any algorithm for grouping, and is not limited herein. In this embodiment, a Median-Cut Algorithm is used. The grouping method of the median cutting algorithm is a conventional technique, and is not described in detail here. It is only briefly described as follows: when the processor 2 receives the image data and obtains the plurality of operation points, the number can be The operation points are defined in a pixel space, wherein the three axes of the pixel space are respectively R, G, and B of the three primary colors, and each of the operation points has a coordinate value of R, G, and B. The median cutting algorithm system calculates the distribution of the plurality of operation points in the three axes of R, G, and B, and determines the axis of the plurality of R, G, and B axes. The distribution is the widest, and the most widely distributed axis is used as the grouping reference axis to group the plurality of operation points, and the median points of the distribution points of the plurality of operation points are used as the grouping boundary to calculate the number of operations. The points are grouped into two color areas. Next, judge each color zone The distribution of the plurality of operation points on the grouping reference axis, and the color regions having the widest distribution of the plurality of operation points are further grouped, and the median points of the distribution points of the plurality of operation points in the color region are used as the grouping boundary. The color is divided into two color regions, and the above procedure is repeatedly executed until the number of the color regions reaches a predetermined number. When the number of color regions of the plurality of operation points has reached the preset number, the image may have The color represented by the color region of the operation point is at most the judgment color.

Since the color of the detection area of the paper detecting device 5 is usually the same color system, after the median cutting algorithm, the color distortion caused by noise or other factors in the detection area can be effectively filtered. The color analysis area closest to the true color of the color of the detection area of the paper detecting device 5 can be used as the color analysis result. For example, when the detection droplet is in the detection area of the paper detecting device 5, the detection area may change color due to a chemical reaction (for example, darkening from white to purple), when the detection liquid diffusion effect is not good, It is possible to present different shades of color in the detection zone (for example, the center of the detection zone is dark purple, and the edge of the detection zone is light purple or other colors). By performing the median cutting algorithm, deep purple and non-deep can be performed. The plurality of operation points of purple are grouped into different color regions. Since most of the operation points in the detection area are still dark purple, the deep purple color region must have the largest number of operation points and has the most operations. The color represented by the color zone of the dot can also be closest to the true color of the detection zone. Thereby, the color judging step S3 can have the color represented by the color region of the operation point at most as the judging color, and has the effect of improving the correctness of the color recognition.

In addition, after performing the color determining step S3, the present invention may further perform a matching step S4, wherein the comparing step S4 reads the plurality of sample colors of the database 3 by the processor 2, and compares Comparing the formula with the determined color and the plurality of sample colors, the sample color having the smallest difference value from the determined color is used as a color comparison result, and the processor 2 outputs the comparison result.

In this embodiment, the comparison formula is as follows:

Wherein, d for the difference value; red color of the sample values for q _r; q _g of sample color values for green; blue values for color samples of the q _b; p for _{R & lt} Analyzing the red color value; p _{g is} the green value of the judgment color; p _{b is} the blue value of the judgment color.

More specifically, when the judgment color is generated, the operator can estimate the patient's degree of illness or other information based on the judgment color. Alternatively, the color change of the detection area of the paper detecting device 5 may also be set to a plurality of color levels, and the plurality of color levels respectively have a sample color, and when the processor 2 compares the color with the comparison formula And the plurality of sample colors, the sample color having the smallest difference from the judgment color can be found, and the sample color is used as the color comparison result. Thereby, the operator can accurately distinguish the color level of the detection area of the paper detecting device 5 according to the color comparison result, and has the effect of improving the correctness of the color comparison.

In summary, the color identification method of the paper type detecting device of the present invention can remove unnecessary non-detection areas or noise points in the image data through the image pre-processing step S2, and then pass the color determining step S3 to determine The main color of the detection area and the color level of the detection area of the paper detecting device 5 are accurately distinguished by the comparison step S4, which has the effect of improving the correctness of color recognition.

While the invention has been described in connection with the preferred embodiments described above, it is not intended to limit the scope of the invention. The technical scope of the invention is protected, and therefore the scope of the invention is defined by the scope of the appended claims.

S1‧‧‧Image capture steps

S2‧‧‧Image pre-processing steps

S3‧‧‧Color judgment steps

S4‧‧‧ alignment steps

Claims (2)

A method for color identification of a paper detecting device comprises the following steps: an image processing device having an image capturing device and a processor, the steps comprising: an image capturing step, wherein the image capturing device captures a paper Detecting area of one of the detection devices to generate an image data, and transmitting the image data to the processor; in an image pre-processing step, the processor performs a gray-scale conversion on a plurality of pixels of the image data, Obtaining a grayscale value of the plurality of pixel points, and setting a threshold value, and determining, by the plurality of pixel points, which grayscale values of the pixel points are smaller than the threshold value, and then each pixel point that meets the foregoing judgment result Commonly regarded as a plurality of operation points; and a color judging step, the processor performs a color quantization process on the plurality of operation points, to restore the plurality of operation points to the original color represented by the image data, and obtain the The color values of the plurality of operation points are further grouped into a plurality of color regions by a splitting algorithm according to the color value, and a judgment color is output, and the judgment color has the most a color represented by the color region of the operation point; after performing the color determination step, performing another comparison step, wherein the comparison step reads, by the processor, a plurality of sample colors of a database, and Comparing the formula to the determined color and the plurality of sample colors, and comparing the color of the sample having the smallest difference value with the determined color as a color comparison result, and outputting the comparison result; wherein the comparison The formula is: Wherein, d for the difference value; red color of the sample values for q _r; q _g of sample color values for green; blue values for color samples of the q _b; p for _{R & lt} Analyzing the red color value; p _{g is} the green value of the judgment color; p _{b is} the blue value of the judgment color. a color identification method for a paper type detecting device according to claim 1, wherein In the color judging step, the splitting algorithm is a median cutting algorithm.