JPH0546733A - Microorganism recognition device - Google Patents

Abstract

PURPOSE:To furthermore restrict an area to be applied to template scanning, to shorten an image processing time and to obtain high recognition accuracy by providing the microorganism recognizing device with a scanning area extracting circuit. CONSTITUTION:The brightness information of an input image is inputted by an image information inputting means 1 and sent to a smoothing processing circuit 2 to remove noise and the density gradation of the noise-removed brightness information is normalized and sent to a scanning area extracting circuit 8. The circuit 8 scans the brightness information and extracts a distance from a point on which brightness less than a 1st threshold set up lower than a background (water area) brightness level is changed to brightness more than a 2nd threshold set up higher than the background brightness level up to a point on which the brightness more than the 2nd threshold is changed to the brightness less than the 1st threshold on the same scanning line as an area to be scanned. An information conversion circuit 4 and a template comparator 6 execute processing such as template scanning and coincidence calculation only in the area extracted by the circuit 8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a microorganism recognition apparatus for automatically identifying a predetermined microorganism to be recognized from among luminance information of a sample in which a plurality of kinds of microorganisms and background hybrids are mixed. ..

[0002]

2. Description of the Related Art FIG. 4 shows, for example, "Journal of the Institute of Electronics, Information and Communication Engineers D" Vol. J71-D, No. 2 (1988 2
It is a block diagram which shows the conventional microorganism recognition apparatus shown by the paper "Pattern recognition system of a microscope image of microorganisms" published on pages 371 to 378 of Mon.

In the figure, reference numeral 1 is an input image obtained by enlarging a sample in which a plurality of types of microorganisms and background hybrids are mixed and enhancing the contour of the microorganism to be recognized and optically enhancing the internal brightness. Image information input means for obtaining the luminance information of the.

Reference numeral 2 denotes a smoothing processing circuit which removes noise of the luminance information input from the image information inputting means 1 and normalizes the grayscale level thereof, and 3 removes noise and normalizes the grayscale level. It is a mask processing circuit that masks the water area that is the background from the obtained luminance information.

Reference numeral 4 is an information conversion circuit for converting the brightness information into a direction value indicating a direction in which the brightness change with respect to a neighboring pixel is maximum, and 5 is a template for identifying a microorganism to be recognized.

Reference numeral 6 denotes a template comparison circuit which scans the template for a region not masked by the mask processing circuit 3 and obtains the degree of coincidence of the direction values from the information conversion circuit 4. Reference numeral 7 denotes the coincidence from the template comparison circuit 6. It is an identification circuit that determines the presence or absence of a microorganism to be identified based on the degree and the brightness information.

Next, the operation will be described. First, in the image information input means 1, an input image in which a sample in which a plurality of types of microorganisms and background hybrids are mixed is enlarged, and the edges of the microorganism to be recognized are emphasized and optically emphasized so that the internal brightness is increased. Luminance information of is input.

The brightness information of the input image is sent to the smoothing processing circuit 2 to remove spike noise, and the gray scale of the noise is normalized and sent to the mask processing circuit 3.
The mask processing circuit 3 divides the transmitted brightness information into a water area which is a background and an area other than the area including flocs, microorganisms, etc., and masks the water area to perform the information conversion circuit 4
Output to.

The information conversion circuit 4 determines, for each pixel in the area of the received luminance information that is not masked,
Find the maximum direction of change in brightness. The direction (gradient direction) in which the change in the brightness is maximized is obtained by the following equation 1 using, for example, a 3 × 3 Sobel filter.

[0010]

[Equation 1]

This φ (i, j) is the direction angle, and * means that 9 pixels in the 3 × 3 area are weighted according to the values of the elements of the matrix Δx ₁ or Δy ₁ and the sum is obtained. There is. Therefore, for example, the first expression of the above-mentioned expression 1 becomes the following expression 2.

[0012]

[Equation 2]

Note that k ₃ in the matrix Δx ₁ or Δy ₁ is usually "2".

The information conversion circuit 4 quantizes the gradient direction angle φ (i, j) thus obtained in n steps,
It is sent to the template comparison circuit 6 as a direction value. In this case, the facing directions are regarded as the same value, and therefore, the direction value is a value from 1 to n / 2.

When the template comparison circuit 6 receives the information converted into the direction value, it sequentially scans each pixel to be scanned with the center of the template 5 being found, the degree of coincidence of the template 5 is obtained, and the identification circuit 7 is provided. Send to.

The discrimination circuit 7 makes a final determination as to whether or not the microorganism is a discrimination target by utilizing the fact that the brightness is high at the point where the degree of coincidence is higher than a certain threshold value.

FIG. 5 is an explanatory diagram showing a discriminant function for such a determination, in which the horizontal axis represents the degree of coincidence of the template 5 and the vertical axis represents the brightness of the target point. In the figure, the shaded area is the microorganism determination area, and the points entering this area are determined to be the microorganisms to be identified.

Here, in the mask processing by the mask processing circuit 3, a lot of processing time is required for the calculation for obtaining the degree of coincidence of the template 5 by the information conversion circuit 4 and the template comparison circuit 6, so that a background is required. This is executed in order to reduce the processing time by masking the water region of (1) and narrowing down the template scanning region, and is performed by, for example, the region merging method.

This area merging method pays attention to the fact that the gradation change is gentle in the water area, and the absolute value of the gradation difference is set in the points adjacent to the point of interest. An area is created by merging points equal to or less than the threshold value, and an area larger than a predetermined area (for example, 50 pixels) and larger than the average density (luminance) is determined as a water area.

As a result, the input image of the sample is divided into a water region occupying a large area and a large number of small regions due to flocs, microorganisms, etc., and the water region is masked and excluded from the scanning region.

[0021]

Since the conventional microorganism recognition apparatus is configured as described above, template scanning, which requires a long processing time, is performed on all areas other than the masked area which can be clearly identified as a water area. It is necessary to calculate the degree of coincidence, and the mask treatment does not work well for the water region part having a shape similar to the microorganism surrounded by the curved flocs, and the final determination of the microorganism is shown in FIG. When the discrimination function is used, there is a problem that a water region having a shape similar to the microorganism to be recognized and having relatively high brightness may be mistaken for a microorganism.

The present invention has been made to solve the above-mentioned problems, and a microorganism capable of expanding the mask area for template scanning to areas other than water areas such as flocs to further limit the area where template scanning is performed. The purpose is to obtain a recognition device.

[0023]

In the microorganism recognizing apparatus according to the present invention, instead of a mask processing circuit for masking a water region, a first luminance level lower than a background luminance level is set on the same scanning line. From the point of changing from the luminance below the threshold value to the luminance above the second threshold value that is set higher than the background luminance level, the luminance from the second threshold value or more to the first threshold value or less Up to the point where
A scanning area extraction circuit for extracting as a scanning target area is arranged.

[0024]

The scanning area extraction circuit according to the present invention utilizes the fact that the brightness of the outline of the microorganism to be recognized is smaller than that of the background (water area) and the brightness of the inside of the microorganism is higher than that of the background. A first threshold lower than the level,
The second threshold value higher than the background luminance level is set to perform scanning of luminance information, and the luminance on the same scanning line is changed from the luminance equal to or lower than the first threshold value to the luminance equal to or higher than the second threshold value. The region from the point where the brightness is equal to or higher than the second threshold value to the point where the brightness is equal to or lower than the first threshold value is extracted as a scan target region, thereby further limiting the region where the template scan is performed. Realize a microorganism recognition device that can perform.

[0025]

【Example】

Example 1. An embodiment of the present invention will be described below with reference to the drawings. In FIG. 1, 1 is an image information input means, 2 is a smoothing processing circuit, 4 is an information conversion circuit, 5 is a template, 6 is a template comparison circuit, and 7 is an identification circuit. Since they are the same as or equivalent to those, detailed description thereof will be omitted.

Numeral 8 is arranged in place of the conventional mask processing circuit 3, scans the luminance information, and on the same scanning line,
From the point where the brightness lower than the first threshold value set lower than the background brightness level changes to the brightness higher than the second threshold value set higher than the background brightness level, the brightness higher than the second threshold value. It is a scanning area extraction circuit that extracts as a target area for scanning from a point to a point where the luminance changes to a first threshold value or less.

Next, the operation will be described. Since the processing operations other than the template scanning area extraction are the same as those in the conventional case, the description will be focused on the operation of the scanning area extraction circuit 8 here.

FIG. 2 shows a schematic diagram of an example of an input image input from the image information input means 1. In the figure, 11 is a floc-like microorganism (flock part), 12 is a hanging insect (green-haired bolticera) as a microorganism to be recognized, and 13
Is the background water area.

Further, FIG. 3 is an explanatory diagram showing a change in luminance when scanning on the line AA shown in FIG. Water area 1
The luminance level shows a high value S _w during scanning of No. 3, and decreases to a low value S _F when moving to the flock section 11.

The worm 12, which is a microorganism to be recognized, is
The brightness level of the contour line portion is as low as that of the flock portion 11, and an area having a brightness level higher than that of the water area 13 exists therein.

The scanning region extracting circuit 8, sets the threshold S ₁ luminance level S _w is lower than the first in the water area 13, the luminance level S second threshold higher than _w S ₂ in the water area Then, the brightness information of the input image processed by the smoothing processing circuit 2 is scanned.

Then, from the point (point X in FIG. 3) at which the luminance on the scanning line changes from the luminance below the _first threshold value S _{1 to} the luminance above the _second threshold value S ₂ , a second mark is displayed. Threshold S
_{A section} from a luminance of ₂ or more to a point (point Y in FIG. 3) at which the luminance changes to the first threshold S ₁ or less is extracted.

The scanning area extraction circuit 8 executes the above-mentioned processing over the entire area of the luminance information of the input image received from the smoothing processing circuit 2, and the set area obtained by both X and Y points is used as a template scanning area. Extract.

The information conversion circuit 4 and the template comparison circuit 6 execute a process that requires a great deal of processing time, such as template scanning and coincidence calculation, only for the area extracted by the scanning area extraction circuit 8.

[0035]

As described above, according to the present invention, the mask processing circuit is replaced by the scanning area extracting circuit to perform scanning of the luminance information, and the first luminance is set lower than the background luminance level on the same scanning line. From the point of changing from the luminance below the threshold value to the luminance above the second threshold value that is set higher than the background luminance level, the luminance from the second threshold value or more to the first threshold value or less Since it is configured to extract the area up to the point that changes to as the scan target area, it is possible to limit the area that executes processing such as template scanning and coincidence calculation that require a lot of processing time. Not only can the image processing time be shortened, but the water area that is surrounded by flocs and has a shape similar to the microbes, and the area inside the flocs that has a lower brightness than the flocs that is similar to the microbes can be excluded from the scan area. Therefore, false positives is reduced during the final determination by the recognition function in the identification circuit, the effect of microbial recognition device is obtained that can perform microbial recognition with high accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram showing a microorganism recognition device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing an example of an input image.

FIG. 3 is an explanatory diagram showing a change in luminance when scanning on the line AA shown in FIG.

FIG. 4 is a block diagram showing a conventional microorganism recognition device.

FIG. 5 is an explanatory diagram showing a discriminant function for making a final determination as to whether or not it is a microorganism.

[Explanation of symbols]

 1 Image Information Input Means 2 Smoothing Processing Circuit 4 Information Conversion Circuit 5 Template 6 Template Comparison Circuit 7 Discrimination Circuit 8 Scan Area Extraction Circuit

Claims (1)

[Claims] 1. Image information for obtaining luminance information of an input image obtained by enlarging a sample in which a plurality of types of microorganisms and background hybrids are mixed and optically enhancing the contour enhancement and internal luminance of a recognition-target microorganism. Input means, a smoothing processing circuit for normalizing the gray scale of the brightness information input from the image information input means, and a first threshold set to be lower than the background brightness level on the same scanning line of the brightness information. From the point of changing from the brightness below the value to the brightness above the second threshold set higher than the background brightness level, the brightness above the second threshold and below the first threshold Scan area extraction circuit that extracts as a target area for scanning up to a point that changes to, an information conversion circuit that converts the brightness information into a direction value indicating a direction in which the brightness change with respect to a neighboring pixel is maximum, and the recognition versus A template for identifying phantom microorganisms, a template comparison circuit that scans the template for the target area extracted by the scanning area extraction circuit, and obtains the degree of coincidence of the direction values from the information conversion circuit, and the degree of coincidence. And a discriminating circuit for discriminating the presence or absence of a microorganism to be discriminated based on the luminance information.