JP6981618B2 - Color pattern discrimination probe, color pattern discrimination device - Google Patents

Description

The present invention relates to a color pattern discriminating device capable of visually discriminating a color pattern on a target surface, and a color pattern discriminating probe capable of discriminating a color pattern that is particularly difficult to visually discriminate with high accuracy. Regarding the color pattern discriminating device.

In fields that handle products that are closely related to the user's life, such as architecture, furniture, interiors, clothing, and printing, color patterns (general term for visual features such as colors, patterns, reflective gloss, unevenness, and texture, which will be described later) are products. Often has a significant impact on the value of. If you mistakenly use materials with different colors and patterns, or combine parts with different colors and patterns, the finished product will disappoint the user's expectations and its value will be greatly reduced, and in some cases it will be necessary to remake it. In these fields, various variations of materials and parts are prepared according to the tastes of users, and many of them are difficult to distinguish at first glance. Therefore, at the sites where such products are manufactured, installed, and sold, it is a major issue to prevent the mix-up of materials and parts having different colors and patterns. With the method of visually checking for mix-ups, it is inevitable that mix-ups will occur with a certain probability.

As a method for that, a method of adding a sign indicating the color pattern to a material or a part has been known. Specifically, a method of printing a barcode or attaching a wireless IC tag to a portion invisible to the user is known (see Patent Document 1).

As another method, there is also known a method of imaging a material or a part using a camera and performing image processing to discriminate the color pattern (see Patent Document 2).

Japanese Unexamined Patent Publication No. 2008-79899 Japanese Unexamined Patent Publication No. 2014-202700

In the invention described in Patent Document 1, there is a problem that the sign is lost when the material is cut, or the sign is hidden when the part is incorporated into the product, making it impossible to distinguish.

Further, the invention described in Patent Document 2 is premised on the fact that the configuration for photographing is exposed to the outside and is used in an optically stable environment such as a production line in a factory. Therefore, when this invention is used in a place where the amount of light and the direction of insertion are likely to change, such as an outdoor or indoor construction / construction site, or a sales site in a store, the discrimination performance is determined by a change in the optical environment. There was a problem that it was difficult to correctly distinguish the color pattern. To solve this problem, there were approaches such as using special equipment and researching difficult algorithms, but this created the problem that the equipment tended to be large and expensive.

The present invention has been made in view of the above problems, and a color pattern discriminating device capable of discriminating color patterns of materials and parts that are difficult to discriminate visually with high accuracy and satisfying the following requirements is provided. The challenge is to provide it.
1. 1. Even if the material is cut or the parts are incorporated into the product, the color pattern can be discriminated.
2. 2. High discrimination accuracy can be maintained even in the field where the optical environment is not stable.
3. 3. It can be provided at a low price without depending on special hardware or complicated image processing method.

In order to solve such a problem, the invention according to claim 1 is a color pattern discrimination probe used for discriminating the color pattern on the surface of the object, and is arranged so as to irradiate the surface of the object with light. illumination means which is disposed to face the surface of the object, a photographing means for photographing the surface of the object illuminated with light by the illumination means, and before Symbol illuminating means and said photographing means A shielding means that optically shields from the outside and a main body portion that integrates the photographing means, the lighting means, and the shielding means are provided, and the shielding means is located between the lighting means and the object. The illuminating means holds a plurality of illuminating patterns as a pattern of the brightness of light irradiating the surface of the object. , by a predetermined control signal input receiving by said plurality of illumination patterns that can toggle between lighting control means varying the intensity of the light at Rukoto of the illumination pattern is changed, the imaging unit, the It is characterized by having a function of photographing the surface of the object irradiated with light emitted by the lighting means and having the lighting pattern switched, and acquiring image data for use in discriminating the color pattern.

Invention according to claim 2, in addition to the configuration of claim 1, wherein at least a portion of the front Stories body portion, a user of the colored pattern discrimination probes are grippable formation, the shielding means It is characterized in that at least a part thereof is formed so as to be able to adhere to the surface of the object by an elastically deformable material.

The invention according to claim 3 comprises the configuration according to claim 1 or 2 , wherein the lighting means includes an optical element having a function of receiving an input of a control signal to change optical characteristics. The optical element is characterized in that it is arranged so that the plurality of illumination patterns can be switched by changing the optical characteristics.

The invention according to claim 4 is the color pattern discrimination probe according to any one of claims 1 to 3 , wherein the lighting means has different illuminances of light shining on the surface of the object. It is characterized in that it has the above illumination pattern, and the change in the illuminance due to the difference in the illumination pattern is configured to be lower than the brightness resolution of the photographing means.

In the invention according to claim 5 , in addition to the configuration according to any one of claims 1 to 4 , the lighting means irradiates the surface of the object with light from an oblique direction, and the photographing means It is characterized in that image data including a shadow generated by irradiating the unevenness of the surface of the object with light emitted by the lighting means is acquired.

The invention described in claim 6 is the color patterns discriminating device, a color pattern discriminating probe according to any one of claims 1 to 5, in the illumination unit, to switch the plurality of the illumination pattern said lighting control means for transmitting control signals, said imaging means and imaging control means for controlling the operation of acquiring image data by photographing the surface of the object, image data obtained by the photographing of the photographing means The feature amount acquisition means for acquiring a plurality of feature amounts from the plurality of image data acquired by the photographing means while switching the plurality of illumination patterns, and the feature amount acquisition means. Further, the feature is provided with a category discriminating means for discriminating the category to which the surface of the object belongs by performing statistical processing on the feature amount.

Invention according to claim 7, in addition to the structure according to claim 6, comprising an index calculating means for calculating an index by the image data from the predetermined method of the imaging means, the same of the illumination pattern, wherein A plurality of image data are acquired by shooting at the same shooting position on the surface of the object at different times, and the index calculated from the plurality of image data by the index calculation means deviates from a predetermined standard. It is characterized by providing an index determination means that makes discrimination impossible in some cases.

The invention according to claim 8 is, in addition to the configuration according to claim 6 or 7 , a probe driving means for pressing the color pattern discrimination probe against the surface of the object, and a plurality of the inventions on the surface of a single object. The imaging means has a drive control means for driving the probe driving means by a predetermined control signal and pressing the color pattern discriminating probe against the corresponding imaging position, and the photographing means has the same. A plurality of the image data are acquired for each of the imaging positions to which the color pattern discrimination probe is pressed, and the feature quantity acquisition means obtains a plurality of feature quantities from the plurality of image data acquired at the plurality of imaging positions. It is characterized by acquiring.

According to the first aspect of the present invention, image data can be acquired from the surface of an object for each illumination pattern while switching between a plurality of illumination patterns. Therefore, more information can be acquired than when the illumination pattern is not switched, so that the color pattern on the surface of the object can be discriminated with higher accuracy. Further, by providing the shielding means, the optical system including the lighting means, the photographing means, and the surface of the object can be shielded from the external environment where the light state is unstable, so that the object can be shielded with higher accuracy. The color pattern on the surface can be discriminated.

According to the first aspect of the present invention, the color pattern discrimination probe can be formed in a small size by integrating the photographing means, the lighting means, and the shielding means. That is, the photographing means, the lighting means, and the shielding means can be firmly fixed to the inner surface of the main body, and the positional relationship between them and the way the light travels can be defined. It is possible to form a color pattern discrimination probe having a function and performance. Since the shielding means has an optically formed opening, the opening of the small color pattern discrimination probe can be pressed against the surface of the object to discriminate the color pattern, and the handling can be performed. It will be easier.

According to the second aspect of the present invention, a user who discriminates a color pattern grips and moves a part of the main body of the color pattern discriminating probe, or a part of a shielding means is placed on the surface of an object. It can be used by pressing it, elastically deforming it, and making it adhere to it, which makes it easier to handle the color pattern discrimination probe.

According to the third to fifth aspects of the invention, since a large amount of information can be acquired by switching the illumination pattern, it is possible to discriminate the color pattern of the object with higher accuracy.

According to the sixth aspect of the present invention, it is possible to utilize statistical processing capable of discriminating categories with higher accuracy as the number of feature quantities increases, and to discriminate the color pattern on the surface of an object with higher accuracy. ..

According to the invention of claim 7 , it is possible to eliminate the influence of disturbance such as the position of the color pattern discrimination probe being displaced during shooting and the disturbance light shining on the optical system, so that the accuracy is higher. Can be used to determine the color pattern on the surface of the object.

According to the eighth aspect of the present invention, image data can be acquired from the surface of an object at each of a plurality of imaging positions, and therefore more information can be acquired than when the imaging position is not changed. , The color pattern on the surface of the object can be discriminated with higher accuracy.

According to the invention described in each of the above claims, the following effects can be obtained.
1. 1. Color patterns are discriminated only by the appearance of the object without using signs such as barcodes and wireless IC tags. Therefore, even if the material is cut or the part is incorporated into the product, the color pattern of the original material or part can be discriminated.
2. 2. The light source, which is the light emitting part, and the camera, which is the light receiving part, are close to each other and optically shielded from the outside world. Therefore, high discrimination accuracy can be maintained even in the field where the optical environment is not stable.
3. 3. Since multiple images are taken while switching the illumination pattern in time, a large amount of optical information can be obtained even by using a general-purpose light source / camera, and the discrimination algorithm becomes easier accordingly. In addition, by fixing the positional relationship between the light source and the camera, adjustments on site are not required, and the discrimination algorithm becomes easy. Therefore, it can be provided at a low price without depending on special hardware or complicated image processing method.

It is a figure which shows the hardware composition of the color pattern discriminating apparatus 1A of Embodiment 1. FIG. It is a figure which shows the schematic structure of the color pattern discrimination probe 11 of Embodiment 1. It is a figure which shows the schematic structure of the 1st modification of the color pattern discrimination probe 11 of Embodiment 1. FIG. It is a figure which shows the schematic structure of the 2nd modification of the color pattern discrimination probe 11 of Embodiment 1. FIG. It is a figure which shows the schematic structure of the 3rd modification of the color pattern discrimination probe 11 of Embodiment 1. FIG. It is a functional block diagram which shows the functional outline of the color pattern discriminating probe 11 and the main body apparatus 12 constituting the color pattern discriminating apparatus 1A of Embodiment 1. FIG. It is a figure which shows typically the category discriminated by the color pattern discriminating apparatus 1A of Embodiment 1. FIG. It is a figure which shows the correspondence example of the illumination pattern and a feature amount in the color pattern discrimination apparatus 1A of Embodiment 1. FIG. It is a figure explaining the case which the change amount of the luminance of LED 22 by switching of an illumination pattern 201 is configured to be less than the brightness resolution peculiar to a camera 23. It is a figure explaining the case where the LED 22 is composed of a plurality of LEDs 22 having different peak wavelengths. It is a flowchart which shows the processing procedure of the color pattern discrimination apparatus 1A of Embodiment 1. It is a schematic diagram of the system configuration of the color pattern discrimination apparatus of Embodiment 2. It is a functional block diagram which shows the functional outline of the main body part 61 and the main body device 12 of the color pattern discriminating device 1B of Embodiment 2. FIG. It is a flowchart which shows the processing procedure of the color pattern discrimination apparatus 1B of Embodiment 2. It is a figure which shows the correspondence example of the probe position, the illumination pattern and a feature amount in the color pattern discrimination apparatus 1B of Embodiment 2. FIG.

[Embodiment 1 of the invention]
1 to 11 show Embodiment 1 of the present invention.

[Basic configuration]
FIG. 1 is a diagram showing a hardware configuration of the color pattern discriminating device 1A according to the first embodiment of the present invention. The color pattern discriminating device 1A is a device for discriminating the category to which the object 2 belongs based on the color pattern on the surface of the object 2. The specific contents of the object 2 and the category will be described later.

As shown in FIG. 1, the color pattern discrimination device 1A includes a color pattern discrimination probe 11, a main body device 12, and a connection cable 13.

The color pattern discrimination probe 11 is a small device used by a user (not shown) of the color pattern discrimination device 1A to contact the surface of the object 2 and acquire information about the color pattern from the surface of the object 2. Is. The specific configuration of the color pattern discrimination probe 11 will be described later. In this specification, the word "color pattern" is used as a general term for "features that can be visually observed". Specifically, "color pattern" includes things such as color, pattern, reflective luster, unevenness, and texture.

The main unit 12 has a configuration for transmitting / receiving various signals and data, storing / storing, and processing by a program. The main body device 12 has a function of taking in information about the color pattern acquired by the color pattern discrimination probe 11 and performing predetermined processing on the information to discriminate the category to which the object 2 belongs. The main body device 12 of the first embodiment is a personal computer, but the main body device 12 may be another device having the same function, for example, a mobile terminal, a device with a built-in microcomputer, or the like.

The connection cable 13 is, for example, a USB cable or a LAN cable, and is a flexible cable that connects the color pattern discrimination probe 11 and the main body device 12 so as to be communicable by wire to transmit and receive data and signals. Instead of the connection cable 13 for wired communication, wireless communication (not shown) such as Wi-Fi, wireless LAN, infrared communication, and Bluetooth (registered trademark) may be used.

[Structure of color pattern discrimination probe]
FIG. 2 is a diagram showing a schematic configuration of the color pattern discrimination probe 11 of the first embodiment.

As shown in the figure, the main body 10 of the color pattern discrimination probe 11 of the first embodiment has an LED 22 as a “lighting means” and a camera as a “shooting means” inside the housing 21 as a “housing”. 23 is incorporated and is integrally formed with a skirt 24 below the housing 21. The housing 21 and the skirt 24 are joined to each other so as not to allow light to pass through, and integrally function as a "shielding means". The housing 21 is formed in a size and shape that can be grasped by a user (not shown) with one hand.

The housing 21 is made of a highly rigid material such as resin or metal and is formed in a substantially cylindrical shape, and forms most of the outer shape of the main body 10. The upper surface portion 25 of the housing 21 is formed in a lid shape and is hermetically sealed so as not to allow light to pass through. The housing 21 is formed in a substantially concave shape with an opening downward (lower side in FIG. 2; the same shall apply hereinafter in the present specification), assuming that the housing 21 is used by being pressed against the object 2 from above. Specifically, an opening 26 is formed at the lower end of the housing 21, and an inner surface whose diameter is reduced from the opening 26 to the upper side (upper side in FIG. 2; the same shall apply hereinafter in the present specification) inside the housing 21. The portion 27 is formed. The lower side of the housing 21 below the opening 26 is optically opened via the skirt 24, and the inner surface portion 27 of the housing 21 is formed through the opening 26 and the inner surface and the close contact portion 29 (described later) of the skirt 24. It is open to the outside.

The above-mentioned "optically opened" state includes not only a physically opened state but also a physically closed state through which light is transmitted. In the first embodiment, both the opening 26 of the housing 21 and the close contact portion 29 (described later) of the skirt 24 are physically opened, but the present invention is not limited to this, and for example, the opening 26 and the close contact portion 29 (described later) are not limited to this. ), A part or all of them may be physically closed with a plate material of a transparent or translucent member (glass, resin, etc.) that transmits light.

In the color pattern discrimination probe 11 of the first embodiment, the LED 22 and the camera 23 are arranged close to each other inside the housing 21. That is, the LED 22 is arranged on the inner surface portion 27, and when the color pattern discrimination probe 11 is placed on the surface of the object 2, the upper end portion of the inner surface portion 27, which is a position facing the surface of the object 2, is located on the upper end portion. A camera 23 is arranged.

The LED 22 irradiates the surface of the object 2 with light. Although only two LEDs 22 are shown in FIG. 2, a plurality of other LEDs (not shown) may be provided along the concentric circles centered on the camera 23 of the inner surface portion 27. Further, one or more other LEDs (not shown) may be provided at positions other than the positions along the concentric circles. The detailed configuration of the LED 22 will be described in detail later in the section [Details of LED configuration].

The camera 23 is a digital camera such as a CMOS camera, and photographs the surface of the object 2 facing the color pattern discrimination probe 11. In the first embodiment, the camera 23 captures a still image, but may capture a moving image. The image data acquired by shooting is a two-dimensional array of spatial units called "pixels", and each pixel determines the brightness of the corresponding point in a gradation determined by the camera 23, for example, 256 of 0 to 255. It is held in gradation.

The skirt 24 is arranged in the opening 26 at the lower end of the housing 21, is made of a flexible elastically deformable material such as rubber, and is formed in a substantially cylindrical shape that expands downward. The skirt 24 is formed in a bellows shape provided with a step portion 28 in the horizontal direction. An opening is formed at the lower end of the skirt 24, and a close contact portion 29 that can flexibly contact and adhere to the object 2 is formed. When the color pattern discrimination probe 11 is pressed against the object 2, the skirt 24 is elastically deformed and adheres to the surface of the object 2, so that the highly rigid housing 21 does not come into direct contact with the object 2 and damage it. , Plays a role in preventing ambient light from entering the housing 21.

[Details of LED configuration]
The LED 22 in the first embodiment has a configuration for changing the illumination pattern in order to extract more information from the image data acquired by the camera 23.

For example, as shown in FIG. 2, each LED 22 is arranged so as to irradiate the surface of the object 2 with light in the oblique direction D1. Each LED 22 can be independently turned on / off by controlling the illumination pattern switching unit 41 (see FIG. 6). When the LED 22 to be turned on is switched, the direction of the light radiated to the object 2 changes, and if the surface of the object 2 has irregularities, the way the shadow is formed also changes. Therefore, when the color pattern discrimination probe 11 is used, the unevenness on the surface of the object 2 can be reflected in the image data.

As the LED 22, a color LED that emits white light in which light of three primary colors (red, green, and blue) is blended and whose light color changes depending on the blending ratio can be used. Further, some LEDs 22 that emit ultraviolet rays or infrared rays can also be used. Four or more LEDs 22 may be provided inside the housing 21 so that the peak wavelengths of the respective LEDs 22 are different. By doing so, even if the image sensor of the camera 23 can disperse only in the three primary colors of red, green, and blue, it is possible to acquire image data dispersed in four or more colors.

Each LED 22 may be configured so that the brightness at the time of light emission is changed by the control of the illumination pattern switching unit 41 (see FIG. 6). By changing the brightness of the LED 22, the quantization error caused by the finite brightness gradation of the pixels constituting the image data acquired from the camera 23 is eliminated, and a large amount of information exceeding the number of gradations can be obtained. Can be taken out.

Each LED 22 may be arranged so that the irradiation angle of light changes according to the control information of the illumination pattern switching unit 41 (see FIG. 6). For example, in response to the control information of the illumination pattern switching unit 41 (see FIG. 6), a means for changing the position and orientation of the reflecting mirror (not shown) and the lens (not shown) provided inside is provided, and FIG. The mode of irradiating the surface of the object 2 with light in the diagonal direction D1 and the mode of irradiating the light in the vertical direction D2 may be switched as shown in the above. Further, the irradiation angle may be changed in multiple steps between the oblique direction D1 and the vertical direction D2.

Although the LED 22 is used as the lighting means in the first embodiment, at least a part of the LED 22 may be replaced with another light emitting element such as an EL panel, a laser, or a discharge lamp.

[Modified configuration of color pattern discrimination probe]
In the color pattern discrimination probe 11 shown in FIG. 2, the shape of the inner surface portion 27 of the housing 21 and the arrangement position of various optical components including the LED 22 and the camera 23 on the inner surface portion 27 are objects of light emitted from the LED 22. Any other object may be used as long as it can irradiate the surface of the object 2 well and the camera 23 can appropriately photograph the surface of the object 2 and acquire image data of the required quality. ..

3 to 5 show a modified example of the color pattern discrimination probe 11 of the first embodiment. The color pattern discrimination probe 11 shown in these figures includes various optical elements that form a part of the “illumination means”. These color pattern discrimination probes 11 are representative of a wide variety of configurations in which it is possible to photograph the surface of an object 2 while switching a plurality of illumination patterns and acquire image data of required quality. It's just something like that. Therefore, the arrangement positions and the number of various optical components in FIGS. 3 to 5 are merely examples, and are not limited to those shown in the drawings.

FIG. 3 is a diagram showing a schematic configuration of a first modification of the color pattern discrimination probe 11 of the first embodiment. This configuration example is a color pattern discrimination probe 11 having a structure that guides the light emitted by a plurality of LEDs 22 to the surface of the object 2 by using the light guide material 30 that constitutes a part of the “illumination means”. The light guide material 30 is a member made of a transparent material such as acrylic resin, and guides light incident from one end to the other end. By appropriately arranging these and making it possible to switch the irradiation position and irradiation angle of light, it is possible to design lighting with a high degree of freedom. In this example, the light incident on the direction D3 close to perpendicular to the surface of the object 2, the light incident on the diagonal direction D4, and the light incident on the direction D5 close to the horizontal can be switched and used properly. Therefore, the gloss and reflection directivity of the surface of the object 2 can be reflected in the image data.

FIG. 4 is a diagram showing a schematic configuration of a second modification of the color pattern discrimination probe 11 of the first embodiment. This configuration example is a color pattern discrimination probe 11 having a structure that guides the light emitted by a plurality of LEDs 22 to the surface of the object 2 by using the inner surface portion 27 that constitutes a part of the “illumination means”. In the color pattern discrimination probe 11 shown in the figure, the inner surface portion 27 is formed of metal in a dome shape, and the inside thereof is coated with white paint that diffusely reflects light without absorbing it. The LED 22 is arranged so as to irradiate the installation portion 31 at the lower end of the housing 21 with light substantially upward. As shown in FIG. 4, the light emitted from the LED 22 is irradiated in the upward direction D6, reflected by the inner surface portion 27 at least once, and then becomes indirect illumination to be irradiated in the surface direction D7 of the object 2. Since shadows are less likely to be formed even if the surface of the object 2 is uneven, the influence of the gloss and reflection directivity of the surface of the object 2 should be eliminated, and the original color and pattern of the material should be faithfully reflected in the image data. Can be done. Further, if a plurality of types of LEDs 22 having different peak wavelengths are mixed, subtle differences in colors and patterns can be reflected in the image data with high accuracy. Although FIG. 4 shows a configuration in which only diffused light can be illuminated, an LED 22 (not shown) that directly irradiates the object 2 as shown in FIG. 2 is added to provide indirect illumination and direct illumination. It may be configured so that it can be switched.

FIG. 5 is a diagram showing a schematic configuration of a third modification of the color pattern discrimination probe 11 of the first embodiment. In this configuration example, the arrangement and configuration of the optical elements are arranged so that the optical elements constituting a part of the "illumination means" can switch the illumination pattern by changing the optical characteristics of the LED 22 and the like. I have. Specifically, the color pattern discrimination probe 11 of FIG. 5 is a color pattern discrimination probe 11 having a structure in which a plurality of liquid crystal filters 32 constituting a part of the “illumination means” are arranged in front of the LED 22. In the color pattern discrimination probe 11 shown in the figure, the LED 22 is arranged above the inner surface portion 27 of the housing 21 so as to irradiate light downward, and the inner surface portion 27 below the LED 22 is light transmissive. The liquid crystal filter 32 of the above is arranged, and an opening 33 is formed below the liquid crystal filter 32. When a voltage is applied to the liquid crystal filter 32 by an electric circuit (not shown), the light transmittance and the transmission direction can be changed depending on the voltage. In the modified example of FIG. 5, the light emitted by the LED 22 passes through the liquid crystal filter 32, is modulated, and then irradiates the surface of the object 2. Therefore, if the transmittance of the liquid crystal filter 32 is electrically changed, the irradiation direction of the light transmitted through the liquid crystal filter 32 changes in the directions D8 and D9, and as a result, the surface of the object 2 is irradiated. The direction of the light can be changed as the direction D10 and the direction D11. Not only transmissive elements such as liquid crystal filters and varifocal lenses, but also reflective elements such as galvano mirrors and digital mirror devices (DMDs) are used as optical elements that modulate light, which form part of the "illumination means". You may.

As described above, in the color pattern discrimination probe 11 shown in FIGS. 2 to 5, the shape and size of the housing 21 which is a part of the main body 10 are formed so as to be graspable by the user, but the main body is not limited to this. The other portion of the portion 10 may be formed in a shape and size that can be grasped by the user. For example, the skirt 24 constituting the main body portion 10 may be formed in a shape and size that can be grasped by the user. Further, the main body portion 10 may be configured to have a handle having a shape and a size that can be grasped by the user, which is projected from a part of the housing 21 or the skirt 24. In that case, the handle may be made of a transparent member that allows light to pass through, or may be made of an opaque member.

In any case, the configurations of the color pattern discrimination probe 11 shown in FIGS. 2 to 5 are all examples, and it goes without saying that there may be configurations based on other principles or configurations in which a plurality of principles are combined.

[Configuration of main unit]
FIG. 6 is a functional block diagram showing an outline of the functions of the color pattern discrimination probe 11 and the main body device 12 constituting the color pattern discrimination device 1A of the first embodiment. Based on the figure, the functional outline of the main body device 12 of the first embodiment will be described.

The main unit 12 includes a control means such as a CPU, a RAM or cache memory used by the CPU or the like as a work area or a temporary storage area, and an auxiliary storage medium such as a ROM, EEPROM, or a hard disk in which data or a program is recorded (all of which are shown in the figure). ). The color pattern discrimination program 40 used for the color pattern discrimination device 1A is set up in the recording unit 20 of the main body device 12 composed of a ROM, EEPROM, or the like (not shown). The color pattern discrimination program 40 has a function of controlling various hardware constituting the color pattern discrimination device 1A and discriminating the category to which the object 2 belongs. Further, the main body device 12 has an operation unit used as a user interface for instructing the color pattern discrimination device 1A to start the discrimination operation, and a display used for displaying the discrimination result by the color pattern discrimination device 1A. Have a department. The operation unit is composed of a keyboard 14 (see FIG. 1), a mouse (not shown), a touch panel (not shown), and the display unit is composed of a display 15 (see FIG. 1), a printer (not shown), and the like. To.

As shown in FIG. 6, the main unit 12 has a lighting pattern switching unit 41 as a “lighting control means” and a shooting as a “shooting control means” as functional means realized as a result of executing the color pattern discrimination program 40. It includes a control unit 42, an image acquisition unit 43, a feature amount extraction unit 44 as a "feature amount acquisition means", a category determination unit 45 as a "category determination means", and a general control unit 46. The illumination pattern switching unit 41, the image acquisition unit 43, the feature amount extraction unit 44, and the category determination unit 45 operate in conjunction with each other in response to the operation instruction 55 from the overall control unit 46.

The lighting pattern switching unit 41 sends a lighting control signal 51 to a lighting means or a driver thereof (a dedicated program for controlling hardware) to individually switch lighting / extinguishing of a plurality of LEDs 22 or to turn on each LED 22. It has a function of adjusting the brightness and driving an optical element that modulates the light emitted by the LED 22 to switch the illumination pattern applied to the surface of the object 2. The lighting pattern switching unit 41 has a table (not shown) that describes a combination of lighting control signals 51 corresponding to each lighting pattern, and when the lighting pattern is switched by an operation instruction 55 from the integrated control unit 46. , The lighting control signal 51 of the corresponding combination is sent to the LED 22.

The shooting control unit 42 controls the shooting operation performed by the camera 23. Specifically, a shooting signal for starting / ending the shooting operation is transmitted to the camera 23.

The image acquisition unit 43 constantly captures the image 52 from the camera 23 that is photographing the surface of the object 2 as a moving image, and when the operation instruction 55 is received from the integrated control unit 46, the still image from the image 52 at that time. It has a function of cutting out (frame) and creating an image 53.

The feature amount extraction unit 44 has a function of quantitatively evaluating the image 53 acquired by the image acquisition unit 43 and obtaining the feature amount vector 54 as a result. The feature amount vector 54 is a vector obtained by collecting a predetermined number of feature amounts, which are numerical values that characterize the image 53. Features that can be used for general purposes include pixel brightness / saturation / hue, movement correlation coefficient, spatial frequency, domain anisotropy, and roundness. When the values of these feature amounts are different for each part of the image 53, the average value and the dispersion value can be obtained for each of a plurality of preset areas of interest, and the feature amounts can be set separately for each.

The category discrimination unit 45 statistically compares the feature amount vector 54 obtained from the object 2 with the reference feature amount vector whose category is known, the category to which the object 2 is considered to belong, and the probability of belonging to the category. Has a function of discriminating the category to which the object 2 belongs by estimating. The "discrimination" here includes not only the process of finding the only category having a affiliation probability of 1, but also the process of finding a plurality of categories having a affiliation probability of less than 1 by allowing uncertainty. As a specific method for estimating the affiliation probability for each category, statistical algorithms such as a cluster analysis method, a distribution test method, and an MT method can be used.

The reference feature vector is a feature vector obtained from an object 2 whose category is known by prior learning and registered in advance in the category determination unit 45, and is obtained from an object 2 whose category is unknown. It is used for comparison with the feature vector 54. The work of registering the reference feature amount vector in the category determination unit 45 is called "teaching". Specifically, in this work, the color pattern discriminating device 1A is operated in a special mode, a feature quantity vector is obtained using an object 2 whose category is known, and a category to which the object 2 belongs is input. It is a thing. As a result, the correspondence between the feature amount vector and the category is recorded in a form that can be interpreted by the category discriminating unit 45.

The overall control unit 46 sends an operation instruction 55 to the lighting pattern switching unit 41, the shooting control unit 42, the image acquisition unit 43, the feature amount extraction unit 44, and the category determination unit 45 to operate them in a predetermined order to determine the color pattern. It has a function to link according to the purpose. It also has a function of directly executing some processing that cannot be entrusted to the illumination pattern switching unit 41, the image acquisition unit 43, the feature amount extraction unit 44, and the category determination unit 45.

[Discrimination target]
The object 2 whose color pattern should be discriminated by the color pattern discriminating device 1A of the first embodiment includes a house, a building, a building, a vehicle, a ship, an outer wall, an inner wall, a floor, a pillar, wood, a carpet, a wallpaper, and a door. , Sashes, desks, chairs, tanks, duvets, clothes, fabrics, wrapping paper, special paper, home appliances, electronic parts, stationery, toys, food, etc. The tangible object is not limited to a solid, and a part or all of it may be a liquid or a gas, or a part or all of it may be an intangible object, for example, the display state of a display. Further, the "color pattern" discriminated by the object 2 includes all "features that can be visually observed" such as color, pattern, reflected gloss, unevenness, and texture. The color may be chromatic or achromatic, and may be colored transparent or colorless and transparent. Further, the feature is not limited to the feature perceived by human eyes, but may be a feature usually perceived by tactile sensation such as unevenness and texture.

[category]
The "category" discriminated by the color pattern discriminating device 1A of the first embodiment is a group in which the object 2 is classified according to a predetermined rule. For example, categories such as wood type for wood grain, product number for wallpaper, and fiber type for fabric can be defined. Categories are determined by product and manufacturer-specific guidelines and are not necessarily quantitative or objective. Further, the visual features of the object 2 do not always correspond to one category. A plurality of visually indistinguishable objects 2 may belong to different categories.

FIG. 7 is a diagram schematically showing the categories discriminated by the color pattern discriminating device 1A of the first embodiment. The sample shown in the figure is the one displayed on the following site as an example.
Japan Wood Information Center "Wood net" homepage
http://www.jawic.or.jp/woods/sch.php
Fuchu Furniture Industry Cooperative "Wood Picture Book" Home Page
http://www.fuchu.or.jp/~kagu/mokuzai/mokuji.htm
FIG. 7 shows a case where the surface of the object 2 has a wood grain pattern, and in this case, the type of wood can be categorized. _{That is, the first category 100 1} where the type of tree is sugi, _{the second category 100 2} where the type of tree is red pine, _{the third category 100 3} where the type of tree is cypress, and the type of tree is teak. The fourth category 100 ₄ and the fifth category 100 ₂ whose tree type is agathis form different categories 100, and each category 100 has two or more samples. In the following, for the sake of simplicity, the first to fifth categories 100 ₁ , 100 ₂ , 100 ₃ , 100 ₄ , 100 ₅ are collectively referred to as category 100, unless otherwise necessary to distinguish them. (The same in this specification).

A plurality of visual features can be obtained from each sample. For example, obtain features such as lightness average value and variance value, saturation average value and dispersion value, hue average value and dispersion value, grain parallelism and linearity, light and dark ratio by wood grain, reflectance and diffusion rate. be able to. Further, the moving average and the moving correlation coefficient of these features, the correlation coefficient between two features, and the like can be regarded as different features. Therefore, it is possible to acquire a large number of features ranging from tens to thousands.

The category determination unit 45 determines which of the first category 100 ₁ to the fifth category 100 ₅ the object 2 photographed by the camera 23 belongs to. When one category 100 is selected and feature quantities are obtained from a large number of samples belonging to the category 100, each feature quantity follows a distribution determined for each category 100. For the sample of the object 2 to which the category 100 to which it belongs is unknown, the possibility that the sample belongs to each category 100 is statistically calculated by using the distribution of the feature amount obtained in advance for each category 100. It is possible to do. There are a plurality of categories 100 in which the distributions of the feature amounts almost overlap, and it may be difficult to distinguish the categories only from the feature amounts, but even in that case, the first category 100 ₁ to the fifth category 100 _{5 are} compared with each other. It is possible to calculate the likelihood that a sample will belong to each. It is also possible to extract a plurality of highly probable categories 100 and list them as candidates for the category 100 to which the sample belongs.

[Correspondence between lighting patterns and features]
FIG. 8 shows an example of correspondence between the illumination pattern and the feature amount in the color pattern discriminating device 1A of the first embodiment.

As shown in Table 200 of FIG. 8, in the first embodiment, the illumination pattern switching unit 41, illumination pattern ₂₀₁ 1 of N as the LED _22, 201 2, ......, switches the 201 _N. Note For ease of explanation in the following, unless otherwise there is a need for distinction, illumination pattern _{201 1,} 201 2, · · ·, 201 _N to as illumination pattern 201 collectively (the same in the present specification ).

The shooting control unit 42 causes the camera 23 to shoot for each lighting pattern 201, and acquires the image 53 one by one. Feature extracting unit 44, respectively the M from the image 53 of the feature amount C (e.g. feature quantity _C 11 from the image 53 taken by the original illumination pattern _{201 1, C 12, ......,} C 1M, illumination pattern The _{feature quantities C N1} , C _N2 , ..., C _NM ) are extracted from the image 53 taken under _{201 N.} Therefore, as a whole, N × M feature quantities C are extracted. In the conventional method in which the illumination pattern 201 is not switched, only M feature quantities C can be obtained, whereas the color pattern discriminating device 1A according to the present invention can extract a very large number of feature quantities C. Generally, in a statistical discrimination algorithm, the higher the number of features (dimension of the feature vector), the better the discrimination accuracy. Therefore, in the color pattern discrimination apparatus 1A according to the present invention, the accuracy of discrimination can be remarkably improved as compared with the conventional method.

[Lighting pattern switching and features]
In the color pattern discriminating device 1A of the first embodiment, as a method of obtaining a large number of features from the surface of the object 2, the camera 23 takes a plurality of photographs while the illumination pattern switching unit 41 switches the illumination pattern 201 of the LED 22. We are taking an approach to do.

Several other methods are known as methods for increasing the feature amount obtained from the surface of the object 2. How to use a camera with high optical resolution to increase the number of pixels that make up the acquired image 53, how to use a camera with a high dynamic range to increase the number of gradations of the pixels that make up the acquired image 53, color This is a method of increasing the number of primary colors of the pixels constituting the acquired image 53 by using a type or multispectral type camera. For example, replace a 300,000-pixel camera with a 2-megapixel camera, replace an 8-bit (256) gradation camera with a 16-bit (65,536) gradation camera, or replace a monochrome camera with a color one. If you do, you can get more features in one shot.

However, when the conventional approach is taken, it often costs more than the approach of the color pattern discriminating device 1A. The camera having a high optical resolution, the high dynamic range type camera, and the multispectral type camera are all considerably more expensive than the general-purpose camera 23 used in the color pattern discriminating device 1A of the first embodiment. Further, these special cameras are often larger than general-purpose cameras, and there remains a problem in terms of usability of the color pattern discriminating device 1A.

On the other hand, in the color pattern discriminating device 1A, although a general-purpose camera 23 is used, the same effect as using these special cameras can be obtained.

As one example, consider a case where the amount of change in the brightness of the LED 22 due to the switching of the illumination pattern 201 is configured to be lower than the brightness resolution inherent in the camera 23, and will be described below with reference to FIG. 9. In order to make the explanation here easy to understand, it is assumed that the camera 23 has a 9-step brightness scale of 1 to 9 and can distinguish only 9 kinds of brightness. It is assumed that the brightness of the surface of the object 2 photographed by the camera 23 is _{5.8 for the first object 2 a} and 6.2 for the second object 2 _b (FIG. 9A). ). In this case, the brightness difference between the first object 2 _a and the second object 2 _b of the surface does not appear in the image 53. This is because these surfaces all appear on the image 53 as pixels having a lightness of 6. However, the brightness of the LED 22 is changed from 90% to 110% in 5% increments (a to c in FIG. 9). Then, since the brightness of the surface of the object 2 changes in proportion to the brightness of the LED 22, the brightness of the pixels appearing in the image 53 is (5, 6, 6, 6, 6) in the _{first object 2 a.} in 2 of the object _{2 b} (6,6,6,7,7) next, it is understood to appear differences in five of the mean. That is, by changing the brightness of the LED 22 step by step, it becomes possible to distinguish a subtle difference in brightness on the surface of the object 2, which was originally indistinguishable. The camera 23 actually used by the color pattern discriminating device 1A is a general-purpose product with 8-bit (256) gradation, but by switching the illumination pattern 201 in this way, an expensive high dynamic range type camera 23 is used. The same effect as that of

As another example, consider a case where the LED 22 is composed of a plurality of LEDs 22 having different peak wavelengths, and will be described below with reference to FIG. 10. Again, for the sake of clarity, it is assumed that the camera 23 is a monochrome camera having a peak sensitivity at a wavelength of 500 nm. Further, it is assumed that the light reflected by the surface of the object 2 has the same wavelength distribution and the peak wavelengths are different, that is _{, 480 nm for the first object 2 a} and 520 nm for the second object 2 _b. .. In the description of FIG. 10, when it is necessary to distinguish, the first object 2 is described as an object 2 _a and the second object 2 is described as an _{object 2 b (first object 2 a} , Neither of the second objects _2b is shown).

When the light emitted from the LED 22 is white light, the difference in the brightness of the surfaces of the two objects 2 does not appear in the image 53 ((a) and (b) in FIG. 10A). This is because the peak wavelengths of 480 nm and 520 nm of the light reflected by each object 2 are symmetrically positioned with respect to 500 nm, which is the peak wavelength of the sensitivity of the camera 23.

But the LED 22 _a peak wavelength of 480 nm, the peak wavelength to irradiate switches the LED 22 _b is 520 nm, appear differences in the image 53 in the first object _{2 a} and the second object _{2 b} .. That is, when the former _{LED 22 a} is lit, the _{image 53 on the surface of the first object 2 a} ((a) in a of FIG. 10) is the image 53 on the surface of _{the second object 2 b (FIG. 10).} It becomes brighter than (b) of 10a. On the other hand, when the latter LED 22 is turned on, the _{image 53 on the surface of the second object 2 b} ((a) in c of FIG. 10) is the image 53 on the surface of _{the first object 2 a.} It becomes brighter than ((b) of c in FIG. 10). That is, the difference in color of the object 2 that cannot be distinguished by the combination of the monochrome camera 23 and the white light can be distinguished by using a plurality of LEDs 22 having different peak wavelengths.

The same idea as described above with respect to FIG. 10 applies even if the number of wavelengths having peak sensitivity is two or more. A general color camera has only three types of image sensors corresponding to the three primary colors (red, green, and blue), but by using four or more LEDs 22 with different peak wavelengths and switching the lighting pattern 201, You can get the same effect as using an expensive multispectral camera.

[Processing procedure]
FIG. 11 is a flowchart showing a processing procedure of the color pattern discriminating device 1A of the first embodiment. Hereinafter, the processing procedure of the color pattern discriminating device 1A of the first embodiment will be described with reference to the figure.

First, the user presses the close contact portion 29 of the color pattern discrimination probe 11 against the surface of the object 2 and starts the discrimination process using the keyboard 14 or the like of the main body device 12.

When the process is started, the illumination pattern switching unit 41 sends the illumination control signal 51 to the LED 22 and sets the illumination pattern of the LED 22 to the first illumination pattern 201 (step S1). In this state, the image acquisition unit 43 acquires the image 53 from the camera 23 (step S2).

Next, the illumination pattern switching unit 41 switches the LED 22 to the next illumination pattern 201 (“No” in step S3 to step S4). In this state, the image acquisition unit 43 acquires the image 53 from the camera 23 again (step S2). The above operation, the image acquisition unit 43, all of the illumination pattern ₂₀₁ 1 of LED _22, 201 2, ......, is repeated until obtaining an image 53 corresponding to the 201 _N.

All illumination pattern ₂₀₁ 1 of LED _22, 201 2, ......, an image acquisition unit 43 ( "YES" in step S3) of 201 to acquire an image 53 corresponding to _N are feature extraction unit 44 of the images 53 send. The feature amount extraction unit 44 obtains the feature amount vector 54 from those images 53 (step S5) and sends it to the category determination unit 45. The category determination unit 45 performs predetermined statistical processing to determine the category of the object 2 (step S6).

[Effect of Embodiment 1]
As described above, in the color pattern discriminating device 1A of the first embodiment, many features can be used to discriminate the category to which the object 2 belongs by performing the processes of steps S1 to S6. Higher accuracy can be obtained compared to the conventional method.

Specifically, in the first embodiment, the processes of steps S2 to S4 are repeated for the number of illumination patterns 201. Therefore, as schematically shown in the table 200 of FIG. 8, the illumination pattern switching unit 41 illumination pattern 201 of N as in, that the illumination pattern _{201 1,} 201 2, ..., when 201 _N are prepared, N different images 53 can be used to make one determination. That is, compared to the case where only one image 53 is used, a large amount of information can be used for determining the category. There is no need to use a special and expensive camera to achieve the above effects. Therefore, it is possible to suppress an increase in manufacturing cost and form a low-priced color pattern discrimination probe 11.

That is, in the first embodiment, image data can be acquired from the surface of the object 2 for each illumination pattern while switching a plurality of illumination patterns. Therefore, since more information can be acquired than when the illumination pattern is not switched, the color pattern on the surface of the object 2 can be discriminated with higher accuracy.

In the first embodiment, by providing the housing 21 and the skirt 24, the optical system including the surfaces of the LED 22, the camera 23, and the object 2 is shielded from the external environment where the light state is unstable. Therefore, the color pattern on the surface of the object 2 can be discriminated with higher accuracy.

In the first embodiment, the housing 21, the LED 22, and the camera 23 are integrated, so that the color pattern discrimination probe 11 can be formed in a small size. That is, the LED 22, the camera 23, and various optical components can be firmly fixed to the inner surface portion 27 of the housing 21, and the positional relationship between them and the way light travels can be defined. It is possible to form the color pattern discrimination probe 11 having sufficient functions and performance. Since the skirt 24 has an opening in which an opening is optically formed, the opening of the small color pattern discrimination probe 11 can be pressed against the surface of the object 2 to discriminate the color pattern. Easy to handle.

In the first embodiment, the user who discriminates the color pattern grips and moves a part of the housing 21 of the color pattern discriminating probe 11 or pushes a part of the skirt 24 to the surface of the object 2. It can be used by hitting it, elastically deforming it, and making it adhere to it, which makes it easier to handle the color pattern discrimination probe 11.

In the first embodiment, since a lot of information can be acquired by switching the illumination pattern 201, it is possible to discriminate the color pattern of the object with higher accuracy.

In the first embodiment, the statistical processing that can discriminate the category with higher accuracy as the number of feature quantities increases is utilized, and the color pattern on the surface of the object 2 can be discriminated with higher accuracy.

[Embodiment 2 of the invention]
12 to 15 show Embodiment 2 of the present invention.

[Basic configuration]
FIG. 12 is a schematic diagram of the system configuration of the color pattern discriminating device according to the second embodiment of the present invention. The color pattern discriminating device 1B as the "color pattern discriminating device" of the second embodiment is provided in the main body 61 in addition to the configuration of the color pattern discriminating device 1A described in the first embodiment of the present invention (see FIG. 1). The probe drive mechanism 62 is provided. That is, the color pattern discrimination device 1B of the second embodiment includes a main body portion 61, a main body device 12, and a connection cable 13, and the main body portion 61 includes a color pattern discrimination probe 11 and a probe drive mechanism 62. ..

In the second embodiment, the image acquisition unit 43 has a function as an "index calculation means" and a function as an "index determination means" in addition to the functions of the first embodiment. As an "index calculation means", the image acquisition unit 43 calculates an index from image data acquired by shooting by a predetermined method (specifically, will be described later). Further, as the "index determination means", the image acquisition unit 43 determines whether or not the category can be determined based on the indexes acquired from a plurality of image data acquired at different times (specifically, described later). ..

In the embodiment shown in FIG. 12, the color pattern discrimination probe 11 and the main body device 12 are separated from each other, but a small computer is mounted on the color pattern discrimination probe 11 and the color pattern discrimination program 40 (FIG. 12). It may be configured to operate a program having the same function as (see 6). By doing so, the color pattern discrimination probe 11 and the main body device 12 are integrated, and the main body device 12 and the connection cable 13 as hardware are not required.

[Structure of the main body]
As shown in FIG. 12, in the color pattern discriminating device 1B of the second embodiment, the main body 61 includes a probe drive mechanism 62. The probe drive mechanism 62 includes a transfer unit 63 and a connecting unit 64. Further, the probe drive mechanism 62 and the main body device 12 are connected by a cable (not shown) for transmitting a control signal for driving an actuator (not shown) or the like.

The transfer unit 63 includes a rail unit (not shown) that moves the color pattern discrimination probe 11 in the horizontal direction (left-right direction in FIG. 12), and an actuator (not shown) such as a servomotor or an air cylinder.

The connecting portion 64 connects the color pattern discrimination probe 11 and the transfer portion 63. For example, the connecting portion 64 has a structure in which a plurality of hollow columnar members have a nested structure, and is moved up and down by the power of an actuator (not shown). It is configured to expand and contract in the direction (vertical direction in FIG. 12).

The probe drive mechanism 62 receives an operation instruction from the main body device 12, moves the color pattern discrimination probe 11 along a predetermined path, and presses the color pattern discrimination probe 11 against a plurality of preset locations on the surface of the object 2. Do the action.

[Configuration of main unit]
FIG. 13 is a functional block diagram showing an outline of the functions of the main body 61 and the main body 12 of the color pattern discriminating device 1B according to the second embodiment. Based on the figure, the functional outline of the main body device 12 of the second embodiment will be described.

As shown in FIG. 13, in the color pattern discriminating device 1B of the second embodiment, the main body device 12 includes a probe driving unit 47 in addition to the configuration of the main body device 12 of the color pattern discriminating device 1A of the first embodiment. There is. Further, instead of the integrated control unit 46 of the first embodiment, the integrated control unit 46a is provided with a configuration as a "drive control means" by expanding the function in accordance with the addition of the probe drive unit 47.

The probe driving unit 47 has a function of sending a probe driving signal 56 to the probe driving mechanism 62 to operate the probe driving unit 47 and moving the color pattern discrimination probe 11 to a predetermined position on the surface of the object 2. The probe driving unit 47 also receives the operation instruction 55 of the integrated control unit 46a and operates in conjunction with the lighting pattern switching unit 41, the photographing control unit 42, the image acquisition unit 43, the feature amount extraction unit 44, and the category determination unit 45.

[Processing procedure]
FIG. 14 is a flowchart showing a processing procedure of the color pattern discriminating device 1B of the second embodiment. Hereinafter, the processing procedure of the second embodiment will be described with reference to the figure. The following control by the functional means of the main body device 12 is performed by the control of the integrated control unit 46a.

First, the user places the main body 61 on the object 2 and starts the discrimination process using the keyboard 14 of the main device 12 or the like.

When the process is started, the probe drive unit 47 sends a control signal to the probe drive mechanism 62 to drive an actuator (not shown) of the probe drive mechanism 62 to drive a preset initial position (eg, in FIG. 12). , The color pattern discrimination probe 11 is pressed against the position where the connecting portion 64 comes to the leftmost end of the transfer portion 63 (step S11). After that, the illumination pattern switching unit 41 transmits a illumination control signal to the LED 22 and sets the LED 22 to the preset initial irradiation state of light (step S12). In this state, the image acquisition unit 43 acquires the image 53 from the camera 23 (step S13).

Next, the illumination pattern switching unit 41 switches the LED 22 to the next illumination pattern (“No” in step S14 to step S15). In this state, the image acquisition unit 43 acquires the image 53 from the camera 23 again (step S13). The above operation is repeated until the image acquisition unit 43 acquires the image 53 corresponding to all the illumination patterns of the LED 22.

The image acquisition unit 43 that has acquired the image 53 corresponding to all the illumination patterns of the LED 22 (“Yes” in step S14) returns the control to the overall control unit 46a and causes the probe drive unit 47 to send the operation instruction 55. Upon receiving the operation instruction 55, the probe drive unit 47 drives an actuator (not shown) of the probe drive mechanism 62 and presses the color pattern discrimination probe 11 to the next predetermined position (“No” to step S16). S17). Then, the processes of steps S12 to S15 are performed again. The above operation is repeated until the probe driving unit 47 presses the color pattern discrimination probe 11 against all the predetermined positions and the processes of steps S12 to S15 are completed at each position.

When the probe driving unit 47 presses the color pattern discrimination probe 11 against all the predetermined positions and the processing of steps S12 to S15 is completed at each position (“Yes” in step S16), the probe driving unit 47 is preset. The color pattern discrimination probe 11 is moved to the first position (step S18). After that, the illumination pattern switching unit 41 switches the LED 22 to the first illumination pattern (step S19). In this state, the image acquisition unit 43 acquires the image 53 from the camera 23 (step S20).

The image acquisition unit 43 calculates a predetermined contrast index from the image 53 acquired last (step S20) and the image 53 acquired first (step S13), and compares the two. Here, the contrast index is an index for verifying the identity of the image 53, and may be a part or all of the feature amount vector 54.

When the difference between the two contrast indexes is within the predetermined reference range (“Yes” in step S21), the image acquisition unit 43 does not make a big difference between the first acquired image 53 and the last acquired image 53, that is, It was determined that there was no disturbance such as the position of the object 2 shifting during the discrimination or the disturbance light incident inside the color pattern discrimination probe 11, and all the images 53 acquired by the image acquisition unit 43 up to that point. The feature amount is extracted from (step S22). The category discrimination unit 45 performs statistical discrimination processing on all the extracted features, and calculates the category to which the object 2 is considered to belong and the probability that the object 2 belongs to the category (step). S23), the process ends.

On the other hand, when the difference between the two comparison indexes deviates from the predetermined reference range (“No” in step S21), the image acquisition unit 43 determines that a disturbance that hinders the discrimination has occurred, and all the acquired images are acquired. The image 53 is discarded (step S24). Then, the overall control unit 46a is requested to redo the series of processes starting from step S11.

[Effect of Embodiment 2]
In the second embodiment, the probe driving unit 47 moves the color pattern discrimination probe 11 along the surface of the object 2 and takes a picture with the camera 23, so that the color pattern discrimination probe 11 is placed in one place of the object 2. It is possible to acquire even more features than when the surface of the object 2 is photographed by fixing it to. That is, in addition to acquiring the repetitive image 53 with a plurality of illumination patterns at one place on the surface of the object 2 (step S14), the repetitive image 53 is also acquired at a plurality of places on the surface of the object 2 (step S16). Become. For example, when the position where the color pattern discrimination probe 11 is pressed is set at K points on the surface of the object 2, and N different illumination patterns are prepared for each position, K × for performing one discrimination. N different images 53 can be used. That is, much more information can be used for the discrimination process than when only one image 53 is used. The large amount of features acquired in this way helps to further improve the accuracy of discrimination.

FIG. 15 is a diagram showing an example of correspondence between the probe position, the illumination pattern, and the feature amount in the color pattern discriminating device 1B of the second embodiment. As schematically shown in Table 300 of FIG. 15, in the second embodiment, the positions (position 1, position 2, ..., Position K) at which the color pattern discrimination probe 11 is pressed against the K locations on the surface of the object 2. set, each lighting N as per position pattern 201 (first illumination pattern 201 _1, the second illumination pattern 201 _2, ..., illumination pattern 201 _N of the N) switches the. _{As shown in FIG. 15, images 53 are acquired for each combination 301 (combination 301 11} , combination 301 ₁₂ , ..., combination 301 _KN ) of each position and the illumination pattern 201, and M images are obtained from each image 53. Feature C (for example, from the image 53 taken under _{the combination 301 11} of the position 1 and the illumination pattern 1 _{to the feature C 111} , C ₁₁₂ , ..., C _{11M , the combination 301 KN} of the position K and the illumination pattern N). _{The feature quantities C KN1} , C _{KN 2} , ..., C _KNM ) are extracted from the original image 53. Therefore, as a whole, K × N × M feature quantities C are extracted. The color pattern discriminating device 1A of the first embodiment obtains only N × M feature quantities C, whereas the color pattern discriminating device 1B of the second embodiment extracts a larger number of feature quantities C. be able to. As described above, in general, in a statistical discrimination algorithm, the higher the number of features (dimension of the feature vector), the better the discrimination accuracy. Therefore, in the color pattern discriminating device 1B of the second embodiment, the accuracy of discrimination can be significantly improved as compared with the color pattern discriminating device 1A of the first embodiment.

Further, in the second embodiment, the color pattern discrimination probe 11 is moved along the surface of the object 2 by the actuator (not shown) of the probe drive mechanism 62 based on the control of the probe drive unit 47. By capturing the image 53 of the object 2 while moving it, the color pattern discrimination probe is taken as compared with the case where the user manually moves the color pattern discrimination probe 11 along the surface of the object 2 to capture the image 53. The movement amount and movement position of 11 can be accurately controlled. Therefore, it is possible to acquire the image 53 with high accuracy, which can increase the accuracy of category discrimination, and it is possible to further improve the accuracy of discrimination.

In the second embodiment, when the difference in the contrast index obtained from the plurality of images 53 taken at different times deviates from a predetermined reference range (“No” in step S21 to step S24), the discrimination is hindered. It can be determined that a disturbing disturbance has occurred. As a result, it is possible to prevent the image 53 in a poor shooting state from being used for discriminating the color pattern to reduce the discriminating accuracy, and to discriminate the color pattern of the object 2 with high accuracy. Although the same disturbance detection method can be applied to the color pattern discriminating device 1A of the first embodiment, it is more effective in the color pattern discriminating device 1B of the second embodiment.

It goes without saying that each of the above embodiments is an example of the present invention and does not mean that the present invention is limited to each of the above embodiments.

1A, 1B ... Color pattern discrimination device 2 _{, 2 a} , 2 _b ... Object 10 ... Main body 11 ... Color pattern discrimination probe 12 ... Main device 13 ... Connection cable 14 ... Keyboard 15 ... Display 20 ... Recording unit 21 ... Housing (shielding means)
22 _{, 22 a} , 22 _b ... LED (lighting means)
23 ... Camera (shooting means)
24 ... Skirt (shielding means)
25 ... Upper surface portion 26 ... Opening portion 27 ... Inner surface portion 28 ... Step portion 29 ... Adhesion portion 30 ... Light guide material 31 ... Installation portion 32 ... Liquid crystal filter 33 ... Opening 40 ... Color pattern discrimination program 41 ... Lighting pattern switching unit (lighting control means)
42 ... Shooting control unit (shooting control means)
43 ... Image acquisition unit (index calculation means, index determination means)
44 ... Feature amount extraction unit (feature amount extraction means)
45 ... Category discrimination unit (category discrimination means)
46 ... Integrated control unit 46a ... Integrated control unit (drive control means)
47 ... Probe drive unit 51 ... Lighting control signal 52 ... Video 53 ... Image (image data)
55 ... Operation instruction 56 ... Probe drive signal 61 ... (Consisting of a color pattern discrimination probe and a probe drive mechanism) Main body 62 ... Probe drive mechanism 63 ... Transfer unit 64 ... Connecting unit 100, 100 _{1 to} 100 ₅ ... Category 200, 300 ... Table 201, 201 _{1 to} 201 _N ... Lighting pattern 301, 301 _{11 to} 301 _KN ... Combination C (of position and lighting pattern) C, C _{11 to} C _NM , C _{111 to} C _KNM ... Feature quantity D1 ... Diagonal direction D2 ... Vertical direction D3 ... Near vertical direction D4 ... Diagonal direction D5 ... Near horizontal Direction D6 ... Upward D7 ... Surface direction D8, D9, D10, D11 ... Direction S1 to S6, S11 to S24 ... Step

Claims (8)

A color pattern discrimination probe used to discriminate the color pattern on the surface of an object.
Illumination means arranged to irradiate the surface of the object with light, and
An imaging means that is arranged facing the surface of the object and is irradiated with light by the lighting means, and an imaging means that photographs the surface of the object .
Before Symbol illumination means, and shielding means for shielding optically and the photographing means from the outside,
A main body unit that integrates the photographing means, the lighting means, and the shielding means is provided.
The shielding means has an opening located between the lighting means and the object and optically formed to face the surface of the object.
The illumination means, the illumination pattern as a luminance pattern of the light irradiated on the surface of the object a plurality of holding, the said light to switch between the plurality of illumination patterns by receiving an input of a predetermined control signal The lighting pattern is changed by a lighting control means capable of changing the brightness .
The photographing means has a function of photographing the surface of the object irradiated with the light emitted by the lighting means and having the illumination pattern switched, and acquiring image data for use in discriminating the color pattern. A color pattern discrimination probe characterized by. The color pattern discrimination probe according to claim 1.
At least a part of the main body is formed so that the user of the color pattern discrimination probe can grasp it.
A color pattern discrimination probe characterized in that at least a part of the shielding means is formed of an elastically deformable material so as to be in close contact with the surface of the object. The color pattern discrimination probe according to claim 1 or 2.
The lighting means includes an optical element having a function of changing optical characteristics in response to an input of a control signal.
The color pattern discrimination probe is characterized in that the optical element is arranged so that the plurality of illumination patterns can be switched by changing the optical characteristics. The color pattern discrimination probe according to any one of claims 1 to 3.
The lighting means has the plurality of lighting patterns in which the illuminance of the light applied to the surface of the object is different.
A color pattern discrimination probe characterized in that the change in illuminance due to a difference in the illumination pattern is configured to be lower than the brightness resolution of the photographing means. The color pattern discrimination probe according to any one of claims 1 to 4.
The lighting means irradiates the surface of the object with light from an oblique direction.
The photographing means is a color pattern discriminating probe characterized by acquiring image data including shadows generated by irradiating unevenness on the surface of the object with light emitted by the lighting means. The color pattern discrimination probe according to any one of claims 1 to 5.
The lighting control means for transmitting a control signal for switching the plurality of the lighting patterns to the lighting means, and the lighting means.
An imaging control means for controlling an operation of photographing the surface of the object and acquiring image data of the photographing means.
An image acquisition means for acquiring image data acquired by photographing the photographing means, and an image acquisition means.
A feature amount acquisition means for acquiring a plurality of feature amounts from the plurality of image data acquired by the photographing means while switching the plurality of illumination patterns.
It is provided with a category discriminating means for discriminating the category to which the surface of the object belongs by performing statistical processing on the feature quantity acquired by the feature quantity acquiring means.
A color pattern discriminating device characterized by this. The color pattern discriminating device according to claim 6.
An index calculation means for calculating an index from the above image data by a predetermined method is provided.
The imaging means acquires a plurality of image data by performing imaging at the same imaging position on the surface of the object with the same illumination pattern at different times.
A color pattern discriminating device comprising an index determining means that makes it impossible to discriminate when the index calculated from the plurality of image data by the index calculating means deviates from a predetermined standard. The color pattern discriminating device according to claim 6 or 7.
A probe driving means for pressing the color pattern discrimination probe against the surface of the object,
A plurality of imaging positions are set on the surface of a single object, the probe driving means is driven by a predetermined control signal, and the color pattern discrimination probe is pressed against the corresponding imaging position. Have,
The photographing means acquires a plurality of the image data for each shooting position to which the color pattern discrimination probe is pressed.
A color pattern discriminating device characterized in that the feature amount acquisition means acquires a plurality of feature amounts from a plurality of the image data acquired at the plurality of shooting positions.

Images