JP2003150944A - Image recognizing device, image recognition method, recording medium and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve image recognition efficiency by eliminating unnecessary light from reflected light from an object to reduce noise. SOLUTION: This image recognizing means includes a light source means 20 for applying light to an object, a filter means 40 for extracting only light having a specified light source wavelength from reflected light from the object and passing the same, an image pickup means 50 for converting the light passed through the filter means 40 to an image signal, and an image recognizing means 70 for performing recognizing processing for an image according to an image signal output by the image pickup means 50.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recognition apparatus, an image recognition method, a recording medium and a program capable of recognizing an image of a subject.

[0002]

2. Description of the Related Art An example of a conventional fingerprint matching device is disclosed in Japanese Patent Laid-Open No.
No. 0-11142. According to this conventional example, the light from the light source means is applied to the human finger as the subject through the prism. Light is diffused at the contact portion between the finger and the prism, and is totally reflected at the non-contact portion. Therefore, the reflected light from the prism is captured by the image capturing means as an image in which the concave portion of the fingerprint has a higher luminance value than the convex portion, and a fingerprint image of the finger can be obtained. The fingerprint image on the image pickup means is subjected to image processing such as signal level and contrast so that fingerprint detection can be optimally performed by the signal processing means at the subsequent stage. Then, the fingerprint discrimination means discriminates the person who has collated the fingerprint by detecting the characteristic points of the fingerprint. In addition, Japanese Patent Laid-Open No. 5-1
Japanese Patent No. 43763 discloses a technique relating to a barcode reader.

[0003]

However, the conventional fingerprint collation device or bar code reader as described above has the following problems. In these devices, the light source means is specially provided, but the contrast of the subject image is deteriorated by ambient light such as external natural light or illumination light. In addition, the reflectance or transmittance of the subject is large, so that part of the image signal is saturated and the signal component disappears. Conversely, the reflectance or transmittance of the subject is low, or the light source is deteriorated. An image recognition error may occur due to the reduction of the signal component due to the. As described above, the conventional bar code reader or the like has a drawback that the decoding result becomes NG or the probability that human fingerprint collation cannot be performed becomes large. In addition, the power consumption of the LED and the sensor is large and the battery life is short for portable use, which is not suitable.

The present invention has been made in view of the above problems, and an object of the present invention is to improve the conventional defects and improve the image recognition rate.

[0005]

In order to achieve such an object, an image recognition apparatus according to claim 1 is a light source means for irradiating a subject with light, and a specific light source of reflected light from the subject. Extraction means for extracting and passing only the wavelength,
The image pickup means is arranged in the latter stage of the extraction means and converts the light passing through the extraction means into an image signal, and the image recognition means for recognizing an image based on the image signal output by the image pickup means. It is characterized by

According to a second aspect of the present invention, there is provided the image recognition device according to the first aspect, wherein the light source means is a red light emitting diode.

According to a third aspect of the present invention, in the image recognition apparatus according to the first aspect, the extracting means is composed of a filter corresponding to each of the three primary colors.

Further, in the image recognition apparatus according to a fourth aspect, in the invention according to the first aspect, the image pickup means only includes an image signal of a color corresponding to the light source color of the light source means among the image signals of the three primary colors. Is output.

According to a fifth aspect of the present invention, in the image recognition apparatus according to the fourth aspect, the image pickup means has a plurality of pixels corresponding to each of the three primary colors and corresponds to each of the three primary colors. Among the plurality of pixels, by reading out the image signal of only the pixel corresponding to the color corresponding to the light source color of the light source means by thinning scanning, only the image signal of the color corresponding to the light source color of the light source means is output. Is characterized by.

An image recognition apparatus according to claim 6 is characterized in that, in the invention according to claim 1, the image pickup means sets an arbitrary exposure time.

The image recognition apparatus according to a seventh aspect of the present invention is the image recognition apparatus according to the first aspect, having a removing means for removing an unnecessary light component from one image signal by calculating a difference between the two image signals. The image pickup means outputs the image signals when the light source means does not emit light and when the light source emits light, and the removing means calculates the difference between the two signals output by the image pickup means. The calculation is performed, and an unnecessary light component is removed from the image signal when the light is irradiated based on the calculation result.

Further, in the image recognition method according to the eighth aspect, an irradiation step of irradiating the subject with light by the light source means, and only a specific light source wavelength of the reflected light from the subject is extracted by the extraction means and passed through. An extraction step to
An image capturing step, which is arranged in the latter stage of the extracting means, converts light passing through the extracting means into an image signal by the image capturing means, and image recognition processing by the image recognizing means based on the image signal output by the image capturing means. And an image recognition step to be performed.

According to a ninth aspect of the invention, there is provided the image recognition method according to the eighth aspect, wherein the light source means is a red light emitting diode.

The image recognition method according to claim 10 is
In the invention according to claim 8, the extraction means is constituted by a filter corresponding to each of the three primary colors.

The image recognition method according to claim 11 is:
In the invention according to claim 8, in the imaging step,
Among the three primary color image signals, only the image signal of a color corresponding to the light source color of the light source means is output from the image pickup means.

The image recognition method according to claim 12 is:
The invention according to claim 11, wherein in the imaging step, image signals of only the pixels corresponding to the color corresponding to the light source color of the light source means among the plurality of pixels corresponding to the three primary colors of the imaging means are provided. It is characterized in that only the image signal of the color corresponding to the light source color of the light source means is output from the image pickup means by reading out from the image pickup means by thinning scanning.

The image recognition method according to claim 13 is:
In the invention according to claim 8, in the imaging step,
It is characterized in that an arbitrary exposure time is set for the image pickup means.

The image recognition method according to claim 14 is:
In the invention according to claim 8, there is a removal step of removing an unnecessary light component from one image signal by calculating a difference between the two image signals, and in the imaging step,
Image signals when the light source means does not emit light and when it emits light are respectively output from the image pickup means,
In the removing step, a difference between the two signals output by the image pickup means is calculated, and an unnecessary light component is removed from an image signal when the light is emitted based on the calculation result.

A recording medium according to a fifteenth aspect is characterized by recording a program for causing a computer to function as each unit according to any one of the first to seventh aspects.

A recording medium according to a sixteenth aspect is characterized by recording a program for causing a computer to execute the processing steps of the image recognition method according to any one of the eighth to fourteenth aspects.

A program according to a seventeenth aspect causes a computer to function as each unit according to any one of the first to seventh aspects.

Further, the program according to claim 18 is
A computer is caused to execute the processing steps of the image recognition method according to any one of claims 8 to 14.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows the arrangement of an image recognition apparatus according to an embodiment of the present invention. In FIG. 1, 10 is a subject, 2
Reference numeral 0 is a specially provided light source means for illuminating a subject, 30 is an optical means for collecting reflected light from the subject and guiding it to an image pickup means, 4
Reference numeral 0 is a filter means for passing the light source wavelength of the light source means 20, 50 is an image pickup means for converting a reflected image of the subject into an image signal, and 60 is a fingerprint image signal or a barcode signal required from the image signal from the image pickup means. The image processing means 70 performs image processing so as to facilitate image recognition, 70 is an image recognition means for performing image recognition based on the image signal output from the signal processing means 60, and 80 is a display means such as a monitor for displaying a subject image.

FIG. 2 shows the spectral sensitivity characteristics of the main parts of the system. The image sensor used in the image pickup means 50 is shown in FIG.
Generally, it has the characteristic shown by the sensor of. About 40
It has a spectral sensitivity from 0 nm to about 900 nm. On the other hand, a red LED is generally used as the light source means.
This is because the red LED is cheap in cost and is a bright light source. Since the above-mentioned sensor and LED are used in a normal image recognition device, when natural light enters a subject, red LE
Components other than D become noise that reduces the contrast of the image. Since the energy of natural light has a wide spectral band with respect to the spectral energy of the LED, the unnecessary signal level due to this natural light cannot be ignored for the image signal.
In the present embodiment, disturbance light can be removed by providing the filter means 40 with a filter (R) having a spectral characteristic as shown in FIG. For this reason, almost only the image of the subject is captured, and the image recognition rate is greatly improved.

Although the filter means is provided between the image pickup means and the optical means in FIG. 1, this may be provided in the optical path before the image pickup sensor. Especially when it is used in a thin portable fingerprint collation device or the like, it may be provided on-chip on the image sensor.
When the image sensor is also used as a color camera, it may be a color image sensor. As the filters of the color camera, filters (R), filters (G), and filters (B) as shown in FIG. 2 are used. In this case, since the light source means 20 as an image recognition device is a red LED, it is better to use only the signal for the filter (R) as the signal from the image sensor.

FIG. 3 shows a schematic diagram of signal reading. Figure 3
In the case of using as a camera, signals of respective colors are sequentially read out as shown in (A). When used in the image recognition device, only the red signal is read as in (B). There is a CMOS sensor as an image sensor suitable for such reading. C
As shown in FIG. 4, the MOS sensor includes a pixel unit 110, a horizontal scanning circuit 120 capable of randomly scanning the signals of the pixel unit 110, a vertical scanning circuit 130, and a reading circuit 140.

In the image recognition apparatus having the color filter shown in FIG. 3, the pixels used for image recognition need to use only the R signal. When the luminance signal Y formed from the R, G, and B signals of the signal processing means 60 is used for image recognition, the signals G and B are determined as missing signals of the image signal of the subject, so an error in image recognition occurs (B. In order to read out only the red R signal (thinning-out scanning) such as (1), vertical scanning and horizontal scanning may be skipped for one row and one column. As a result, the number of read pixels can be reduced, and power consumption can be reduced accordingly.

FIG. 5 shows a timing chart in the second embodiment of the present invention. Generally, in the image recognition apparatus, in the case of a bar code reader used by many users, the object on which the bar code is printed or the paper has different reflectance. Also,
The reflectance also varies depending on dirt and deterioration over time. Further, there is deterioration of the light amount of the light source means. For this reason, under a fixed exposure condition, an appropriate signal level necessary for image recognition cannot be obtained, and the recognition rate deteriorates. In the present embodiment, the image sensor changes the exposure condition to take an image to improve the above-mentioned drawbacks.

If the exposure conditions are changed exponentially, an appropriate signal level can be obtained from at least one image even if the signal level changes as described above. In the embodiment of FIG. 5, the exposure ratio is 1: 1/2: 1/4. In FIG. 5, fv1, fv2 ... Are vertical drive frequencies of the image sensor, and v1, v2 ... Are horizontal pixel rows.

In the figure, (A) shows an example of driving when the exposure time is constant. Each pixel row is driven (C / R / T / R) every 1 fv. That is, the resetting (C) of the pixel amplifier, the noise transfer (R) of the pixel amplifier, the transfer of the photodiode to the pixel amplifier (T), and the transfer of the signal (R) from the pixel amplifier are repeated for each row pixel. By this repeated driving, the signal of the accumulation time TS1 is read from the image sensor.

(B) shows the above-mentioned exposure time by T for each fv.
In this example, s1: Ts2: Ts3 = 1: 1/2: 1/4. The exposure time is changed by clearing the pixel (T / C) during storage.

FIG. 6 shows a pixel configuration example of the image sensor and a timing chart. The photodiode corresponds to each color filter, and the pixel amplifier is shared by the two photodiodes. In the period C, the gate of the pixel amplifier is cleared, and the residual charges of the vertical signal line and the memory Cs (Cn) are reset before reading the noise of the pixel amplifier. Next, the noise of the pixel amplifier is transferred to the memory Cn during the R period, and then the signal of the photodiode is transferred from the pixel amplifier to the memory Cn.
s. The noise and the signal from the memory are transferred to a differential amplifier (not shown) during the next horizontal scanning period to remove the noise.

A third embodiment of the present invention is shown in FIG. 5- (C). In this embodiment, the influence of ambient light is further improved as compared with the first embodiment. The image sensor is LE
The accumulated signal when D is not lit is read during the fv1 period and LE
The signal at the time of lighting D is read during the fv3 period, the removing means in the subsequent stage calculates the difference between both signals, and L is calculated based on the result.
The unnecessary light component is removed from the signal when the ED is turned on. In this case, the lighting period of the LED requires 2fv period. To reduce the power consumption of the LED, if the gate portion of the pixel amplifier is used as a pixel memory, the LED lighting period of 1fv period can be shortened. FIG. 7 is a diagram when the subject of FIG. 1 is a finger and the image recognition device of the present embodiment is applied to a fingerprint collation device.

According to the above embodiment, when the specific light source means is used, the filter means 40 is provided in the image pickup system, so that the noise due to unnecessary light can be remarkably reduced. In addition, a signal suitable for image recognition can be obtained by performing image recognition from a plurality of images with different exposure conditions. In addition, when the LED is lit and when it is not lit, images are taken, and the noise of unnecessary light is reduced by the arithmetic processing. As a result, the recognition rate of the image recognition device can be improved. Furthermore, power consumption can be reduced by thinning scanning.

Further, an object of the present invention is to supply a storage medium having a program code of software for realizing the functions of the above-described embodiments to a system or apparatus, and to supply the computer (or CPU or M of the system or apparatus).
It is needless to say that it is also achieved by the PU) reading and executing the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a flexible disk, hard disk, optical disk, magneto-optical disk, CD-ROM, C
DR, magnetic tape, non-volatile memory card, ROM
Etc. can be used.

Further, by executing the program code read by the computer, not only the functions of the above-described embodiment are realized, but also the OS (operating system) operating on the computer based on the instruction of the program code. It is needless to say that this also includes the case where () etc. perform part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.

Further, after the program code read from the storage medium is written in the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, based on the instruction of the program code, Needless to say, this also includes the case where the CPU or the like included in the function expansion board or the function expansion unit performs some or all of the actual processing and the processing realizes the functions of the above-described embodiments.

[0041]

As is apparent from the above description, according to the present invention, since the unnecessary light component is not transmitted from the reflected light from the subject by the extracting means, the noise due to the unnecessary light is significantly reduced. Therefore, the image recognition rate can be improved.

[Brief description of drawings]

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

FIG. 2 is a diagram showing spectral sensitivity characteristics of main components of the image recognition device according to the embodiment of the present invention.

FIG. 3 is a diagram schematically showing a signal reading operation of the image recognition device according to the embodiment of the present invention.

FIG. 4 is a diagram showing an example of a CMOS sensor applicable to the image recognition device according to the embodiment of the present invention.

FIG. 5 is a timing diagram showing an operation of the image recognition device according to the embodiment of the present invention.

FIG. 6 is a diagram showing a pixel configuration example of an image sensor and an operation timing.

FIG. 7 is a diagram when the image recognition device according to the embodiment of the present invention is applied to a fingerprint matching device.

[Explanation of symbols]

10 subject 20 light source means 30 Optical means 40 Filter means 50 Imaging means 60 signal processing means 70 Image recognition means 80 display means 110 pixels 120 horizontal scanning circuit 130 vertical scanning circuit 140 read circuit

Claims (18)

[Claims] 1. A light source unit for irradiating a subject with light, an extracting unit for extracting and passing only a specific light source wavelength of the reflected light from the subject, and an extracting unit arranged after the extracting unit. An image recognition apparatus comprising: an image pickup unit that converts light that has passed through the unit into an image signal; and an image recognition unit that performs image recognition processing based on the image signal output by the image pickup unit. 2. The image recognition apparatus according to claim 1, wherein the light source means is a red light emitting diode. 3. The extracting means comprises a filter corresponding to each of the three primary colors.
The image recognition device described. 4. The image recognition apparatus according to claim 1, wherein the image pickup unit outputs only an image signal of a color corresponding to the light source color of the light source unit among the image signals of the three primary colors. 5. The image pickup means has a plurality of pixels corresponding to each of the three primary colors, and corresponds to a color corresponding to the light source color of the light source means among the plurality of pixels corresponding to each of the three primary colors. The image recognition apparatus according to claim 4, wherein only the image signal of the color corresponding to the light source color of the light source means is output by reading out the image signal of only the pixels to be thinned out. 6. The image recognition apparatus according to claim 1, wherein the image pickup unit sets an arbitrary exposure time. 7. A removing means for removing an unnecessary light component from one image signal by calculating a difference between the two image signals, wherein the image pickup means is provided when the light source means does not emit light. And output the image signal when irradiated with light,
The removing unit calculates a difference between the two signals output by the image pickup unit, and removes an unnecessary light component from an image signal when the light is irradiated based on the calculation result. The image recognition device according to item 1. 8. An irradiation step of irradiating a subject with light by a light source means, an extraction step of extracting only a specific light source wavelength of reflected light from the subject by an extracting means, and a post-stage of the extracting means. And an image recognizing step for performing image recognition processing by the image recognizing means on the basis of the image signal output by the image capturing means. An image recognition method comprising: 9. The image recognition method according to claim 8, wherein the light source means is a red light emitting diode. 10. The image recognition method according to claim 8, wherein the extraction means is composed of a filter corresponding to each of the three primary colors. 11. The image pickup device according to claim 8, wherein in the image pickup step, of the image signals of the three primary colors, only the image signal of a color corresponding to the light source color of the light source device is output from the image pickup device. Recognition method. 12. In the imaging step, an image signal of only a pixel corresponding to a color corresponding to a light source color of the light source means among a plurality of pixels corresponding to respective three primary colors of the imaging means is subjected to thinning scanning. 12. The image recognition method according to claim 11, wherein only the image signal of a color corresponding to the light source color of the light source means is output from the image pickup means by reading out from the image pickup means. 13. The image recognition method according to claim 8, wherein in the imaging step, an arbitrary exposure time is set for the imaging means. 14. A step of removing an unnecessary light component from one of the image signals by calculating a difference between the two image signals, wherein in the imaging step, when the light source means does not emit light. And an image signal at the time of irradiating light is output from the image pickup means, and in the removing step, the difference between the two signals output by the image pickup means is calculated, and the light is radiated based on the calculation result. 9. An unnecessary light component is removed from the image signal at that time.
Image recognition method described. 15. A computer-readable recording medium in which a program for causing a computer to function as each unit according to any one of claims 1 to 7 is recorded. 16. A computer-readable recording medium on which a program for causing a computer to execute the processing steps of the image recognition method according to claim 8 is recorded. 17. A program for causing a computer to function as each unit according to any one of claims 1 to 7. 18. A program for causing a computer to execute the processing steps of the image recognition method according to any one of claims 8 to 14.