JP2013045149A - Information code reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information coder reader which rarely fails to read information from an information code in an image.SOLUTION: Corner detection processing is performed to a picked-up image (S12), and candidates for corners of an information code are extracted on the basis of edge line vectors and positions from a plurality of detected corners (S14). Furthermore, estimated cell size is calculated for each of information code corner candidates (S16). Then, two to four candidates are selected as information code corners from the information code corner candidates on the basis of degrees of similarity of the estimated cell size, and the code outer shape of an information code D is estimated on the basis of the information code corners (S18).

Description

  The present invention relates to an information code reader that captures an information code by a camera and reads information from the information code, and more particularly to a technique for specifying the position of the information code in a captured image.

  There are many known information code readers that capture an information code with a camera and analyze the captured image to read information from the information code. When analyzing an image in an information code reader, in order to speed up the processing, the position of the information code in the captured image is specified, and a decoding process is performed on a part of the image including the position of the information code. There are many cases.

  As a method of specifying the position of the information code, a method of searching for a characteristic pattern included in the information code from the entire image is widely known. For example, when the information code is a QR code (registered trademark), since the QR code includes a cut-out symbol (finder pattern), first, processing for searching the cut-out symbol from the entire image is performed.

  Patent Document 1 discloses a technique for specifying the position of a two-dimensional code by creating an envelope surrounding a two-dimensional code image. Specifically, in Patent Document 1, after binarizing a two-dimensional code image and removing noise, a black cell existing at the outermost periphery is found in the binarized image, and a space between adjacent black cells is sequentially detected. By connecting, an envelope is created (paragraph 0016 of Patent Document 1).

2004-362053

  In Patent Document 1, in order to create an envelope, it is necessary to find a black cell that is present at the outermost periphery of the two-dimensional code image, but it is described how to determine the outermost periphery of the two-dimensional code image. Not. Actually, it is difficult to determine whether or not the black portion is a cell before the position of the two-dimensional code is specified.

  In addition, in the method of searching for a characteristic pattern such as a finder pattern in the QR code from the entire image, if the characteristic pattern cannot be found due to dirt or specular reflection, the position of the information code may be specified. Can not. Therefore, for example, the specular reflection is only a part of the finder pattern portion and the like, and even if the image is decodable, the finder pattern cannot be detected, resulting in a result that information cannot be read.

  The present invention has been made based on this situation, and an object of the present invention is to provide an information code reading apparatus in which information cannot be read from an information code in an image.

In order to achieve the object, the invention according to claim 1
(A) An information code reading device that includes a camera, images an information code by the camera, and reads information from the information code based on the captured image,
(B) feature point detecting means for detecting a feature point including at least a corner from an image photographed by the camera;
(C) Information code angle for extracting information code angle candidates that are candidates for the information code corners based on the edge line vector passing through the feature points and the positions of the feature points from the feature points detected by the feature point detection means Candidate extraction means;
(D) Corner candidate information calculation means for calculating an estimated cell size for each information code angle candidate based on a luminance change in a predetermined range around the extracted information code angle candidate;
(E) Based on the degree of similarity of the estimated cell size calculated by the corner candidate information calculation means, 2 to 4 candidates are selected as information code angles from the information code angle candidates, and based on the selected information code angles, An outer shape estimating means for estimating a quadrangular code outer shape as an outer shape of the information code;
(F) A code that determines an edge angle at a plurality of points inside the code outline estimated by the outline estimation means, and determines a code axis angle that is a reference angle of the information code based on the determined plurality of edge angles Shaft angle determining means;
(G) Calculate the black and white change number and the black and white change width inside the code outline estimated by the external shape estimation means and on lines orthogonal to each other, and based on the calculated black and white change number and black and white change width, Cell size determining means for determining the cell size;
(H) decoding processing means for decoding an information code based on the code outer shape estimated by the outer shape estimating means, the code axis angle determined by the code axis angle determining means, and the cell size determined by the cell size determining means; , Including.

  In the present invention, in order to specify the position of the information code, the information code angle candidate extracting means extracts information code angle candidates (information code angle candidates). In order to extract the information code angle candidates, feature points are detected from the image captured by the camera. Information code angle candidates are extracted based on the edge line vectors passing through the feature points and the positions of the feature points.

  The feature points that can be detected from the image include not only corners that define the outer shape of the information code, but also corners formed by cells in the information code and parts other than the corners. However, since the information code angle candidate extracting means extracts the information code angle candidates from the feature points based on the edge line vectors and the positions of the feature points, it is difficult to extract a portion other than the corners or corners in the information code. . Therefore, the information code angle candidates are narrowed down to a certain number.

  However, if there is another figure (character or the like) having a corner shape in addition to the information code in the imaging range, the corner on the figure may be extracted as an information code corner candidate. Therefore, the estimated cell size is further calculated for each information code angle candidate based on the luminance change around the information code angle candidate. Among the information code angle candidates, for the real information code angle, the luminance change around it reflects the cell size of the information code. On the other hand, for candidates that are extracted as information code corner candidates but are not actually information code corners (for example, the corners of the above-mentioned figure), the luminance change around them reflects the cell size of the information code. Because it is not, it is very different from the actual information code angle. Accordingly, the estimated cell size calculated based on the luminance change around the information code angle candidate is greatly different between the case of the information code angle and the case of not being the information code angle. In other words, the estimated cell sizes of information code angle candidates that are the angles that define the outer shape of the same information code are similar to each other. Therefore, the outer shape estimation means selects 2 to 4 candidates as information code angles from the information code angle candidates based on the degree of similarity of the estimated cell sizes. Based on this, a quadrangular code outer shape that is the outer shape of the information code is estimated. As described above, in the present invention, the code outline is estimated without detecting a feature pattern such as a finder pattern. Therefore, the code outline can be estimated even for an image in which the feature pattern cannot be detected. In addition, four information code angles are not necessarily used, and the code outer shape is estimated from two or three information code angles. As a result, the number of cases in which the outer shape of the code cannot be determined is smaller than in the past.

  Further, the code axis angle and the cell size are also determined without using the feature pattern. That is, the code axis angle is determined based on the determined edge angles at the plurality of points inside the code outer shape. The cell size is also determined based on the number of black and white changes and the black and white change width on mutually orthogonal lines inside the code outline. The decoding processing means performs decoding processing based on the code outline, code axis angle, and cell size determined in this way. Therefore, even if the image cannot detect the feature pattern, the decoding process can be performed, so that it is less likely that the information code cannot be read.

In the invention according to claim 2, the edge angle frequency distribution determined based on the edge angles determined for the plurality of edges by the code axis angle determination means is the edge angle frequency distribution in the one-dimensional code and the two-dimensional code. A code type determining means for determining whether the information code is a one-dimensional code or a two-dimensional code by determining which of the frequency distributions of the edge angle;
The decoding processing means determines which one of the one-dimensional code decoding algorithm and the two-dimensional code decoding algorithm to use first based on the determination result of the code type determining means.

  Since the one-dimensional code has only lines arranged in one direction in parallel, the frequency distribution of edge angles in the one-dimensional code concentrates the frequency in one angle range. On the other hand, since the two-dimensional code has cells arranged in two directions, the vertical direction and the horizontal direction, the frequency distribution of edge angles in the two-dimensional code has at least two angular ranges in which the frequencies are concentrated. . Therefore, the code type can be determined by using the frequency distribution of the edge angle. Based on the determined code type, it is determined which one of the decoding algorithm for the one-dimensional code and the decoding algorithm for the two-dimensional code is used first, so that the decoding can be performed at high speed.

In invention of Claim 3, it comprises the rough image creation means which creates the rough image which extracted pixel information from the image photoed with the camera at fixed pixel intervals,
The feature point detection means and the information code angle candidate extraction means detect the feature points using the coarse image created by the coarse image creation means, and extract the information code angle candidates.

  When feature points are detected using a coarse image in this way, unnecessary feature points such as small feature points inside the information code are difficult to detect, so there are few unnecessary candidates included in information code angle candidates. Become. As a result, the information code angle can be easily selected.

  In the third aspect of the present invention, the roughness of the coarse image can be determined as in the fourth and fifth aspects. According to a fourth aspect of the present invention, cell size storage means for storing the cell size at the time of reading the information code is provided, and the coarse image creating means is based on the cell size at the time of reading the previous information code stored in the cell size storage means. The roughness of the rough image to be created is determined. In Claim 5, the cell size input means for a user to input a cell size is provided, The said rough image production | generation means is the roughness of the coarse image produced based on the cell size input into the said cell size input means. To decide.

  In the invention according to claim 6, the outer shape estimating means has a range larger than the code outer shape estimated based on the information code angle selected from the information code angle candidates by the estimated cell size calculated by the candidate angle information calculating means. Is used as a specified range, an expansion filter of a size determined based on the estimated cell size is applied, and corner detection is performed using the image of the specified range after the expansion filter is applied, thereby re-determining the outer shape of the information code It is characterized by doing.

  In this way, by applying the expansion filter, the information code becomes an image that is closer to being filled than the original image, and small corners in the information code are reduced. Since corner detection is performed using the image after applying the expansion filter, it is difficult to detect small corners in the information code, and it is possible to accurately detect four corners that define the outer shape of the information code. As a result, the reading accuracy of the information code is improved.

  For example, as described in claims 7 and 8, the above-described code type determination means determines whether the frequency distribution of the edge angle in the one-dimensional code or the frequency distribution of the edge angle in the two-dimensional code. According to a seventh aspect of the present invention, in the frequency distribution of the edge angle, the code type determining unit, when the frequency ratio within a predetermined angle range centered on the maximum frequency is equal to or higher than a predetermined ratio, It is determined that the frequency distribution is a one-dimensional code. According to another aspect of the present invention, the code type determination unit traverses the code outline in a vertical scanning axis direction, and a frequency distribution of edge angles from one end of the code outline to the other end, and the code outline horizontally When the degree of coincidence with the frequency distribution of the edge angle from one end to the other end of the code outer shape crosses in the scanning axis direction is a certain value or less, the frequency distribution is a frequency distribution of a one-dimensional code judge.

The object can also be achieved by the invention of claim 9. The invention according to claim 9 is
(A) An information code reading device that includes a camera, images an information code by the camera, and reads information from the information code based on the captured image,
(B) feature point detecting means for detecting a feature point including at least a corner from an image photographed by the camera;
(C) for each feature point detected by the feature point detection means, feature information determination means for determining feature information that is at least one of edge strength and edge line vector,
(D) Based on the feature information, from the feature point detected by the feature point detection means, the corner is a corner that is close to another corner and close to the corner and the feature information is also approximated A corner group determining means for determining a group;
(E) outer shape estimation means for estimating the outer shape of the information code based on the overall shape of the corner group determined by the corner group determination means;
(F) A code that determines an edge angle at a plurality of points inside the code outline estimated by the outline estimation means, and determines a code axis angle that is a reference angle of the information code based on the determined plurality of edge angles Shaft angle determining means;
(G) Calculate the black and white change number and the black and white change width inside the code outline estimated by the external shape estimation means and on lines orthogonal to each other, and based on the calculated black and white change number and black and white change width, Cell size determining means for determining the cell size;
(H) Decoding processing means for decoding an information code based on the code outline estimated by the outline estimation means, the code axis angle determined by the code axis angle determination means, and the cell size determined by the cell size determination means It is characterized by including these.

  Also in the invention according to claim 9, feature points including at least a corner are detected from an image photographed by a camera, and at least one of edge strength and edge line vector size is detected for each feature point. On the other hand, certain feature information is determined. Further, the corner group determining means determines a corner group that is a corner from a feature point and that is close in distance to other corners and that approximates the adjacent corner and the feature information. If the corners belong to the same two-dimensional code, the feature information is a value approximate to each other, and the distance to other corners is also short. Therefore, the corner group determined by the corner group determining means can be considered as a group of corners included in the two-dimensional code. Therefore, the outer shape estimation means estimates the outer shape of the information code based on the overall shape of the corner group. Thus, in the present invention, since the code outline is estimated without detecting a feature pattern such as a finder pattern, the code outline can be estimated even for an image in which the feature pattern cannot be detected. it can.

  Further, the code axis angle and the cell size are also determined without using the feature pattern. That is, the code axis angle determines the edge angle at a plurality of points inside the code outline, and determines the code axis angle based on the determined plurality of edge angles. The cell size is also determined based on the number of black and white changes and the black and white change width on mutually orthogonal lines inside the code outline. The decoding processing means performs decoding processing based on the code outline, code axis angle, and cell size determined in this way. Therefore, since the decoding process can be performed even if the feature pattern cannot be detected, it is less likely that the information code cannot be read.

It is a fragmentary longitudinal cross-section which shows the structure outline | summary of the housing 11 grade | etc., Of the information code reader 10 which concerns on embodiment of this invention. It is a block diagram which shows the structure outline | summary of the circuit part 20 of the optical information reader 10 which concerns on embodiment. It is explanatory drawing which shows the structure of the marker light projector. 3 is a flowchart illustrating an information reading process executed by a control circuit 40 in the first embodiment. It is a figure which shows an example of the image which the area sensor 23 took. It is a figure explaining the calculation formula of an edge angle. (A) is an example of a histogram of edge angles, and (B) is a two-dimensional code image in which the histogram of (A) is created. 10 is a flowchart illustrating an information reading process executed by a control circuit 40 in Modification 2 of the first embodiment. (A) is an original image before the expansion filter processing, and (B) is an image after the expansion filter processing. 6 is a flowchart illustrating an information reading process executed by a control circuit 40 in the second embodiment.

(First embodiment)
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing a schematic configuration of an information code reading apparatus 10 to which the present invention is applied.

  As shown in FIG. 1, the information code reading device 10 is formed integrally with the housing body 11 in the shape of a round, thin, substantially rectangular box, and the lower surface of the housing body 11 approximately near the rear end of the center. And a gun-type housing comprising the grip portion 12. The grip portion 12 is set to have an outer diameter that allows an operator to hold the grip portion 12 with one hand, and an illumination light Lf and a marker light M1 described later are placed on a portion where the index finger of the operator holding the grip portion 12 abuts. A trigger switch 14 is provided for instructing the emission of light.

  A circuit unit 20 (to be described later) is accommodated in the housing main body 11, and a reading port 11a capable of emitting the illumination light Lf and receiving the reflected light Lr is formed at the tip of the housing main body 11. Has been. In the vicinity of the reading port 11a in the housing body 11, a marker light projector 60 that projects the marker light M1 is disposed. Although not shown in FIG. 1, a light emitting diode 21 (see FIG. 2) that irradiates illumination light Lf is also disposed near the reading port 11 a in the housing body 11. FIG. 1 also shows printed wiring boards 15 and 16 constituting the circuit unit 20, an area sensor 23 mounted on the printed wiring board 16, an imaging lens 27, and the like.

  As shown in FIG. 2, the circuit unit 20 mainly includes an optical system such as a marker light projector 60, a light emitting diode 21, an area sensor 23, and an imaging lens 27, a memory 35, a control circuit 40, an operation switch 42, and the like. Microcomputer system (hereinafter referred to as “microcomputer”), and a power supply system such as a power switch 41 and a battery 49, and mounted on the above-described printed wiring boards 15 and 16 or built in the housing body 11. Yes.

  As described above, the optical system includes the marker light projector 60, the light emitting diode 21, the area sensor 23, the imaging lens 27, and the like. The light emitting diode 21 is configured to be able to irradiate the illumination light Lf toward the reading surface R through the reading port 11 a of the housing body 11. 2 shows only one light emitting diode 21 for the sake of convenience, a plurality of light emitting diodes 21 are provided as appropriate. An information code D such as a one-dimensional code or a two-dimensional code is recorded on the reading surface R illuminated by the light emitting diode 21. The recording form includes printing, engraving, and the like.

  The area sensor 23 is configured to be able to receive the reflected light Lr irradiated and reflected on the reading surface R. For example, the area sensor 23 includes a light receiving element that is a solid-state image pickup element such as a C-MOS or a CCD. It is arranged in two dimensions of m rows and n columns in the order of 10,000. The light receiving element of the area sensor 23 is a photoelectric conversion element that converts received light into an electrical signal and outputs the electric signal, and may be hereinafter referred to as a pixel. Each light receiving element outputs a signal indicating the light received by the element. In the present embodiment, the area sensor 23 is assumed to be a monochrome sensor.

  The light receiving surface 23a of the area sensor 23 is positioned so as to be visible from the outside of the housing body 11 through the reading port 11a. The area sensor 23 receives incident light incident through the imaging lens 27 as the light receiving surface 23a. It is mounted on the printed wiring board 16 so that it can receive light.

  The imaging lens 27 functions as an imaging optical system capable of condensing incident light incident from the outside through the reading port 11a and forming an image on the light receiving surface 23a of the area sensor 23. It is comprised by the cylinder and the some condensing lens accommodated in this barrel. In the present embodiment, the illumination light Lf emitted from the light emitting diode 21 is reflected by the information code D, and the reflected light Lr incident on the reading port 11a is collected, whereby a code image is formed on the light receiving surface 23a of the area sensor 23. Imaging is possible.

  FIG. 3 is an explanatory diagram showing the configuration of the marker light projector 60. The marker light projector 60 projects the positioning marker light M indicating the reading position of the information code D onto the reading surface R, and is provided on the laser diode 62 and the emission side of the laser diode 62 as shown in FIG. A collimating lens 64, a diffraction grating plate 66 capable of forming a pattern of the marker light M for positioning, a field stop 68 provided with a slit having a predetermined shape, and the like.

  The field stop 68 is configured to be switchable to various slit shapes, and by switching the slit shape, the pattern of the positioning marker light M can be switched to various patterns. Note that switching of the pattern of the positioning marker light M by switching the slit shape is described in detail in Japanese Patent Application Laid-Open No. 2008-191999, and thus the description thereof is omitted here. In this embodiment, the color of the positioning marker light M is red.

  Returning to FIG. 2, the outline of the configuration of the microcomputer system will be described next. The microcomputer system includes an amplification circuit 31, an A / D conversion circuit 33, a memory 35, an address generation circuit 36, a synchronization signal generation circuit 38, a control circuit 40, an operation switch 42, an LED 43, a buzzer 44, a communication interface 48, and the like. Yes. As its name suggests, this microcomputer system is composed mainly of a control circuit 40 and a memory 35 that can function as a microcomputer (information processing device), and the image signal of the code image captured by the optical system described above is hardware. It can perform signal processing in terms of hardware and software. The control circuit 40 also performs control related to the entire system of the optical information reading apparatus 10.

  The image signal (analog signal) output from the area sensor 23 of the optical system is amplified by a predetermined gain by being input to the amplifier circuit 31, and then input from the analog signal when input to the A / D conversion circuit 33. Converted into a digital signal. The digitized image signal is input to the memory 35 and stored. The synchronization signal generation circuit 38 is configured to generate a synchronization signal for the area sensor 23 and the address generation circuit 36. The address generation circuit 36 is based on the synchronization signal supplied from the synchronization signal generation circuit 38. Thus, the storage address of the image signal stored in the memory 35 can be generated.

  The memory 35 is a semiconductor memory device, and corresponds to, for example, a RAM (DRAM, SRAM, etc.) or a ROM (EPROM, EEPROM, etc.). In addition to the image processing program, the ROM stores in advance a system program that can control each hardware such as the marker light projector 60, the light emitting diode 21, and the area sensor 23.

  The control circuit 40 is a microcomputer that can control the entire optical information reader 10 and includes a CPU, a system bus, an input / output interface, and the like. . As the information processing function, for example, there is an information reading process in which an image signal stored in the memory 35 is processed and information is read from an information code included in the image signal.

  Further, the control circuit 40 is configured to be connectable to various input / output devices (peripheral devices) via a built-in input / output interface, and includes a power switch 41, operation switch 42, LED 43, buzzer 44, and communication interface 48. The marker light projector 60 and the like are connected.

  Thereby, for example, monitoring and management of the power switch 41 and the operation switch 42, turning on / off the LED 43 functioning as an indicator, turning on / off the buzzer 44 capable of generating a beep sound and an alarm sound, and serial communication with an external device The communication control of the communication interface 48 enabling the control, the projection control of the marker light M from the marker light projector 60, and the like are enabled. The operation switch 42 includes the trigger switch 14 described above.

  The power supply system includes a power switch 41, a battery 49, and the like. When the power switch 41 managed by the control circuit 40 is turned on and off, the conduction of the drive voltage supplied from the battery 49 to each device and each circuit described above is established. Or shut off is controlled. The battery 49 is a secondary battery that can generate a predetermined DC voltage, and corresponds to, for example, a lithium ion battery.

  Next, the information reading process of the control circuit 40 will be described in detail. FIG. 4 is a flowchart showing this information reading process. The process shown in FIG. 4 is executed each time the area sensor 23 captures an image including the information code D represented on the reading surface R and an image signal indicating the image is accumulated in the memory 35.

  First, in step S10, an image signal is fetched from the memory 35 and edge detection processing is performed. In this edge detection process, the luminance value of each pixel is determined from the image signal, and an edge on the image is detected based on the luminance value. Various known edge detection methods can be used.

  The subsequent step S12 corresponds to the feature point detection means in the claims, and detects a corner on the image using a known corner detection method based on the edge detected in step S10. The corners that can be detected here are the intersections of the edges, and various corners are detected in addition to the four corners that define the outer shape of the information code D. In addition to corners, feature points other than corners such as inflection points may be detected. Furthermore, some or all of the four corners (hereinafter also referred to as information code corners) that define the outer shape of the information code D may not be detected by this processing. FIG. 5 is a diagram showing an example of an image captured by the area sensor 23, and the image shows the corner detected in step S12 with a circle.

  The subsequent step S14 corresponds to the information code angle candidate extracting means in the claims, and extracts information code angle candidates that are candidates for the information code angle from the corner detected in step S12. In the extraction process here, for example, the length of the edge is obtained using Hough transform. Further, there are two conditions for extraction, an edge line vector condition and a corner position condition, and those satisfying both conditions are extracted as information code angle candidates. Specifically, the condition of the edge line vector is that the length of the edge line vector is not less than a predetermined length, and the intersection angle of two edge lines that intersect at the corner is approximately 90 degrees (90 degrees ± α). Range). On the other hand, the corner position condition is specifically a position with respect to another corner, and more specifically, the direction in which the other corner exists is concentrated in a range of about 90 degrees when viewed from the target corner. (When the direction of other corners is calculated, the range of 90 degrees falls within a predetermined ratio).

  The number of information code angle candidates obtained by executing step S14 is considerably small. However, only the information code angle is not always extracted. For example, when a character including a shape close to a right angle is included in the image, the shape may be extracted as an information code angle candidate.

  Therefore, in the subsequent step S16, for each of the information code angle candidates extracted in step S14, the information code angle is estimated as the angle candidate information that is different from the candidate but not the information code angle. Calculate the cell size.

  For the calculation of the estimated cell size, a change in edge strength within a predetermined range including the information code angle candidate is used. The estimated cell size is determined based on a preset condition from the black and white change number and the black and white change width of the edge intensity in this range. The preset condition is, for example, in a histogram in which the black-and-white change width is the horizontal axis (class) and the black-and-white change number is the vertical axis (frequency), the change width having the maximum frequency is the number of cells (for example, 1 Cell or 2 cells), and the estimated cell size is determined from the number of cells determined from the histogram and the monochrome change width. This step S16 corresponds to corner candidate information calculation means in the claims.

  In step S18 corresponding to the outer shape estimation means in the claims, two to four information code corner candidates whose corner candidate information (ie, estimated cell size) calculated in step S16 is similar to each other are selected, and the information is selected. Determine as the chord angle. The determination that the estimated cell sizes are similar to each other is made, for example, by determining that the difference or ratio between the estimated cell sizes calculated for each information code angle candidate is within a predetermined difference or within a predetermined ratio range. . If the information code angle candidates cannot be narrowed down to 2 to 4 even if narrowing within a predetermined difference or based on a predetermined ratio, the candidates having the largest difference or ratio with other candidates are excluded in order.

  In step S18, the code outer shape of the information code is further estimated using the determined two to four information code angles. When four information code angles are determined by the above-described processing using the estimated cell size, a quadrilateral passing through the four information code angles is set as a code outline. On the other hand, when only two or three information code angles are determined, the code outer shape is estimated using the information code angles and the information code being rectangular. If only two are determined, for example, the two information codes are in a diagonal relationship or the same, based on the positional relationship between the straight line connecting the two information code angles and the determined corner. It is determined whether it is at both ends of the side. That is, when there is a diagonal relationship, corners are distributed on both sides of a line connecting two information code angles, but when they are both ends of the same side, one of the lines connecting two information code angles The corners are concentrated on the side. Using this, it is determined whether the two information codes are in a diagonal relationship or both ends of the same side.

In the following step 20, first, an edge angle is determined at a plurality of points in the code outline estimated in step S18. This edge angle is obtained from the luminance value of 3 × 3 pixels centering on the edge point, and is calculated from the following equation 1. In the following formula, “angle” is an edge angle. Dx and dy have the meanings shown in FIG. That is, dx is a difference between the total luminance value of the right three pixels sandwiching the edge point E and the total luminance value of the right three pixels, and dy is the total luminance value of the lower three pixels sandwiching the edge point E and the upper three pixels. This is the difference from the total luminance value.
(Formula 1) angle = arctan (dx / dy)
A histogram is created by using edge angles obtained at a plurality of points in the code outline using the above equation 1 and a predetermined angle width (for example, 1 degree, 5 degrees, etc.) as one section. The angle having the highest frequency in the histogram is defined as the code axis angle. Note that the plurality of points change depending on the size of the code outline, and as an example, the lattice point for each predetermined pixel in the code outline is a point for obtaining the edge angle. This step S20 corresponds to the code axis angle determining means in the claims.

  In subsequent step S22, the cell size is determined. Specifically, scanning is performed along a set of lines that are orthogonal to each other in the code outline estimated in step S18 and one line is parallel to the code axis determined in step S21. Then, the number of black and white changes in the edge strength and the black and white change width are calculated. The black-and-white change number and the black-and-white change width may be calculated a plurality of times while shifting the set of lines. Then, the cell size is determined based on the preset condition from the calculated monochrome change number and monochrome change width. The preset condition is, for example, creating a histogram in which the black-and-white change width is divided in a certain unit to be a horizontal axis, and the number of each division is a vertical axis (frequency). In this histogram, the change width having the maximum frequency is set to a certain number of cells (for example, 1 cell or 2 cells is determined in advance), and the estimated cell size is determined from the number of cells determined from the histogram. This step S22 corresponds to the cell size determining means in the claims.

  The subsequent step S24 corresponds to the code type determining means in the claims, and determines the code type of the information code D. The code type here means whether it is a one-dimensional code or a two-dimensional code. In this determination, an edge angle histogram (for example, one created in step S20) is used. FIG. 7A is an example of the histogram of the edge angle, and FIG. 7B is a two-dimensional code image in which the histogram of FIG. 7A is created.

  Whether the captured information code D is a one-dimensional code or a two-dimensional code by determining whether the histogram is an edge angle histogram in a one-dimensional code or an edge angle histogram in a two-dimensional code Determine. As a specific determination method, for example, a reference value of a feature amount that can be calculated from a histogram is set in advance for each of a one-dimensional code histogram and a two-dimensional code histogram. In the case of a one-dimensional code, the edge angle is concentrated in a certain angle range, whereas in the case of a two-dimensional code, the edge angle has at least two angle ranges in which the frequency is concentrated. Therefore, for example, as the feature amount, there is a ratio of the total frequency in the range of 45 degrees with respect to the maximum value of the histogram to the total frequency. If this ratio is greater than or equal to a preset reference ratio, it means that the edge angle is concentrated in a certain range, and therefore it is determined that the histogram is a one-dimensional code. Also, the histogram of the edge angle from one end of the code outline to the other end and the code outline across the horizontal scan axis direction across the code outline in the vertical scanning axis direction and from one end of the code outline A histogram of edge angles to the other end is obtained, and when the degree of coincidence is equal to or less than a certain value, the histogram can be determined to be a frequency distribution of a one-dimensional code. In addition, a standard one-dimensional histogram or two-dimensional histogram shape itself is used as a reference pattern, and a one-dimensional code histogram or two-dimensional is calculated based on a correlation value between the reference pattern and the shape created in step S20. It may be determined whether the histogram is a code.

  Step S26 corresponds to the decoding processing means in the claims, and performs decoding processing. In the decoding process, information on the code size, the code inclination, and the cell size constituting the code is required. In other words, if these are known, decoding is possible. In the flowchart of FIG. 4, the code size, the code inclination, and the cell size constituting the code are obtained in steps S18, S20, and S22, respectively. Therefore, decoding processing is performed using the values obtained in steps S18, S20, and S22. In the decoding process, a plurality of decoding algorithms determined for a specific type of code (for QR code, etc.) are sequentially used. In this step S26, the decoding algorithm belonging to the code type determined in step S24 is first used. Used for decoding.

  As described above, according to the present embodiment described above, information code angle candidates are extracted in order to specify the position of the information code D in the image (step S14). In order to extract the information code angle candidate, a corner is detected from the captured image (step S12), and the information code angle candidate is extracted based on the edge line vector passing through the corner and the position of the feature point (step S14). .

  The corners that can be detected from the image include not only the corners that define the outer shape of the information code D but also the corners that are formed by the cells in the information code. May be included to some extent. However, in step S14, information code angle candidates are extracted based on the edge line vector and the position of the corner, and therefore, parts other than the corner and corners in the information code D are not extracted. Therefore, the information code angle candidates are narrowed down to a certain number.

  However, if there is another figure (such as a character) having a corner shape in addition to the information code D in the imaging range, the corner on the figure may be extracted as an information code corner candidate. Therefore, an estimated cell size is further calculated for each information code angle candidate based on the luminance change around the information code angle candidate (step S16). Among the information code angle candidates, for the real information code angle, the surrounding luminance change reflects the cell size of the information code D. On the other hand, for candidates that are extracted as information code corner candidates but are not actually information code corners (for example, the corners of the graphic described above), the luminance change around them reflects the cell size of the information code D. However, it is very different from the actual information code angle. Accordingly, the estimated cell size calculated based on the luminance change around the information code angle candidate is greatly different between the case of the information code angle and the case of not being the information code angle. In other words, the estimated cell sizes of information code angle candidates that are the angles that define the outer shape of the same information code are similar to each other. Therefore, in step S18, based on the degree of similarity of the estimated cell sizes, 2 to 4 candidates are selected from the information code angle candidates as information code angles, and the code outer shape of the information code D is estimated based on this. Yes.

  As described above, in the present embodiment, the code outer shape is estimated without detecting a feature pattern such as a finder pattern, and therefore the code outer shape can be estimated even for an image in which the feature pattern cannot be detected. . In addition, four information code angles are not essential, and the code outer shape is estimated from two or three information code angles. Therefore, there are fewer cases where the code outline cannot be determined.

  Further, the code axis angle and the cell size are also determined without using the feature pattern. That is, as for the code axis angle, the edge angle is determined at a plurality of points inside the code outline, and the code axis angle is determined from the histogram of the edge angle (step S20). The cell size is also determined based on the number of black and white changes and the black and white change width on the mutually orthogonal lines inside the code outline (step S22). In step S26, decoding is performed based on the code outline, code axis angle, and cell size determined in this way. Therefore, even if the image cannot detect the feature pattern, the decoding process can be performed, so that information cannot be read from the information code.

Further, according to the present embodiment, it is determined whether the information code D is a one-dimensional code or a two-dimensional code by using a histogram of edge angles, and based on the determination result, decoding for the one-dimensional code is performed. Since it is determined which one of the algorithm and the decoding algorithm for the two-dimensional code is used first, decoding can be performed at high speed.
(Modification 1 of the first embodiment)
In the first modification, the pixel information of all the pixels stored in the memory 35 is not acquired, but the pixel information is acquired from the pixel information of all the pixels at a constant pixel interval (that is, acquired by thinning out the pixel information). ). In this way, a rough coarse image can be created for the captured image. This acquisition process corresponds to the rough image creating means in the claims.

  Then, the edge detection process in step S10, the corner detection process in step S12, and the information code angle candidate extraction process in step S14 are performed using the coarse image.

  When a coarse image is used in this way, unnecessary feature points such as small feature points inside the information code D are difficult to be detected, so that unnecessary candidates included in information code angle candidates are reduced. As a result, the information code angle can be easily selected.

Note that the pixel information acquisition interval in creating a rough image is determined based on the cell size. More specifically, the acquisition interval is increased as the cell size increases. As the cell size, the cell size determined at the time of reading the previous information code is stored in a storage unit such as the memory 35, and the cell size stored in the storage unit is used. Alternatively, the cell size may be input from the operation switch 42 by the user, and the pixel information acquisition interval may be determined based on the input cell size.
(Modification 2 of the first embodiment)
In the second modification, the flowchart shown in FIG. 8 is executed instead of the flowchart shown in FIG. Although FIG. 8 shows step S18 and subsequent steps, the same processing as in FIG. 4 is executed from step S10 to step S18.

  In this modification 2, as shown in FIG. 8, after performing step S18, it progresses to step S19-1. In step S19-1, a process of applying an expansion filter determined based on the estimated cell size estimated in step S16 to the information code angle determined in step S18 is performed. The designated range designated as the range to be subjected to the expansion filter is a range larger than the code outer shape estimated in step S18 by the estimated cell size estimated in step S16 with respect to the information code angle determined in step S18. Since there are four estimated cell sizes corresponding to the four information code angles, the average value is used. An expansion filter having a size based on the estimated cell size (for example, an average size of four estimated cell sizes) is applied to the designated range. The image in FIG. 9 is an image for explaining the effect of the expansion filter processing. FIG. 9A is an original image before the expansion filter processing, and FIG. 9B is an image after the expansion filter processing. is there. As can be seen from FIG. 9B, by performing the expansion filter process, the information code becomes an image that is entirely painted black, and there are almost no small corners.

  In subsequent step S19-2, the corner detection process is performed again on the image after applying the expansion filter in step S19-1. As shown in FIG. 9B, by applying an expansion filter, there are almost no small corners, so that four corners that define the outer shape of the information code can be detected with high accuracy. In the following step S19-3, the code outline is re-determined from the four corners detected in step S19-2. In this process, the four corners can be used as they are as the four corners of the code outline, or the positions of the four corners are moved by the amount expanded by the expansion filter to determine the corner of the code outline. May be.

  Subsequent steps S20 to S26 are the same as those in FIG. 4 except that the code outline redetermined in step S19-3 is used.

According to the second modification, corner detection is performed using an image in which small corners in the information code are reduced by applying an expansion filter. Therefore, it is possible to detect the four corners that define the outer shape of the information code with high accuracy, thereby improving the reading accuracy of the information code.
(Second Embodiment)
Next, a second embodiment will be described. The second embodiment has the same mechanical configuration as the first embodiment. However, in the second embodiment, information is read for a two-dimensional code. Also in the second embodiment, the control circuit 40 performs the information reading process, but the specific content of this process is different from the first embodiment.

  FIG. 10 is a flowchart showing information reading processing in the second embodiment. First, in step S30, the same edge detection process as in step S10 of FIG. 4 is performed. In the subsequent step S32, the same corner detection process as in step S12 of FIG. 4 is performed. This step S32 corresponds to feature point detection means.

  The subsequent step S34 corresponds to the feature information determining means in the claims, and the feature information is determined for each of the plurality of corners detected by the corner detection processing in step S32. In the present embodiment, this feature information is an edge line vector and an edge strength. This step S34 corresponds to the characteristic information determining means in the claims.

  In the subsequent step S36, the corners included in the two-dimensional code are defined as the corner group from the feature information determined in step S34 and the positions of the corners from the corner detected in step S32. Since the two-dimensional code has a large number of cells arranged vertically and horizontally, when corner detection processing (S32) is performed, a large number of corners are detected in the two-dimensional code. Moreover, since they are in the same code, the corners in the two-dimensional code are similar in edge strength and edge line vector size (the above-mentioned feature information). In other words, it can be estimated that the range where the corners having similar feature information are dense is the range of the two-dimensional code. Therefore, in step S36, the corner is extracted based on the condition that the distance is close to other corners and the feature information determined in step S34 is also approximated to the close corner. Then, the entire extracted corner is set as a corner group. Whether or not the distance is close is determined by, for example, the number of pixels between corners being within a predetermined number of pixels. Further, whether or not the feature information is close is determined by the difference or ratio of the feature information being within a predetermined value. This step S36 corresponds to the corner group determining means in the claims.

  The subsequent step S38 corresponds to the outer shape estimation means in the claims, and estimates the outer shape of the two-dimensional code based on the overall shape of the corner group determined in step S36. In the present embodiment, first, a range for applying the expansion filter is determined from the overall shape of the corner group. For example, a rectangle including all corners in the corner group (a circumscribed rectangle in the corner group) is set as the designated range. Then, an expansion filter is applied to the specified range. The size of the expansion filter is, for example, the size of the estimated cell size (or a size determined based on the size), and the estimated cell size is the same as that of the first embodiment with respect to part or all of the estimated outer shape. It calculates | requires similarly to step S22. When the expansion filter is applied in this way, corner detection processing is performed as in step S19-2 of FIG. Further, the code outer shape is re-determined as in step S19-3.

  In step S40, the code shaft angle is determined in the same manner as in step S20 of FIG. In step S42, the cell size is determined in the same manner as in step S22 of FIG. In step S44, decoding is performed using the code outline, code axis angle, and cell size determined in steps S38 to S42. These steps S40, S42, and S44 correspond to the code axis angle determining means, the cell size determining means, and the decoding processing means, respectively, in the claims.

  As described above, according to the second embodiment described above, after corner detection processing, feature information (edge strength, edge line vector) of each corner is determined (step S34). Then, the corners of the two-dimensional code that are close in distance to the other corners and approximate the feature information of the adjacent corners are determined as the corner group of the two-dimensional code (step S36), and the two-dimensional code is based on the overall shape of the corner group. Is estimated (step S38). As described above, in the second embodiment as well, the code outline is estimated without detecting the feature pattern. Therefore, the code outline can be estimated even for an image in which the feature pattern cannot be detected. Further, the code axis angle and the cell size are also determined without using the feature pattern. Therefore, even if the feature pattern cannot be detected, the decoding process can be performed, so that information cannot be read from the information code.

As mentioned above, although embodiment of this invention was described, this invention is not limited to the above-mentioned embodiment, The following embodiment is also contained in the technical scope of this invention, and also the summary other than the following is also included. Various modifications can be made without departing from the scope.
However, only one of them may be determined as feature information.

  For example, in the above-described embodiment, the area sensor 23 is a monochrome sensor, but a color sensor may be used as the area sensor 23. However, since the information code is a cell (or line) and the other two colors, it is preferable to process with a pair of colors (for example, black and white) when reading the information code.

  In the second embodiment described above, the edge line vector and the edge strength are determined as the feature information. However, only one of them may be determined as the feature information.

10: Information code reading device 11: Housing body 11a: Reading port 12: Grip part 14: Trigger switch 15: Printed wiring board 16: Printed wiring board 20: Circuit part 21: Light emitting diode 23 : Area sensor (light receiving sensor), 23a: light receiving surface, 27: imaging lens, 31: amplification circuit, 33: A / D conversion circuit, 35: memory, 36: address generation circuit, 38: synchronization signal generation circuit, 40 : Control circuit, 41: power switch, 42: operation switch, 43: LED, 44: buzzer, 48: communication interface, 49: battery, 60: marker light projector, 62: laser diode, 64: collimating lens, 66: Times Lattice plate, 68: Field stop, D: Information code, E: Reading range, Lf: Illumination light, Lr: Reflected light, M: Positioning marker light, R: Reading surface, S12: Feature point detection means, S14: Information Code angle candidate extraction means, S16: Corner candidate information calculation means, S18: Outline estimation means, S20: Code axis angle determination means, S22: Cell size determination means, S24: Code type determination means, S26: Decode processing means, S32: Feature point detection means, S34: Feature information determination means, S36: Corner group determination means, S38: Outline estimation means, S40: Code axis angle determination means, S42: Cell size determination means, S44: Decode processing means

Claims (9)

An information code reading device that includes a camera, images an information code by the camera, and reads information from the information code based on the captured image;
Feature point detecting means for detecting a feature point including at least a corner from an image photographed by the camera;
Information code angle candidate extraction means for extracting, from the feature points detected by the feature point detection means, information code angle candidates that are candidates for the corners of the information code based on edge line vectors passing through the feature points and the positions of the feature points When,
Corner candidate information calculation means for calculating an estimated cell size for each information code angle candidate based on a luminance change in a predetermined range around the extracted information code angle candidate;
Based on the degree of similarity of the estimated cell size calculated by the corner candidate information calculation means, 2 to 4 candidates are selected as information code angles from the information code angle candidates, and the information code is selected based on the selected information code angle. An outer shape estimating means for estimating the outer shape of the quadrangular code that is the outer shape of
A code axis angle determination that determines an edge angle at a plurality of points inside the code outline estimated by the outline estimation means, and determines a code axis angle that is a reference angle of the information code based on the determined plurality of edge angles Means,
The black and white change number and the black and white change width within the code outline estimated by the outline estimation means and on the lines orthogonal to each other are calculated, and the cell size is calculated based on the calculated black and white change number and black and white change width. A cell size determining means for determining;
A decoding processor that decodes an information code based on a code outer shape estimated by the outer shape estimating unit, a code axis angle determined by the code axis angle determining unit, and a cell size determined by the cell size determining unit. An information code reading device. In claim 1,
The frequency distribution of the edge angle determined based on the edge angle determined for the plurality of edges by the code axis angle determination means is either the frequency distribution of the edge angle in the one-dimensional code or the frequency distribution of the edge angle in the two-dimensional code. A code type determining means for determining whether the information code is a one-dimensional code or a two-dimensional code by determining whether the information code is present;
The decoding processing means determines which one of the one-dimensional code decoding algorithm and the two-dimensional code decoding algorithm to use first based on the determination result of the code type determining means. Reading device. In claim 1 or 2,
A coarse image creating means for creating a coarse image obtained by extracting pixel information at a fixed pixel interval from an image photographed by the camera;
The feature code detection means and the information code angle candidate extraction means detect the feature points using the coarse image created by the coarse image creation means, and extract the information code angle candidates. Reading device. In claim 3,
Cell size storage means for storing the cell size at the time of reading the information code;
The coarse image creating means determines the roughness of the coarse image to be created based on the cell size at the time of reading the previous information code stored in the cell size storage means. In claim 3,
A cell size input means for a user to input a cell size;
The coarse image creating means determines the roughness of the coarse image to be created based on the cell size input to the cell size input means. In any one of Claims 1-5,
The outer shape estimation means uses the estimated cell as a specified range that is larger than the code outer shape estimated based on the information code angle selected from the information code angle candidates by the estimated cell size calculated by the corner candidate information calculation means. An information code characterized by applying an expansion filter of a size determined based on the size and performing corner detection using the image in the specified range after the expansion filter is applied, thereby re-determining the outer shape of the information code Reading device. In claim 2,
In the frequency distribution of the edge angle, the code type determining unit determines that the frequency distribution is a frequency of a one-dimensional code when a ratio of the frequency within a predetermined angle range centered on the maximum frequency is equal to or higher than a predetermined ratio. An information code reading device characterized in that it is determined as a distribution. In claim 2,
The code type determination means crosses the code outline in the vertical scanning axis direction, crosses the code outline in the horizontal scanning axis direction, and the frequency distribution of edge angles from one end to the other end of the code outline, When the degree of coincidence with the frequency distribution of the edge angle from one end of the code outer shape to the other end is equal to or less than a certain value, it is determined that the frequency distribution is a frequency distribution of a one-dimensional code Information code reader. An information code reading device that includes a camera, images an information code by the camera, and reads information from the information code based on the captured image;
Feature point detecting means for detecting a feature point including at least a corner from an image photographed by the camera;
For each feature point detected by the feature point detection means, feature information determination means for determining feature information that is at least one of edge strength and edge line vector,
Based on the feature information, a corner group is determined from the feature points detected by the feature point detection means, which is a corner and is close in distance to other corners, and which also approximates the adjacent corner and the feature information. A corner group determining means to perform,
An outer shape estimating means for estimating the outer shape of the information code based on the overall shape of the corner group determined by the corner group determining means;
A code axis angle determination that determines an edge angle at a plurality of points inside the code outline estimated by the outline estimation means, and determines a code axis angle that is a reference angle of the information code based on the determined plurality of edge angles Means,
The black and white change number and the black and white change width within the code outline estimated by the outline estimation means and on the lines orthogonal to each other are calculated, and the cell size is calculated based on the calculated black and white change number and black and white change width. A cell size determining means for determining;
Decoding processing means for decoding information code based on the code outer shape estimated by the outer shape estimating means, the code axis angle determined by the code axis angle determining means, and the cell size determined by the cell size determining means, An information code reading device comprising: