JP2001209751A - Adaptive binarization system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an adaptive binarization system capable of performing binarization processing being more faithful to an original image by simple processing. SOLUTION: This adaptive binarization system consists of a code reading part 1 that optically reads a two-dimensional code printed on a printing medium such as paper a zero crossing point detecting part 2 that applies secondary differentiation to an image signal read by the code reading part 1 and detects the zero crossing point of the secondary differentiation signal, a binarization threshold calculating part 3 that samples the luminance value of an image signal corresponding to the zero crossing point detected by the zero crossing point detecting part 2 to define it as a binarization threshold and a binarizing part 4 that binarizes the image signal with the binarization threshold calculated by the binarization threshold calculating part 3. Thus, binarization being faithful to the contour of an input image can be possible by utilizing a secondary differentiation value for the threshold calculation of binarization.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an adaptive binarization method most suitable for a two-dimensional code reader for optically reading a two-dimensional code printed on a printing medium such as paper.

[0002]

2. Description of the Related Art Generally, in optically reading a two-dimensional code printed on a printing medium such as paper, a multi-valued image signal obtained by optically capturing the two-dimensional code is binarized,
Various processes are often performed. There are many methods for the binarization technique of the multi-level image signal, and typical ones include a “fixed threshold method” and a “dynamic threshold method”. The “fixed threshold method” binarizes a multi-valued image signal with a fixed threshold value, and is characterized by very simple processing. However, the “fixed threshold value method” is vulnerable to uneven brightness of the entire screen (shading). Undesirable phenomena such as the appearance of black (or white) painted spots often occur. There is also a problem that a minute change in luminance cannot be binarized.

[0003] On the other hand, the "dynamic threshold method" is a method of calculating an optimum threshold value for each pixel, and does not cause a problem as in the "fixed threshold method", but requires an enormous processing time. Therefore, it is not suitable as a binarization method for a system that requires responsiveness, such as a two-dimensional code reader. Note that these methods are described in detail in many image processing reference books, for example, “Image Analysis Handbook (published by The University of Tokyo (1991)”).

[0004] In order to solve the above problems,
Japanese Patent Laid-Open No. 7-57080 proposes an adaptive binarization control method. This is, as shown in FIG.
, Trigger generation circuit 202, level setting circuit 203, comparison 2
A binarization circuit 204 is provided, and a steep rising portion of a luminance change point of the image signal is detected by the differentiating circuit 201, and the image signal at that portion is used as a threshold for binarization, thereby following the luminance change of the image signal. This is to enable binarization. Further, according to this method, the processing time can be significantly reduced as compared with the “dynamic threshold method”.

Next, the operation of the conventional example shown in FIG. 16 will be described with reference to FIG. The image signal a shown in FIG. 17A is differentiated by the differentiating circuit 201 in FIG. 16 to become a differentiated signal b shown in FIG. This differentiated signal b is compared by the trigger generation circuit 202 with a constant level signal LS in FIG. 17B, and when the differentiated signal b is larger than LS, the constant signal shown in FIG. It is converted into a pulse signal c having a width. The level setting circuit 203 in FIG. 16 samples and holds the input image signal a using the pulse signal c, and sets it as a threshold for binarization ((D) in FIG. 17). This binarization threshold d is input to the comparison binarization circuit 204 of FIG. 16 together with the input image signal a, and is converted into a binarization signal f as shown in FIG. FIG. 17E is a diagram in which the input image signal a and the binarization threshold d determined by the level setting circuit 203 are superimposed, and FIG.
FIG. 18 is a diagram showing the relationship between the pulse signal c of FIG. 17C and sample points of the input image signal a and the binarization threshold d.

[0006]

By the way, in the conventional binarization method (Japanese Patent Laid-Open No. 7-57080),
The differential signal b of the image signal a is a spike signal as shown in FIG. 17B, but the differential signal of the image signal obtained by capturing the two-dimensional code is not necessarily spike. For example, in the case of the image signal 210 shown in FIG. 18A, the differential signal is as shown by the reference numeral 211 in FIG. 18B. In the conventional example, the differential signal 211 is compared with the constant level signal LS210 to output a trigger signal. However, as shown in FIG. 18C, the trigger signal 212 is considerably ahead of the peak of the differential signal 211. Output. As a result, the binarization threshold TH210 is set, and the binarized signal becomes as indicated by reference numeral 213 in FIG. The binarized signal 213 lacks a portion indicated by a dotted line 214 as compared with a case where ideal binarization is performed.

In the prior art, only the rising portion of the luminance change is used. That is, in the case of the image signal 220 as shown in FIG. 19A, the differential signal is as shown by reference numeral 221 in FIG. 19B. As a result of comparing the differentiated signal 221 and the constant level signal LS220, as shown in FIG. 19C, the threshold signal TH22 is generated by the trigger signals 222 and 223.
0 and TH221 are set. However, this binarization threshold T
H220 is not suitable for binarizing the luminance change 226 in the image signal 220. Therefore, this binarization threshold TH220,
In the binarized signal 224, which is the binarized result by TH221, the dotted line portion indicated by 225 in FIG. 19D is missing.

The present invention has been made in order to solve the above-mentioned problems in the conventionally proposed adaptive binarization control system.
It is an object of the present invention to provide an adaptive binarization method capable of performing a binarization process more faithful to an original image while maintaining the simplicity of the process.

[0009]

According to a first aspect of the present invention, there is provided a code reading unit for optically reading a two-dimensional code printed on a printing medium such as paper, and a code reading unit for reading the code. A second-order differentiation of the image signal read by the section, a zero-crossing point detecting section for detecting a zero-crossing point of the second-order differential signal, and the image corresponding to the zero-crossing point detected by the zero-crossing point detecting section A binarization threshold calculation unit that samples a luminance value of a signal to be a binarization threshold, and a binarization unit that binarizes the image signal using the binarization threshold calculated by the binarization threshold calculation unit And constitute an adaptive binarization method.

In the adaptive binarization system configured as described above, the two-dimensional code read by the code reader is converted into a multi-level image signal by a zero-crossing point detector, a binarization threshold calculator, and a binarizer. The multi-level image signal is secondarily differentiated in a zero-crossing point detection unit, and a point at which the sign of the second-order differential signal changes (zero-crossing point) is detected. Then, this zero-crossing point is input as a trigger signal to the binarization threshold calculation unit, and the binarization threshold calculation unit samples the luminance of the input multi-level image signal at the position corresponding to the trigger signal position, and Is used to determine a binarization threshold up to the next trigger signal generation position and output it to the binarization unit. The binarization unit binarizes the input multi-level image signal using the binarization threshold.

In the invention according to claim 1 in which the above operation is performed, the following advantages can be obtained. That is, the second derivative of the image signal is widely used for extracting the contour component of the image signal, and by using this second derivative for calculating the threshold for binarization, the second derivative is faithful to the contour of the input image. Binarization becomes possible. Also, the processing target is not the “luminance value of the image signal” but the “degree of change in the luminance value”. Therefore, even if there is luminance unevenness in the same screen due to the influence of shading, etc. If the brightness unevenness is such that the outline of the dimensional code can be recognized, there is no influence. Therefore, the configurations of the illumination system and the optical system can be simplified. In addition, many of the widely used methods for binarizing an image include some means (such as creating a histogram, detecting a maximum value and a minimum value from an area of a certain size including a pixel to be binarized). And the like, statistical information is collected, and an optimal threshold for binarization is determined for binarization. However, according to the method of the present invention, it is sufficient to have information for several pixels in the vicinity of the pixel to be binarized, and responsiveness close to that of the fixed threshold method can be obtained.

According to a second aspect of the present invention, in the adaptive binarization method according to the first aspect, a stage prior to the zero-crossing point detecting section is provided.
A smoothing filter for removing a noise component superimposed on the image signal read by the code reading unit.

In the adaptive binarization method configured as described above, the two-dimensional code read by the code reading unit is input to a smoothing filter as a multi-valued image signal, and is subjected to smoothing by the smoothing filter. The multi-valued image signal from which the noise component has been removed is input to the zero-crossing point detection unit. The zero-crossing point detector performs second-order differentiation of the smoothed image signal, detects a zero-crossing point of the second-order differentiated signal, and the zero-crossing point is input to the binarization threshold calculator as a trigger signal. You. In the binarization threshold value calculation unit, the image signal before performing the smoothing, or the image signal after performing the smoothing, sample the luminance of the position corresponding to the trigger signal position, this value as a binarization threshold Output to the binarization unit. The binarization unit binarizes the image signal before performing the smoothing or the image signal after performing the smoothing using the binarization threshold. According to the second aspect of the present invention, the smoothing filter is arranged in front of the zero-crossing point detection unit, and the binarization threshold value calculation unit and the input signal to the binarization unit are not particularly limited. .
Therefore, both the binarization threshold value calculation unit and the binarization unit may use either the signal before filtering or the signal after filtering.

According to the second aspect of the present invention in which the above operation is performed, the following advantages can be obtained. That is, as a property of the second derivative signal, it has a disadvantage that it reacts sensitively to the contour component of the image signal, but also reacts to noise such as a high frequency component superimposed on the image signal. In order not to react to such noise, unnecessary noise components may be removed from the image signal before inputting the image signal to the second differentiator in the zero-crossing point detector. Therefore, the risk of erroneously recognizing noise as a contour component can be reduced by providing a smoothing filter, which is a typical noise removal filter, as in the present invention and removing high-frequency components.

According to a third aspect of the present invention, in the adaptive binarization method according to the first aspect, the zero-crossing point detecting section includes a first-order differentiating section for performing a first-order differentiation of the image signal, and a first-order differentiating section of the image signal. When the value exceeds a predetermined positive / negative threshold value, the zero crossing point of the second derivative signal of the image signal is made effective.

In the adaptive binarization system configured as described above, the two-dimensional code read by the code reader is converted into a multi-level image signal by a zero-crossing point detector, a binarization threshold calculator, a binarization threshold, The zero-crossing point detection unit performs second-order differentiation of the input multi-valued image signal and, in parallel with detecting the zero-crossing point of the second-order differential signal, performs first-order differentiation on the input image signal. I do. The primary differential signal is binarized by two positive and negative thresholds having a predetermined value, and a maximum point and a minimum point having a certain magnitude or more are detected. The zero crossing point of the second derivative signal is masked by an extreme value detection signal obtained by binarizing the first derivative signal, and the resulting signal has a first derivative value having a maximum value, a minimum value, and Only a zero-crossing point generated when the extreme value has a certain magnitude or more is a signal detected, and the other zero-crossing points are deleted. The zero-crossing point masked in this way is input as a trigger signal to the binarization threshold calculation unit, which samples the luminance value of the image signal corresponding to the trigger signal position, and Is output to the binarization unit as a binarization threshold. The binarization unit binarizes the input image signal using the binarization threshold.

In the invention according to claim 3 in which the above operation is performed, the following advantages can be obtained. That is, the first derivative takes an extreme value at a point where the luminance change of the image signal sharply changes (the contour of the image), and the larger the degree of the luminance change becomes, the larger the amplitude becomes. This is a derivative, and becomes zero at the extreme value of the primary differential value. That is, if the secondary differential value is zero (zero crossing point) and the primary differential value has a certain magnitude or more, the location can be regarded as not a noise but a contour component of the image signal. The smoothing filter in the invention according to claim 2 is for suppressing the occurrence of an unnecessary zero-crossing point. However, the invention according to claim 3 considers the unnecessary zero-crossing point as inevitable. The main focus is on extracting a truly significant zero-crossing point from them. Therefore, also in the invention according to the present invention, it is possible to avoid erroneous recognition of a contour component due to noise.

According to a fourth aspect of the present invention, in the adaptive binarization method according to the first aspect, the binarization threshold value calculating section calculates the image signal corresponding to a zero crossing point of a second derivative signal of the image signal. Is sampled, and the binarization threshold value of each pixel between the zero-crossing points is determined by interpolating between the zero-crossing points with the sampled luminance value.

In the adaptive binarization system configured as described above, the two-dimensional code read by the code reading unit is converted into a multi-valued image signal by a zero-crossing point detection unit, a binarization threshold calculation unit, and a binarization unit. At the zero-crossing point detector,
The image signal is secondarily differentiated, and a zero crossing point of the second differentiated signal is detected. The zero-crossing point is input as a trigger signal to the binarization threshold calculator, and the binarization threshold calculator samples the luminance value of the input image signal corresponding to the trigger signal point. At this time, the binarization threshold calculation unit interpolates the binarization threshold of the pixel between the immediately preceding trigger signal point and the current trigger signal point between the two trigger signal points using the sampled luminance value. Is set in units of one pixel. That is, in the input image signal, a line connecting the luminance values corresponding to the trigger signal points is a binarization threshold for binarizing the image signal. However, the line connecting the trigger signal points (binarization threshold) does not need to be a straight line. The binarization unit binarizes the input image signal using the binarization threshold.

In the invention according to claim 4 in which the above operation is performed, the following advantages can be obtained. That is, as a result of adaptively calculating the binarization threshold for each contour position of the image signal, a large difference may occur between adjacent binarization thresholds. In such a case, the shape of the binarized signal may be distorted at the portion where the binarization threshold is switched. However, if the binarization thresholds are set so as to connect the edge components as in the invention according to the fourth aspect, the binarization thresholds are continuously switched, so that distortion of the shape does not occur.

According to a fifth aspect of the present invention, in the adaptive binarization method according to the first aspect, when the secondary differential signal of the image signal exceeds a predetermined positive / negative threshold value, the binarization unit performs the binary conversion. And a signal synthesizing unit for correcting the binarized signal by using the second derivative signal.

In the adaptive binarization system configured as described above, the two-dimensional code read by the code reading unit is converted into a multi-valued image signal by a zero-crossing point detection unit, a binarization threshold value calculation unit, and a binarization unit. At the zero-crossing point detector,
The image signal is secondarily differentiated, and a zero crossing point of the second differentiated signal is detected. This zero-crossing point is output as a trigger signal to the binarization threshold calculator, which samples the luminance value of the input multi-level image signal corresponding to the trigger signal position, and converts this value into a binary value. It is output to the binarization unit as a binarization threshold. The binarization unit binarizes the input multi-level image signal using the binarization threshold, and outputs the binarized signal to the signal synthesis unit. On the other hand, the zero-crossing point detection section also outputs a secondary differential signal generated in the process in addition to the trigger signal, and the secondary differential signal is input to the signal synthesis section. The signal synthesizing unit binarizes the input secondary differential signal with a predetermined threshold value, and shapes the binarized signal binarized by the binarizing unit using the secondary differential binarized signal. I do.

In the invention according to claim 5 in which the above operation is performed, the following advantages can be obtained. In other words, the secondary differential signal becomes zero at a portion where the luminance change of the image signal is the largest (contour portion of the image signal), but forms positive and negative peaks before and after that. Therefore, if the peak of the second derivative signal near this contour (zero crossing point) is used, it is possible to reproduce the contour buried under the influence of the switching of the binarization threshold to some extent. Becomes The effect obtained by the invention according to claim 5 is similar to the effect of the invention according to claim 4, but is characterized in that it can be realized with a simpler configuration and processing.

[0024]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment] Next, an embodiment of the present invention will be described. FIG. 1 is a block diagram showing a first embodiment of the adaptive binarization system according to the present invention. This embodiment corresponds to the first aspect of the present invention. This embodiment is configured as follows. In FIG. 1, reference numeral 1 denotes a code reading unit, and the code reading unit 1 is means for optically reading a two-dimensional code, and is realized by an image sensor such as a CCD or a CMOS image sensor. Numeral 2 denotes a zero-crossing point detecting unit. The zero-crossing point detecting unit 2 comprises a second-order differentiating unit 5 and a sign change point detecting unit 6 as shown in FIG. The image signal is secondarily differentiated by the second differentiator 5, and a point (zero crossing point) at which the second differential signal greatly changes from positive to negative or from negative to positive is detected by the sign change point detector 6. Has become. The detected zero-crossing point is input to the binarization threshold value calculation unit 3 as a trigger signal. The binarization threshold calculator 3 samples and holds the luminance value of the multi-valued image signal output from the code reader 1 in response to a trigger signal from the zero-crossing point detector 2, and sets a binary threshold as a binarization threshold. Transformation part 4
Output to The binarization unit 4 compares the binarization threshold determined as described above with the multi-valued image signal output from the code reading unit 1 and binarizes it.

Next, the operation of this embodiment will be described with reference to the signal waveform diagram shown in FIG. It is assumed that the two-dimensional code captured by the code reading unit 1 becomes a multi-valued image signal 7 as shown in FIG.
A signal obtained by secondarily differentiating the multi-level image signal 7 (hereinafter referred to as a second-order differential signal) is as indicated by reference numeral 9 in FIG. In order to obtain the secondary differential signal 9, a multi-level image signal may be masked by a Laplacian operator as shown in FIG. 4A, but depending on the image signal, only the X direction is considered in FIG. The mask processing by the operator as shown in FIG. Also, an operator who widens the pixels so as to correspond to the data cycle of the two-dimensional code as shown in FIG. 4C is effective in reducing the influence of noise. The zero-crossing point detector 2 detects the zero-crossing points indicated by 10, 11, 12, and 13 in FIG. 3D from the secondary differential signal 9, and as shown in FIG. Trigger signals 14, 15, 16, 17 corresponding to the respective zero-crossing points are output. As shown in FIG. 3A, the binarization threshold calculator 3 samples and holds the corresponding point of the multi-level image signal 7 by using the trigger signal, and stores the binarization thresholds TH1 and TH.
2, TH3 and TH4 are output to the binarization unit 4. The binarization unit 4 performs the binarization thresholds TH1, TH2, TH3, T
By H4, the multi-level image signal 7 is binarized, and FIG.
Is obtained. Note that the signal waveform diagram shown in FIG. 3 is a conceptual diagram for clearly illustrating the principle of operation, and in practice, calculation of the second derivative requires a certain amount of time. Therefore, the binarization threshold calculation unit 3 and the binarization unit 4 have a function of adjusting the timing of the multi-level image signal, the trigger signal, and the timing of the binarization threshold.

On the other hand, the signal 8 in FIG. 3B is a signal obtained by differentiating (primary differential) the multi-level image signal 7 (hereinafter, a primary differential signal). According to the conventional example, the primary differential signal 8 is compared with the threshold value LS1, and a portion where the primary differential signal 8 exceeds the threshold value LS1 [19, 20 in FIG. 3B] is a threshold sampling point. The threshold value at this time is T in FIG.
H5 and TH6, and the binarized multi-level image signal 7 at the threshold value is as indicated by reference numeral 21 in FIG. When this binarized signal 21 is compared with the binarized signal 18 according to the present embodiment, the shape of the hatched portion 22 in FIG. 3C is not correctly binarized, and further the portion of the hatched portion 23 is missing. You can see that it is.

As described above, by using the zero-crossing point of the secondary differential signal to calculate the binarization threshold as shown in the present embodiment, it is possible to perform binarization more faithful to the original image.

[Second Embodiment] Next, a second embodiment of the present invention will be described.
An embodiment will be described. FIG. 5 is a block diagram showing the second embodiment, and the same components as those of the first embodiment shown in FIG. 1 are denoted by the same reference numerals. This embodiment corresponds to the first and second aspects of the present invention, and is characterized in that a smoothing filter 30 is provided in a stage preceding the zero-crossing point detection unit 2 of the first embodiment. .

Next, the operation of the present embodiment will be described with reference to the signal waveform diagrams of FIGS. By the way, the present invention is characterized in that a secondary differential signal of a multi-level image signal is used for calculating a binarization threshold. However, the second derivative signal is suitable for extracting features (particularly, contour components) of the image signal, but has another aspect in that it is sensitive to a slight change in the image signal. For example, the secondary differential signal of the multi-valued image signal 31 whose luminance changes slightly as shown in FIG. 6A is as indicated by reference numeral 32 in FIG.
The secondary differential signal 32 responding to the slight luminance change 33 of the multi-level image signal 31 causes zero crossing at positions 34 and 35 in FIG. 6B. In response to this zero crossing, binarization thresholds TH31 and TH32 which are originally unnecessary are set, and as shown in FIG. 6C, a noise 37 is generated in the binarization result 36. The hatched portion 38 is missing.

This embodiment suppresses the occurrence of an extra zero crossing in the second derivative signal by removing a high frequency noise component superimposed on the multi-level image signal.
That is, the two-dimensional code read by the code reading unit 1 is input to the smoothing filter 30 as a multi-valued image signal 31 indicated by a dotted line in FIG. The image signal from which high-frequency components have been removed by the smoothing filter 30 is
In FIG. 7A, the curve becomes a smooth curve like an image signal 39 indicated by a solid line. The zero-crossing point detection unit 2 performs second differentiation and zero-crossing point detection on the image signal 39 from which the high-frequency component has been removed.
As shown in (B) of FIG. 6, unnecessary zero-crossing as shown by 34 and 35 in (B) of FIG. 6 does not occur. Therefore, the binarization thresholds set by the second derivative signal 40 of the image signal 39 from which the high-frequency component has been removed are TH33 and TH as shown in FIG.
34, and unnecessary binarization thresholds such as TH31 and TH32 shown in FIG. 6A are not generated. Therefore, unnecessary noise does not occur in the binarization result 41 based on the binarization thresholds TH33 and TH34 as shown in FIG. 7C.

Note that the present embodiment can be variously modified. In the present embodiment, the output of the smoothing filter 30 is input only to the zero-crossing point detection unit 2, and the binarization threshold value calculation unit 3,
Although the output of the code reading unit 1 is input to the binarization unit 4, the output of the smoothing filter 30 is input to the binarization threshold calculation unit 3 or the binarization unit 4. It may be configured. Further, the output from the code reading unit 1 and the smoothing filter 30 are sent to the binarization threshold value calculation unit 3 or the binarization unit 4.
And inputting both of them, and selecting the most suitable one, or combining the two by calculation. When the output of the smoothing filter 30 is input only to the zero-crossing point detector 2, the second-order differentiator (5 in FIG. 2) in the zero-crossing point detector 2 and the smoothing filter are combined. With one filter, the overall configuration can be simplified.

[Third Embodiment] Next, a third embodiment of the present invention will be described.
An embodiment will be described. This embodiment corresponds to the first and third aspects of the invention, and the configuration thereof is the same as that of the first embodiment in the basic configuration.
The zero-crossing point detection unit 2 in the first embodiment shown in FIG. 1 has the configuration shown in FIG. 8, and uses a first-order differential signal for zero-crossing point detection processing of a second-order differential signal. It is characterized by the following. Therefore, FIG. 1 is used for the basic configuration of the third embodiment.

Next, the configuration and the operation of the third embodiment will be described with reference to a signal waveform diagram shown in FIG. 9. The basic operation other than the zero-crossing point detecting section 2 is the same as that of the first embodiment. This is the same as the embodiment, and the description here is omitted. Therefore, here, the configuration and operation of the zero-crossing point detecting unit 2 shown in FIG. 8 will be mainly described. The two-dimensional code data read by the code reading unit 1 is input as a multi-valued image signal 60 shown in FIG. 9A to a first differentiating unit 51 and a second differentiating unit 52 constituting the zero-crossing point detecting unit 2. You. First derivative
Reference numeral 51 first-order differentiates the image signal 60, and outputs a first-order differential signal 61 shown in FIG. The primary differential signal 61 is further input to a positive direction binarization unit 54 and a negative direction binarization unit 55.

The positive direction binarization section 54 compares the binarization threshold TH61 of FIG. 9B with the primary differential signal 61 to perform a binarization process on the primary differential signal, and then performs a binarization process of FIG. ) Is output. The primary differential positive direction binarized signal 62 becomes “1” when the primary differential signal 61 is larger than the threshold value TH61. In the negative direction binarization section 55, similarly to the positive direction binarization section 54, the primary differential signal 61 is obtained by the threshold value TH62.
Is binarized, and a primary differential negative direction binarized signal 64 shown in FIG. 9D is output. This first derivative negative direction binarized signal
64 becomes “1” when the primary differential signal 61 is smaller than the threshold value TH62.

On the other hand, the secondary differentiating section 52 performs secondary differentiation on the image signal 60 and outputs a secondary differential signal 66 shown in FIG.
The secondary differential signal 66 is input to the sign change check detection unit 56, and the point at which the secondary differential signal 66 changes from positive to negative or from negative to positive is the secondary differential sign change point in FIG. This is output as a pulse train indicated by the signal 67. However, it is clear from the comparison between the image signal 60 in FIGS. 9A and 9F and the pulse train 67, and this pulse train 67 includes a signal which is clearly not related to the contour component. I have.

The primary differential positive direction binary signal 62, the primary differential negative direction binary signal 64, and the secondary differential sign change point signal (pulse train) 67 are input to a selector 57. Selector 57
The second differential sign change point signal 67 is converted to the first differential positive direction binarized signal.
Mask processing with 62 and the first derivative negative direction binarized signal 64,
The trigger signal 68 shown in FIG. 9 (G) is output. By this masking process, of the pulse train of the secondary differential sign change point signal 67, 63 in the primary differential positive direction binarized signal 62 shown in FIG. 9C and the primary differential negative signal shown in FIG. The pulses other than 69 and 70 pulses corresponding to the 65 positions in the direction binarized signal 64 are deleted. Therefore, the final trigger signal 68 has 71,
Only 72 pulses will appear, and any useless zero-crossings in response to small changes in the image signal 60 will be ignored. With this trigger signal 68, binarization is performed by the same processing as in the first embodiment of the present invention, and the binarization thresholds TH63 and TH64 shown in FIG. H)
Is obtained.

The present embodiment utilizes the property that the secondary differential value becomes zero at a position where the primary differential value takes a maximum and a minimum. As can be seen from FIGS. 9A and 9B, the first derivative value 61 is obtained at a position corresponding to the contour component of the original image signal 60.
Has a maximum and a minimum, and the value increases as the luminance changes more steeply. Therefore, the sign change point (zero crossing point) of the secondary differential value at a location (extreme value) where the primary differential value has a certain magnitude or more may be regarded as a contour component in the image. This makes it possible to suppress the property of easily reacting to noise at the same time while taking advantage of the characteristic of the second derivative that reacts to a slight change in luminance.

The present embodiment can be variously modified and applied. In this embodiment, the zero crossing point detecting unit is provided with the primary differentiating unit and the secondary differentiating unit. However, instead of the secondary differentiating unit, the output of the primary differentiating unit is further differentiated by the first order. You may. Also, the first derivative is not binarized by the positive binarization unit and the negative binarization unit, but is binarized only by the positive binarization unit by taking the absolute value of the first derivative. Etc. can also be used.

[Fourth Embodiment] Next, a fourth embodiment of the present invention will be described.
An embodiment will be described. FIG. 10 is a block diagram showing the fourth embodiment, in which the same components as those of the first embodiment shown in FIG. This embodiment corresponds to the first and fourth aspects of the present invention.

In each of the embodiments described above, as shown in FIGS. 11A and 11B, the binarization threshold values TH91, TH91, Assume that TH92 is obtained. When the image signal 90 is binarized using the binarization thresholds TH91 and TH92, a binarized signal 92 shown in FIG. 11C is obtained. However, the rising edge of the luminance level in the image signal 90 is at the position indicated by reference numeral 95 in FIG. 11A, and the binary signal is desirably indicated by reference numeral 93 in FIG. 11D. .

Considering the above viewpoints,
This is the present embodiment. Hereinafter, the configuration and operation of the present embodiment shown in FIG. 10 will be described. The two-dimensional code read by the code reading unit 1 is input to the zero-crossing point detection unit 2 and the line buffer 82. The zero-crossing point detection unit 2 includes a smoothing filter as exemplified in the second embodiment described above, or a secondary filter as exemplified in the third embodiment. It is desirable to have a differential noise elimination function. Line buffer 82
Is used to delay the image signal input from the code reading unit 1 by the time required for calculating the binarization threshold.

Then, similarly to the above-described embodiments, a trigger signal for sampling the binarization threshold is generated by the zero-crossing point detection unit 2 and the binarization threshold calculation unit 81
Is input to The binarization threshold calculation section 81 samples the image signal in response to the trigger signal, but does not immediately output the value as the binarization threshold in the present embodiment.
The binarization threshold of the pixel from the current trigger signal generation point to the immediately preceding trigger signal generation point is set for each pixel by interpolating the sampling luminance value at the two trigger signal generation points, and the binarization threshold is set. The data is stored in the storage buffer 83. That is, the present embodiment is characterized in that a line connecting the luminance values of the input image signal at the position where the trigger signal is generated is set as the binarization threshold.

FIG. 12 is a signal waveform diagram for explaining an example of the binarization threshold value set according to the present embodiment.
(A) and (B), a line TH93 connecting the luminance value (the cross mark in (A) of FIG. 12) corresponding to the zero crossing point of the second derivative signal 97 of the image signal 96 is a binary threshold value. Become. The result of binarizing the image signal 96 with the binarization threshold TH93 is the binarization signal 98 shown in FIG. On the other hand, TH94 in FIG.
Is a binarization threshold value when interpolation is not performed between trigger occurrence points. When the image signal 96 is binarized by the threshold value TH94,
A binary signal 99 shown in FIG. 12D is obtained. But,
In this case, the portion corresponding to the shaded portion 100 is not correctly binarized.

As described above, if the binarization threshold value is determined so as to interpolate between trigger occurrence points according to the present embodiment, a binarized signal in which the contour shape of the original image signal is reproduced more correctly can be obtained.

However, in the present embodiment, the threshold cannot be calculated until the trigger signal is input (the contour component of the image signal is detected), and the binarization cannot be performed. Therefore, in order to perform the binarization, the timing of the image signal to be binarized must be aligned with the timing of the calculated binarization threshold. Therefore, the image signal input to the binarization unit 4 is processed by the line buffer 82. The signal is delayed by the required time. The binarization unit 4 sets a binarization threshold necessary for binarizing the delayed image signal,
The data is sequentially read from the binarization threshold value storage buffer and binarized. The capacity of the line buffer required to delay the image signal is not easy to adjust because the interval between the trigger signals depends on the input image (that is, the two-dimensional code). Therefore, it is realistic to provide a capacity for one line.

In this embodiment, a straight line is used to interpolate between trigger signal generation points. However, it is needless to say that a curve may be used for interpolation. Further, in the present embodiment, as a means for matching the timing between the binarization threshold and the image signal to be binarized, a line buffer and a binarization threshold storage buffer are used. Alternatively, a modification may be used in which the image data is stored in the memory and the image signal corresponding to the threshold value is read out as appropriate after the binarization threshold value is determined, and the binarization is performed.

[Fifth Embodiment] Next, a fifth embodiment of the present invention will be described.
An embodiment will be described. This embodiment corresponds to the first and fifth aspects of the present invention. FIG. 13 is a block diagram showing the fifth embodiment, in which the same components as those of the first embodiment shown in FIG. The two-dimensional code read by the code reading unit 1 is input to the zero-crossing point detection unit 2 as a multi-valued image signal, and the zero-crossing point of the secondary differential signal is detected. Note that the zero-crossing point detection unit 2 includes a smoothing filter as exemplified in the second embodiment described above, or includes a smoothing filter as exemplified in the third embodiment. It is desirable to have a second differential noise removal function. From the zero-crossing point detection unit 2, a secondary differential signal generated in the process of the zero-crossing point detection process is output in addition to the trigger signal.
The signal is input to the signal synthesis unit 110. The configurations and operations of the binarization threshold value calculation unit 3 and the binarization unit 4 are the same as those in the first embodiment described above, and a detailed description thereof will be omitted.

Next, the operation of the fifth embodiment will be described with reference to signal waveform diagrams shown in FIGS. As shown in FIGS. 14A and 14B, the binarization thresholds TH110 and TH111 are calculated from the zero-crossing points of the second-order differential signal 112 obtained by secondarily differentiating the image signal 111 in the zero-crossing point detection unit 2. Is done. The binarizing section 4 performs the binarization thresholds TH110 and TH111.
Thus, the image signal 111 is binarized to obtain a binarized signal 115 shown in FIG. On the other hand, the secondary differential signal 112 is binarized by a positive binarization threshold TH112 and a negative binarization threshold TH113, and the secondary differential positive binarization signal shown in FIGS. 113, and the second derivative negative direction binarized signal 114. The second derivative positive direction binarized signal 113 is the second derivative signal 11
2 is greater than the positive binarization threshold TH112, and becomes “1”. The second derivative negative binarization signal 114 is “1” when the second derivative signal 112 is smaller than the negative binarization threshold TH113. 1 ". Next, a diagram showing a logical OR of the second derivative negative direction binarized signal 114 and the binarized signal 115
A signal 116 shown in FIG. 14 (F) (in this case, a binarized signal 115
Is directly converted into a signal 116) and the second differential positive direction binarized signal 113 is subjected to the operation shown in Table 1 to obtain a final binarized signal 118 shown in FIG. As a result, the final binarized signal 118 is converted into a binarized signal 115 by the binarizing section 4.
The portion corresponding to the hatched portion 117 of the second-order positive-direction binarized signal 113 is removed from FIG. The binarized signal 118 thus obtained is compared with the binarized signal 115 binarized by the binarizing unit 4 to obtain a binarized result closer to the contour component of the original image signal 111. Has become.

[0049]

[Table 1]

FIG. 15 is a signal waveform diagram showing a binarization process according to the present embodiment for an image signal 121 having an edge (falling edge) opposite to the original image signal 111 shown in FIG. As shown in (A) and (B) of FIG.
1 from the second derivative signal 122, the binarization thresholds TH114 and TH114
15 and the binarization unit 4 outputs a binarization signal 125 shown in FIG. Meanwhile, the figure
As shown in FIG. 15 (B), the second derivative signal 122 is compared with the positive direction binarization threshold TH116 and the negative direction binarization threshold TH117, and the second derivative signal 122 shown in FIGS. 15 (C) and (D) is obtained. A direction binarized signal 123 and a second derivative negative direction binarized signal 124 are generated. Next, the second derivative negative direction binarized signal 124 and the binarized signal
The signal 127 shown in FIG.
The same operation as in the case of the binarization processing on the original image signal 111 shown in FIG. 14 is performed between the second differential positive direction binary signal 123 and the final binary signal shown in FIG. Obtains a binarized signal 128 (in this case, the signal 127 remains unchanged as the final binarized signal 12
8). In this case, the second differential positive direction binarized signal 113 in Table 1 may be replaced with 123, the OR signal 116 with 127, and the final binary signal 118 with 128.
The final binarized signal 128 thus obtained is obtained by adding the components of the hatched portion 126 of the second-order differential negative-direction binarized signal 124 to the binarized signal 125. In the present embodiment,
As high accuracy as the binary signal obtained in the fourth embodiment of the present invention cannot be obtained. However, the processing content is simplified accordingly, and the overall configuration can be further simplified. Further, the responsiveness is superior to that of the fourth embodiment.

[0051]

As described above with reference to the embodiment, according to the first aspect of the present invention, the threshold value for binarizing the secondary differential value used for extracting the contour component of the image signal is calculated. It is possible to perform binarization faithfully on the contour of the input image by using this function.Also, since the degree of change in the luminance value of the image signal is the object of processing, it is possible to perform two-dimensional If the luminance unevenness is such that the outline of the code can be recognized, it is not affected, and the configurations of the illumination system and the optical system can be simplified. In addition, if there is information on several pixels in the vicinity of the pixel to be binarized, the binarization process can be performed sufficiently, and responsiveness approaching that of the fixed threshold type binarization method can be obtained. According to the second aspect of the present invention, the noise component superimposed on the image signal is removed by the smoothing filter so as to suppress the occurrence of unnecessary zero-crossings, so that the risk of erroneously recognizing the noise as a contour component is reduced. Can be done. According to the third aspect of the present invention, even in a case where occurrence of unnecessary zero-crossing due to a noise component superimposed on an image signal is unavoidable,
Only significant zero-crossings are extracted from the zero-crossings, so that erroneous recognition of contour components due to noise can be avoided. Claim 4
According to the invention according to the above, since the interval between the zero-crossing points is interpolated with the sampling luminance value of the corresponding image signal to determine the binarization threshold value of each pixel between the zero-crossing points, Can be switched continuously to prevent the distortion of the shape of the binarized signal. According to the invention according to claim 5, since the binarized binarized signal is corrected using the secondary differential signal, the outline buried under the influence of the switching of the binarization threshold is reproduced. Thus, it is possible to prevent the shape of the binarized signal from being distorted by a simpler configuration and processing.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of an adaptive binarization method according to the present invention.

FIG. 2 is a block diagram showing a configuration of a zero-crossing point detecting unit in the embodiment shown in FIG.

FIG. 3 is a signal waveform diagram for explaining the operation of the first embodiment shown in FIGS. 1 and 2;

FIG. 4 is a diagram illustrating an example of an operator used to obtain a second derivative signal in a zero-crossing point detection unit.

FIG. 5 is a block diagram showing a second embodiment of the present invention.

FIG. 6 is a signal waveform diagram for explaining the operation of the second embodiment.

FIG. 7 is a signal waveform diagram for explaining the operation of the second embodiment.

FIG. 8 is a block diagram illustrating a configuration of a zero-crossing point detection unit according to a third embodiment of this invention.

FIG. 9 is a signal waveform diagram for explaining the operation of the third embodiment.

FIG. 10 is a block diagram illustrating a fourth embodiment of the present invention.

FIG. 11 is a signal waveform diagram for explaining an operation of the fourth embodiment.

FIG. 12 is a signal waveform diagram for explaining an operation of the fourth embodiment.

FIG. 13 is a block diagram illustrating a fifth embodiment of the present invention.

FIG. 14 is a signal waveform diagram for explaining the operation of the fifth embodiment.

FIG. 15 is a signal waveform diagram for explaining the operation of the fifth embodiment.

FIG. 16 is a block diagram showing a configuration of a conventionally proposed adaptive binarization method.

17 is a signal waveform diagram for explaining the operation of the conventional example shown in FIG.

18 is a signal waveform diagram for describing a problem of the conventional example shown in FIG.

FIG. 19 is a signal waveform diagram for explaining a problem of the conventional example shown in FIG. 16;

[Explanation of symbols]

 Reference Signs List 1 Code reading unit 2 Zero crossing point detecting unit 3 Binary threshold calculating unit 4 Binarizing unit 5 Secondary differentiating unit 6 Sign change point detecting unit 30 Smoothing filter 51 Primary differential unit 52 Secondary differential unit 54 Forward direction 2 Value conversion unit 55 Negative direction binarization unit 56 Sign change point detection unit 57 Selector 81 Binarization threshold calculation unit 82 Line buffer 83 Binarization threshold storage buffer 110 Signal synthesis unit

Claims (5)

[Claims] A code reading section for optically reading a two-dimensional code printed on a printing medium such as paper, a second derivative of an image signal read by the code reading section, and a zero of the second derivative signal. A zero-crossing point detecting unit that detects a crossing point, and a binarization threshold calculating unit that samples a luminance value of the image signal corresponding to the zero-crossing point detected by the zero-crossing point detection unit and uses the sampled value as a binarization threshold value And a binarizing section for binarizing the image signal with a binarization threshold calculated by the binarization threshold calculating section. 2. The apparatus according to claim 1, further comprising a smoothing filter provided before said zero-crossing point detecting section for removing a noise component superimposed on an image signal read by said code reading section. Adaptive binarization method. 3. The image processing apparatus according to claim 1, wherein the zero-crossing point detection unit includes a first-order differentiating unit that performs a first-order differentiation of the image signal, and when the first-order differential value of the image signal exceeds a predetermined positive / negative threshold value, 2. The adaptive binarization method according to claim 1, wherein a zero crossing point of the differential signal is made valid. 4. The binarization threshold value calculating section samples a luminance value of the image signal corresponding to a zero-crossing point of a secondary differential signal of the image signal, and uses the sampled luminance value to determine a distance between the zero-crossing points. 2. A binarization threshold value of each pixel between the zero crossing points is determined by interpolation.
The adaptive binarization method according to. 5. When the second derivative signal of the image signal exceeds a predetermined positive / negative threshold value, the binarized signal binarized by the binarization unit is corrected using the second derivative signal. The adaptive binarization method according to claim 1, further comprising a signal synthesizing unit.