JP5299196B2 - Marker detection device and program for marker detection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To quickly detect a marker with a small processing load by identifying a predetermined marker image from a photographed image. <P>SOLUTION: The marker includes four triangles as marker elements and is configured by combining a first element pair P1 having adjacently-disposed identical triangle elements (A, A') and a second element pair P2 having different identical triangle elements (B, B') as shown in Fig.1(B). The pair P1 and the pair P2 are disposed so that a straight line C connecting the center of gravity of each triangle (A, A') of the pair P1 and a straight line D connecting the center of gravity of each triangle (B, B') of the pair P2 are intersected at a prescribed angle as shown in Fig.1(C). The marker can be detected by recognizing the elements in a simple process such as calculating a shape parameter (circularity) of a white area from a photographed image and acquiring a simple relationship between the elements. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a marker detection device and a marker detection device program for detecting a marker or an object to which a marker is attached depending on whether or not a marker image previously determined as a detection target can be recognized from an image photographed by a camera or the like.

As a method for detecting whether or not a marker defined as a detection target exists from an image taken by a video camera, a method for detecting the presence or absence of a marker attached to an object through image recognition of the marker has been conventionally known. Yes.
The conventional detection method is to make the marker shape unique and to recognize the pattern of the image using template matching. However, when pattern recognition is performed, the marker image is enlarged, reduced, rotated, tilted, or distorted. It is necessary to improve the recognition accuracy by performing geometric transformation by the above.

As described above, the conventional marker detection method based on template matching has to search for an image that matches a template having a unique shape over the entire screen. Processing load increases.
The present invention has been made in view of the above-described problems of the prior art in which a marker or an object with a marker is detected by recognizing a marker image previously determined as a detection target from the captured image. This is to enable detection of a marker with a small processing load and to perform detection quickly.

  The present invention is a marker detection device that detects a marker by recognizing a marker image that is previously determined as a detection target from a captured image, and includes a binarizing unit that binarizes the captured image, and the binarizing unit. The shape recognition means for calculating the shape parameter of the connected pixel area appearing in the binarized image processed by the above, and the first shape parameter predetermined as the detection target from the connected pixel area for which the shape parameter is calculated by the shape recognition means A first marker element candidate selecting means for selecting a connected pixel area having a close shape parameter as a candidate for a marker element; and a second shape parameter predetermined as a detection target from the connected pixel area for which the shape parameter has been calculated by the shape recognizing means. Second marker element that selects a connected pixel region having a shape parameter close to that as a candidate marker element A first pair candidate that determines, as a marker element pair candidate, a pair that satisfies a constituent condition of a marker element pair that is predetermined as a detection target from the marker element candidate group selected by the supplementary selection means and the first marker element candidate selection means A second pair candidate decision that decides, as a marker element pair candidate, a pair that satisfies a constituent condition of a marker element pair that is determined in advance as a detection target from the marker element candidate group selected by the determination unit and the second marker element candidate selection unit A combination of marker element pair candidates satisfying a marker constituent condition in which the relationship between the marker element pair candidates determined by both the first pair candidate determination unit and the second pair candidate determination unit It is characterized by having marker confirmation means for confirming as a marker to be detected.

  According to the present invention, when detecting a marker set as a detection target in advance from a captured image, marker detection can be performed quickly with a small processing load, cost performance is improved, and the field of use of marker detection is expanded. be able to.

It is a figure which shows the structural example of the marker which the marker detection apparatus of this invention detects, and the figure explaining the structural condition of the marker. It is a figure explaining the fluctuation | variation of the marker image which appears in a picked-up image. It is a block diagram showing a schematic structure concerning an embodiment of a marker detection device of the present invention. FIG. 6 is a diagram showing a part of a marker detection processing flow according to the present invention (leading to the processing flow of FIG. 5). It is a figure which shows a part of processing flow of marker detection which concerns on this invention (it is connected from the processing flow of FIG. 4). It is a figure which shows an example of the result of the intermediate process obtained by the marker detection apparatus of this invention, and a marker detection result.

Hereinafter, an embodiment according to the marker detection device of the present invention will be described.
It is based on the method of detecting the marker by recognizing the image of the marker that has been set as the detection target in advance from the captured image. This marker detection method has a large processing load, and if it involves geometric transformation, the processing load is further increased.
Therefore, in the marker detection device of the present invention, markers that are predetermined as detection targets are configured according to the following conditions.

"Marker Configuration"
In the present invention, an element constituting a marker (hereinafter referred to as “marker element”) is an element represented by a simple shape parameter, and this marker element is arranged in a predetermined relationship to constitute a marker.
As the simple shape parameter, a degree of circularity that expresses a triangle, a polygon, a circle, or the like by a constant numerical value can be used. The circularity is
Circularity = (4π × area) / (perimeter length × perimeter length)
Since it is a parameter that does not depend on the size, it is not affected by enlargement / reduction or rotation. In addition, the simple shape can reduce the burden of the process of recognizing the marker element based on the circularity calculation result and detecting the marker.
As the shape parameter of the marker element, in addition to the circularity, a filling rate and an aspect ratio may be adopted. The filling rate is the ratio of the area of the image to the area of the rectangle circumscribing the image of the recognition target (marker element), and the aspect ratio is the ratio of the length and width of the circumscribed rectangle. The filling rate and aspect ratio are also parameters that do not depend on the size of the recognition target (marker element), and thus are not affected by enlargement / reduction or rotation.

In addition, since the marker element is represented by a simple shape parameter, depending on the processing conditions, the image of the same shape parameter other than the recognition target is recognized only by the marker element, and recognition accuracy does not appear. For this reason, arranging the marker elements represented by simple shape parameters in the following predetermined relationship is a constituent condition of the mark.
That is, marker elements having the same shape parameter are paired (hereinafter referred to as “marker element pairs”), the marker elements are arranged close to each other within a predetermined distance, and two types of marker element pairs ( Marker parameters are different), and the marker element pairs are arranged in a predetermined positional relationship. By defining such a configuration condition, the marker recognition accuracy is maintained. For the predetermined positional relationship between the marker element pairs, a method is adopted in which a relative position is determined based on the position coordinates of the center of gravity of each marker element (refer to the description of FIG. 1C showing an example of the marker configuration described later).

Here, a configuration example of the marker of the present embodiment will be described with reference to the accompanying drawings. In the following embodiments described with reference to the attached drawings, an example is shown in which a triangle is used as a marker element represented by a circularity as a simple shape parameter. This is a configuration example. Yes, it is not limited to triangles.
FIG. 1 shows an example of a marker. In FIG. 1, (A) shows the configuration of the marker, and (B) and (C) are diagrams for explaining the configuration conditions of the marker.
As shown in FIG. 1A, the marker of this embodiment forms a marker by white areas of four triangles (marker elements). The four triangles (marker elements) are, as shown in FIG. 1B, a first marker element pair P1 (hereinafter simply referred to as “pair P1”) composed of marker elements (A, A ′) of the same triangle arranged in proximity to each other. And a second marker element pair P2 (hereinafter also simply referred to as “pair P2”) composed of marker elements (B, B ′) of the same triangle (a triangle different from A, A ′) arranged in proximity to each other. Are combined.
Further, as a marker configuration condition, the positional relationship between the pair P1 and the pair P2 is, as shown in FIG. 1C, the straight line C connecting the centroids of the triangles (A, A ′) of the pair P1 and the pair P2. It arrange | positions so that the straight line D which connects each gravity center of a triangle (B, B ') may cross | intersect at a predetermined angle (in the illustration, a right angle).

The marker configuration shown in FIG. 1, that is, the marker is composed of the marker element pairs P1 and P2 in accordance with predetermined conditions, so that the relationship between the marker element represented by a simple shape parameter and the simple marker element can be obtained. Even if it is adopted, the marker recognition accuracy can be maintained. In addition, when finding a marker element pair from the captured image, it is possible to give a measure of the other area or shape from the area or shape of one side, and it is also possible to give a measure of the distance from each other.
Therefore, if one marker element is obtained, the relationship with the remaining elements can be understood from the area, shape, distance, etc. in a daisy chain, and this is also an effective technique for reducing erroneous recognition.

In addition, since it is normal that a certain shooting condition cannot be obtained between the marker to be detected and the camera for shooting the marker, a difference in shooting conditions appears in the image. For example, as shown in FIG. In contrast to (A in FIG. 2), reduction ((B) in FIG. 2), inclination ((C) in FIG. 2), and distortion ((D) in FIG. 2) occur in the captured image. .
Reduction and enlargement can be dealt with by using a circularity that does not depend on the size of the marker element.
Inclination and distortion can be dealt with by lowering the reference value for determining whether or not the marker element is a circularity. In other words, even if there is inclination or distortion, one marker element pair is often found, so if one marker element pair is found, the approximate position and size can be obtained for the other, and in such cases Can be dealt with by enlarging the determination width of the marker element determined from the area and circularity with respect to surrounding marker elements.

"Configuration of marker detection device"
FIG. 3 is a block diagram illustrating a schematic configuration of the marker detection device according to the present embodiment.
As shown in FIG. 3, the marker detection device 10 includes a control unit 11, an image processing unit 12 that operates under the control of the control unit 11, a marker parameter setting unit 13, a storage unit 14, a display unit 15, an input unit 16, and Each unit of the interface unit 17 and the above-described units 11 to 17 have a bus 18 connected to exchange data with each other.
Although not shown, the control unit 11 controls the entire marker detection device and performs data processing for executing a marker detection operation or the like, but a CPU (Central Processing Unit), a program and data required for processing of the CPU A RAM (Random Access Memory) for temporarily storing and the like, and a ROM (Read Only Memory) storing a program for driving the CPU to execute operations and processes, etc. .

The image processing unit 12 is a unit that performs a process of detecting a marker from a captured image, and includes a unit that executes a detection process that will be described in detail later with reference to the process flow of FIGS. 4 and 5. The image processing unit 12 may be configured by a dedicated image processing circuit that performs marker detection processing (FIG. 3 illustrates this example), but a program for executing the marker detection processing is driven by the computer of the control unit 11. You may carry out by doing.
The marker parameter setting unit 13 is a unit that receives a processing (control) condition of a marker detection process instructed by a user from an input unit 16 (described later) and sets a processing (control) parameter in the image processing unit 12. Note that the marker parameter setting unit 13 may also be implemented using the computer of the control unit 11 as execution means.
The storage unit 14 is a unit that stores various data including image data such as a photographed image and a processed image. In addition, the storage unit 14 stores various programs necessary for marker detection processing, image processing, and the like provided to the computer of the control unit 11. To do.

The input unit 16 is used to input various data including setting data necessary for marker detection, such as instructions and operations for the marker detection device 10 or setting of the processing (control) conditions of the marker detection processing described above, for example, a mouse. Or a keyboard or a touch panel.
The display unit 15 is a means for providing a function as a user interface together with the input unit 16, such as data input through the input unit 16, a captured image by the camera 1, or a processed image processed by the image processing unit 12. It is a display means such as a monitor for displaying data, and is used when displaying the detection result of the marker on the photographed image, and when displaying the setting data at the time of input, the detection result, etc. with the character.
The interface unit 17 is a means for exchanging data with an external device, and in this embodiment, a captured image from the camera 1 that captures a marker to be detected is taken into the marker detection device 10 in this embodiment. .

"Marker detection process"
Details of the marker detection process performed by the marker detection device 10 will be described below.
The marker detection process is a process for detecting the marker described in the “marker configuration” from the photographed image of the camera 1, and is a process performed by the image processing unit 12 under the control of the control unit 11.
The outline of the marker detection process is performed according to the flow shown in the following (1) to (6).
(1) A camera image is converted into a binary image.
(2) Labeling the connected pixel areas (white area in the case of the marker in FIG. 1; hereinafter, “white area” will be described as an example)) appearing in the binarized image obtained in (1) above; Exclude too much from processing.
(3) The shape parameter (circularity) of each white area obtained in (2) above is calculated to recognize the shape of each white area.
(4) From each white region for which the shape parameter (circularity) has been calculated in (3) above, white regions close to two types of marker elements having different shape parameters determined as detection targets are selected as marker element candidates. .
(5) From the candidate group of the two types of marker elements selected in (4) above, a pair satisfying the configuration conditions (the relationship between the distance between the marker elements, the area, and the shape parameter) of the marker element pair determined as a detection target in advance Each is determined as a marker element pair candidate.
(6) Marker configuration conditions in which the relationship between the two types of marker element pair candidates determined in (5) above is determined in advance (the length of the line connecting the marker elements in each pair, the intersection and intersection of the connecting lines A combination of marker element pair candidates satisfying (angle) is determined as a marker to be detected.

“Marker detection process flow”
One embodiment of the marker detection process outlined in the above flows (1) to (6) will be described based on the process flow shown in FIGS. Note that this processing flow recognizes marker elements using circularity as a shape parameter and performs marker detection, and thus can be applied not only to the example of a marker having white triangles in FIG. For the sake of convenience, a processing example will be described with reference to the white triangle in FIG. In the description of the detection process, reference is made to FIG. 6 showing intermediate data generated in the process, its explanatory diagram, and the detection result.
The image processing unit 12 receives a photographed image of the camera 1 as a processing target image together with an instruction for marker detection processing from the control unit 11, and starts a marker detection processing flow.
When the processing is started, first, the photographed image is converted into a binary image by binarization processing in order to extract a white region of the photographed image (step S101). A threshold value of a luminance level at which a white area can be extracted is set, and a captured image signal exceeding this threshold value is obtained as a binarized image. FIG. 6A shows an image after binarization of the original captured image shown in FIG. In FIG. 9A, the extracted white area is represented by a black image.

Next, labeling is performed on the image showing the white area extracted by the binarization processing of the captured image (step S102). In this step, the image obtained by extracting the white region is divided into connected pixel regions (regions in which white pixels are connected to form one cluster), and a label for identifying the region is attached to each region and used as management information.
Next, by removing the too large and too small of the connected pixel area that can be regarded as noise, it is excluded from the processing to be performed thereafter (step S103). An excessively large one and an excessively small one to be removed can be determined in advance from the size of the marker to be detected in the captured image.
Next, a process of recognizing the shape is performed on the connected pixel area after removing the noise (step S104). This process is a process of calculating a circularity as a shape parameter, and performs a calculation of circularity = (4π × area) / (perimeter length × perimeter length). In this embodiment, since the marker to be detected has a triangle as an element, the circularity calculated from the marker element obtained from the connected pixel area on the photographed image is included in a substantially constant numerical value range. In FIG. 6B, the calculated numerical value is written in addition to the connected pixel region where the circularity is calculated (note that the numerical value in the figure is not a circularity itself but a converted value). This numerical value becomes larger as it is closer to a circle, and in the case of a triangle, it becomes larger as it is closer to a regular triangle.

  After calculating the circularity as the shape parameter of the connected pixel region, based on the calculation result, two types of marker elements that are predetermined as detection targets, that is, pairs of marker elements that are distinguished by the circularity (FIG. 1 ( Regions close to the circularity used for the pair P1 (A, A ′) and the pair P2 (B, B ′) in C) are selected as marker element candidates (steps S111 and S121). Also in the case of configuring a pair having triangles having different shapes as marker elements, the triangles can be distinguished by the circularity, so that the pair P1 (A, A ′) and the pair P2 (B, B ′) in FIG. ) A candidate group of marker elements is selected according to a circularity range determined corresponding to each. In the example of FIG. 6B, the marker elements that are candidates for the pair P1 (A, A ′) are approximately 50 to 80, and the marker elements that are candidates for the pair P2 (B, B ′) are approximately 30 to 50. , And a candidate group of marker elements for the pair P1 and the pair P2 is selected according to this numerical range.

After selecting a candidate group of marker elements of the pair P1 (A, A ′) in step S111, processing for determining a candidate for the pair P1 is performed. In this process, it is determined whether the two elements extracted from the selected marker element candidate group are pairs that satisfy the constituent conditions of the marker element pair P1 (A, A ′) determined in advance, A pair that satisfies the configuration condition is determined as a candidate for the pair P1.
In the pair P1 candidate determination procedure, first, one element is selected as a candidate A from the candidate group of marker elements (step S112), and the candidate A constituting the pair P1 (A, A ′) with the selected candidate A is selected. As a first constituent condition between ', it is checked whether the distance between candidates is close, that is, whether there is a candidate A' in the distance range estimated from the area of the marker element of candidate A (step S113). The distance between the marker elements obtained at this time is obtained by calculating the distance between the gravity center positions of the elements. In addition, the distance range is estimated from the area, but this may be incorrect if the marker configuration conditions determined in advance as detection targets are applied as they are to the actually captured image. This is because the correction amount for obtaining the result is estimated from the area.

If there is no candidate A ′ within the distance range in step S113 (step S113-NO), the configuration condition of the pair P1 (A, A ′) is not satisfied, so the process returns to step S112 to select candidate A from the marker element candidate group. Then, the next element in the order is selected, and it is checked whether or not the pair P1 is configured by the same procedure as that for the previous element.
On the other hand, if there is a candidate A ′ within the distance range in step S113 (step S113—YES), then, as a second constituent condition of this pair P1 (A, A ′), the area between the marker element candidates Is close, that is, whether or not the difference between the two areas is within a predetermined range (step S114).
Here, if the difference between the two areas is out of the predetermined range (step S114-NO), the second constituent condition of the pair P1 (A, A ′) is not satisfied, so the process returns to step S112, and the marker element Next, an element in the next order is selected as a candidate A from the candidate group, and it is examined whether or not the pair P1 is configured by the same procedure as that for the previous element.

On the other hand, if the difference between the two areas is within the predetermined range (step S114-YES), then, as a third constituent condition of the pair P1 (A, A ′), the shape parameter between the marker element candidates It is checked whether or not the circularity is close, that is, whether or not the difference between the two is within a predetermined range (step S115).
Here, when the difference in the circularity between the two is outside the predetermined range (step S115-NO), the third constituent condition of the pair P1 (A, A ′) is not satisfied, so the process returns to step S112, and the marker The element in the next order is selected as a candidate A from the element candidate group, and it is examined whether or not the pair P1 is configured by the same procedure as that for the previous element.

On the other hand, if the difference in the circularity between the two is within the predetermined range (step S115-YES), a pair that satisfies the configuration condition of the pair P1 (A, A ′) is determined as a candidate for the pair P1 (step S116). . That is, the pair P1 (A, A ′) determined at this time satisfies all the first to third configuration conditions, and is determined as a candidate for the pair P1 if none of the first to third configuration conditions is satisfied. do not do. In the example of FIG. 6B, two pairs having circularity of (79, 70) (72, 68) are determined as candidates for the pair P1 (A, A ′) as shown in FIG. ing.
Thereafter, it is confirmed whether or not the processing for checking the constituent conditions of the pair P1 (A, A ′) is completed with all the candidate groups of marker elements of the pair P1 (A, A ′) as candidates A (step S117). ). If this check process is not completed (step S117-NO), there is an unprocessed element, so the process returns to step S112, and the next element is selected as a candidate A from the marker element candidate group. Then, it is checked whether or not the pair P1 is configured by the same procedure as that for the previous element.
On the other hand, if the processing for checking the configuration condition of the pair P1 (A, A ′) is completed (step S117—YES), the next intersection check (the relationship between the pair P1 and the pair P2 is determined in advance as a detection target). To check whether the marker configuration condition is satisfied).

The steps S111 to S117 are procedures for determining a candidate for the pair P1, but the procedure for determining a candidate for the pair P2 that is the other marker element pair is the same as that for the pair P1. That is, steps S121 to S127 corresponding to steps S111 to S117, which are procedures for determining a pair P1 candidate, are performed as a procedure for determining a pair P2 candidate in parallel. In the example of FIG. 6B, one set having a circularity of (33, 40) is determined as a candidate for the pair P2 (B, B ′).
It should be noted that the description of the procedure for determining the candidate for the pair P2 refers to the description of the procedure for determining the candidate for the previous pair P1, and the description is omitted here.
However, when both the process for checking the configuration condition of the pair P1 (A, A ′) and the process for checking the configuration condition of the pair P2 (B, B ′) are completed, the process proceeds to the next intersection check. Since it is a condition, after confirming the completion of both processes, the process proceeds to the intersection check process flow (FIG. 5).

In the intersection check processing flow (FIG. 5), it is checked whether or not the relationship between the pair P1 and the pair P2 satisfies a marker configuration condition that is set in advance as a detection target.
First, one set is selected from each pair determined as a candidate for each of pair P1 and pair P2 in the preceding process (FIG. 4) (step S131).
Next, the centroid of each of the selected marker elements A and A ′ of the pair P1 is obtained, and a straight line connecting the obtained centroids is obtained (step S132). A straight line connecting the centers of gravity is a straight line C shown in FIG. 1C, which shows an example in which the marker elements A and A ′ are triangular.
Similarly to the pair P1, the center of gravity of each of the selected marker elements B and B ′ of the pair P2 is obtained, and a straight line connecting the obtained center of gravity is obtained (step S133). A straight line connecting the centroids is a straight line D shown in FIG. 1C, which shows an example in which the marker elements B and B ′ are triangular.

As a marker configuration condition, a predetermined relationship between the pair P1 and the pair P2 is that a straight line connecting the centroids obtained in the previous stage intersects and the intersection angle is within a predetermined range as a detection target. .
Therefore, next, it is checked whether or not the straight line connecting the centroids obtained in step S132 and the straight line connecting the centroids obtained in step S133 intersect (step S134).
Here, if there is no intersection between the straight lines (step S134-NO), the process returns to step S131 to check the combination of the next pair P1 and pair P2.
On the other hand, if there is an intersection between straight lines (step S134-YES), it is checked whether the intersection angle, which is another constituent condition, is within an allowable range that can be recognized as a range that has been previously determined as a detection target (step S134). S135).
If the intersection angle is not within the allowable range (step S135-NO), the process returns to step S131 to check the next combination of the pair P1 and the pair P2.

  On the other hand, if the intersection angle is within the allowable range (step S135: YES), the combination of the pair P1 and the pair P2 satisfies the marker constituent condition, and is thus determined as a marker to be detected (step S136). Data of the connected pixel area, which is an element constituting the determined marker, is stored in the storage unit 14 as detected marker information. FIG. 6C is an example showing a marker detection result. In this example, the detected marker portion is surrounded by a line on the screen of the display unit 15 that displays the original photographed image, and the above detection result can be used for such display. Note that FIG. 6C illustrates an example of use. The present invention is not limited to this example, and an existing display method capable of identification display can be employed. Further, data obtained as a detection result may be used in another control system.

After determining the pair P1 and pair P2 as markers in step S136, it is confirmed whether or not all combinations of the pairs determined as candidates for the pair P1 and pair P2 have been checked (step S137). If there is a combination (step S137-NO), the process returns to step S131 to check the combination of the next pair P1 and pair P2.
On the other hand, if all the combinations of the candidates of the pair P1 and the pair P2 have been checked (step S137-YES), it is confirmed whether all the marker pairs have been checked (step S138). This step is a procedure corresponding to the case of detecting a plurality of markers.
If there is an unchecked marker (step S138-NO), the process returns to step S131 to select a candidate for this marker pair.
For the new marker, the detection process in steps S131 to S137 is performed in the same manner as in the previous marker, and if it is confirmed that all the marker pairs have been checked (step S138-YES), the marker detection process flow ends. .

  DESCRIPTION OF SYMBOLS 1 ... Camera 10 ... Marker detection apparatus 11 ... Control part 12 ... Image processing part 13 ... Marker parameter setting part 14 ... Storage part 15 ... Display part 16 ... Input part , 17 · Interface part, 18 · Bus.

Claims (5)

A marker detection device that detects a marker by recognizing an image of a marker that is determined in advance from a captured image,
Binarization means for binarizing the captured image;
Shape recognition means for calculating shape parameters of connected pixel regions appearing in the binarized image processed by the binarization means;
First marker element candidate selection means for selecting, as a marker element candidate, a connected pixel area having a shape parameter close to the first shape parameter determined in advance from the connected pixel area for which the shape parameter is calculated by the shape recognition means; ,
Second marker element candidate selecting means for selecting, as a marker element candidate, a connected pixel area having a shape parameter close to a second shape parameter determined in advance from the connected pixel area whose shape parameter is calculated by the shape recognition means; ,
First pair candidate determination means for determining, as a marker element pair candidate, a pair that satisfies a constituent condition of a marker element pair previously determined as a detection target from the marker element candidate group selected by the first marker element candidate selection means;
Second pair candidate determining means for determining, as marker element pair candidates, a pair that satisfies the constituent conditions of a marker element pair that has been previously determined as a detection target from the marker element candidate group selected by the second marker element candidate selecting means;
A combination of marker element pair candidates satisfying a marker configuration condition in which a relationship between marker element pair candidates determined by both the first pair candidate determining unit and the second pair candidate determining unit is set as a detection target in advance is detected. A marker detection device comprising marker determination means for determining as a marker. In the marker detection device according to claim 1,
Each of the first and second pair candidate determination means is within a distance range in which the marker element candidates are determined in advance as detection targets, the difference in area between both is in a range determined in advance as detection targets, and both A marker detection device characterized in that a difference in shape parameters is within a range that is determined in advance as a detection target, as a pair candidate determination condition. In the marker detection device according to claim 1 or 2,
The marker determining means intersects a line connecting the centroids of marker elements constituting one of both marker element pair candidates and a line connecting the centroids of marker elements constituting the other, and the intersection angle is set as a detection target in advance. A marker detection apparatus characterized in that a marker constitution condition is that it is within a defined range. In the marker detection device according to any one of claims 1 to 3,
The marker detection device, wherein the shape parameter is circularity.   The program for functioning the computer of a marker detection apparatus as each means of the marker detection apparatus described in any one of Claims 1 thru | or 4.

Images