JPH11183159A - Image processor and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor for reading the tag information of a baggage moving on a belt conveyer. SOLUTION: This processor 10 is provided with a shape measuring part 2 for processing the images of a baggage 7 picked up by cameras 11, 12, and obtaining the focus information necessary for picking up the image of the baggage 7 in focus, a camera control part 3 for controlling the focus of a line scan camera 6 by using the measured focus information, an in-focus image input part 4 for inputting in image for which the focus is adjusted by the camera control part 3, and a baggage tag reading part 5 for reading the information recorded on the baggage tag from the image inputted in the image input part 4.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus for reading a package tag and a method thereof.

[0002]

2. Description of the Related Art In the transportation business and the postal business, packages and mail are once collected at a collection and distribution center, where the packages are sorted by destination and sent to another collection and distribution center. The sorting work at this collection and distribution center requires a large amount of luggage and requires a short period of work, so automation is required.

[0003] Among them, an apparatus for reading a postal code from a mail conveyed by a belt conveyor or a conveyer and automatically sorting mail having a small thickness such as an envelope or a postcard has been developed.

[0004] Further, for thick luggage such as a cardboard box, a bar code is printed on the luggage, a sticker on which the bar code is printed is attached, and the bar code is read from the luggage moving on the belt conveyor at the collection and delivery center. Automatic sorting devices have been developed.

[0005] However, the destination classification based on the barcode requires that the destination information of a package to be sent out by a business operator be previously stored in a database or the like, and this method cannot be applied to general packages.

[0006] There are also some problems in reading information on a tag by using a method used for a mail sorting machine for envelopes and postcards in a cardboard box. Postal sorters use a contact-type scanner that is used for faxing, etc. to read the postal code.However, for general luggage such as cardboard boxes, use a contact-type scanner because the shape of the luggage is not constant. Can not. The contact type scanner can be applied if the side where the tag of the package is attached is down, but generally the tag is attached to the top of the package, so most of the package is facing downward with the tag attached. The operation of transposing the worker so as to be performed occurs. In addition, this method cannot be applied because some items are prohibited from being vertically transposed.

One of the means for reading the information on the tag attached to the upper surface without transposing the package is a TV camera.
When the image is shot at W640 × H480, which is the resolution of the TSC camera, the image is shot at 1 pixel / mm, and the resolution is insufficient to read the information on the tag.

In such a case, it is possible to take an image with a sufficient resolution by using a line scan camera. A line scan camera can capture an image at a high resolution by using a one-dimensional camera using an image sensor having a large number of pixels.

If a luggage is placed on a belt conveyor or the like and is moved under the line scan camera with the line scan camera photographing from above, the tag attached to the luggage upper surface can be photographed with sufficient resolution. .

This method also has a problem. In a normal TV camera, the luggage enters the field of view of the luggage before passing under the camera. At that point, the focus can be adjusted and an image focused on the upper surface of the luggage can be taken. In the case of (1), since the image is taken only on one scanning line, the baggage will enter the camera's field of view only when the baggage comes under the camera. Therefore, even if the focus adjustment is started, there is a possibility that the adjustment cannot be completed before the tag to be photographed enters the camera view.

When a tag is stuck on an oblique plane such as the carrying bag 100 in FIG. 5 or a tag is stuck on a curved surface with cylindrical baggage or the like, the baggage moves under the camera. Therefore, since the focal length changes, there is a problem that even if focus adjustment is performed with only one scanning line image captured by a line scan camera, defocus occurs.

[0012]

As described above, the conventional image processing apparatus focuses on a tag attached to a thick object such as a cardboard box or a carrying bag with sufficient accuracy when photographing and processing the image. There is a problem that it is not possible to input and process an image that fits.

Accordingly, the present invention provides an image processing apparatus and method for reading tag information of a package.

[0014]

According to a first aspect of the present invention, there is provided an image input means for inputting an image of a load, a focused image input means for inputting an image obtained by adjusting a focus on the load, and the image inputting means. From the image photographed by the means, a shape measuring means for obtaining focus information necessary for photographing the luggage in an in-focus state by the in-focus image input means, and using the focus information measured by the shape measuring means. An image processing apparatus comprising: focus control means for controlling the focus of the focused image input means; and a tag reading means for reading information described on the tag of the package from the image input by the focused image input means. is there.

An image processing apparatus according to the first aspect of the present invention will be described.

The image input means inputs an image of the package.

The in-focus image input means inputs an image in which the focus is adjusted for the package.

The shape measuring means obtains, from the image photographed by the image input means, focus information necessary for photographing the package in the focused state by the focused image input means.

The focus control means controls the focus of the focused image input means using the focus information measured by the shape measurement means.

The tag reading means reads information written on the tag of the package from the image input by the in-focus image input means.

As described above, the image photographed by the image input means is processed to obtain the information necessary for photographing the baggage in a focused state. Since the focused image is input by the input means, the information written on the tag can be accurately read from the tag attached to the package. The invention of claim 2 is
The shape measuring means measures a three-dimensional shape of the package,
The measured three-dimensional shape information is sent to the focus control unit as focus information, and the focus control unit determines a distance between the package and the focused image input unit based on the three-dimensional shape information from the shape measurement unit. 2. The image processing apparatus according to claim 1, wherein the focal point of the in-focus image input unit is adjusted based on the distance.

An image processing apparatus according to a second aspect of the present invention will be described.

The shape measuring means measures the three-dimensional shape of the package, and sends the measured three-dimensional shape information to the focus control means as focus information.

The focus control means calculates a distance between the package and the focused image input means based on the three-dimensional shape information from the shape measuring means, and calculates the focused image input means based on the distance. Adjust the focus.

According to a third aspect of the present invention, the load is conveyed by a belt conveyor, the image input means is arranged at a first point on the belt conveyor, and a second position after the first point on the belt conveyor. 2. The image processing apparatus according to claim 1, wherein a line scan camera as said focused image input means is arranged at a point.

An image processing apparatus according to a third aspect of the present invention will be described.

The luggage is conveyed by a belt conveyor, the image input means is disposed at a first point on the belt conveyor, and the focused image input is performed at a second point after the first point on the belt conveyor. A line scan camera as a means is provided. As a result, it is possible to accurately read the tag of the baggage being conveyed by the belt conveyor.

According to a fourth aspect of the present invention, there is provided an image inputting step of inputting an image of a package, a focused image inputting step of inputting an image focused on the package by a tag reading camera, and the image inputting step. From the image captured in the above, using a shape measurement step to obtain focus information necessary to shoot the luggage in focus with the tag reading camera, and using the focus information measured in the shape measurement step An image processing method comprising: a focus control step of controlling a focus of a tag reading camera; and a tag reading step of reading information described on a tag of the package from the image input in the focused image inputting step. Is the way.

According to a fifth aspect of the present invention, in the shape measuring step, the three-dimensional shape of the package is measured, and the measured three-dimensional shape information is sent to the focus control step as focus information. A method of calculating a distance between the package and the tag reading camera based on the three-dimensional shape information from the shape measuring step, and adjusting a focus of the tag reading camera based on the distance. An image processing method according to item 4.

According to a sixth aspect of the present invention, the baggage is conveyed by a belt conveyor, and an image input camera used in the image input step is arranged at a first point of the belt conveyor, and the first bag of the belt conveyor is provided. 5. The image processing method according to claim 4, wherein a line scan camera serving as the tag reading camera is arranged at a second point after the point.

[0031]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment A first embodiment of the present invention will be described below with reference to FIGS.

FIG. 1 is a configuration diagram of an image processing apparatus 10 according to the present embodiment.

The belt conveyor 1 is a device that translates at a constant speed when a load 7 to be photographed is placed thereon.

The camera 11 and the camera 12 are mounted on and beside the belt conveyor 1, and photograph the luggage 7 moving on the belt conveyor 1.

The shape measuring unit 2 measures the three-dimensional shape of the package 7 photographed by the cameras 11 and 12, and outputs the measurement result to the camera control unit 3.

The camera control unit 3 receives the shape measurement result of the object, generates camera control information from the result, and controls the camera so that an image always focused on the object can be taken.

The in-focus image input unit 4 includes the belt conveyor 1
Above the camera, an image is input using the cameras 11 and 12 controlled by the camera control unit 3.

The tag reading section 5 processes the image input from the focused image input section 4, reads information such as a destination, and outputs the information.

(Shape Measuring Unit 2) FIG. 2 is a configuration diagram showing an embodiment of the shape measuring unit 2.

The camera 11 is located above and beside the belt conveyor 1.
And the camera 12. The photo switch 13 is provided at a fixed position on the belt conveyor 1.

The baggage 7 is conveyed on the belt conveyor 1 at an appropriate interval. The photo switch 13 detects that the baggage 7 has arrived at a predetermined position on the conveyor 1 and sends it to the stereo camera controller 20. Send a signal.

Upon receiving the signal from the photo switch 13, the stereo camera controller 20 immediately captures an image with the cameras 11 and 12, digitizes the image, and inputs the digitized image.

The shape measurement control section 21 inputs a digital image of the package 7 and measures the shape of the package 7 from this image.

Hereinafter, this procedure will be described with reference to the flowchart of FIG.

First, a difference process is performed to obtain a difference for each pixel between an input image and an image which has been stored in advance and is taken without the baggage 7 on the belt conveyor 1, and a difference image is created. Basically, the difference is large in the part of the baggage 7 in the image, and the difference is very small in the area other than the baggage 7 (step 101).

Threshold processing is performed on the absolute value of each pixel of the difference image, and a binary image is created with 1 if the absolute value of the pixel is larger than the threshold and 0 if the absolute value of the pixel is smaller than the threshold (step 10).
2).

Next, labeling of the connected area of the binary image is performed. In the labeling process, the values of eight pixels adjacent to the pixel having the pixel value of 1 in the binary image are checked. If there is a pixel having the pixel value of 1 out of the eight pixels, the value of the pixel adjacent to the original pixel is determined. A process of adding the same label number to both is performed. As a result, an independent label number is added to the connection area (step 103).

The area (the number of pixels) is calculated for each labeled connected area, and the area having the largest area is selected. The area having the largest area is the area representing the package 7 (step 104).

Next, a circumscribed polygon (convex hull) of the package 7 area is calculated. The procedure for calculating the convex hull of a two-dimensional region is described in, for example, “Introduction to Computational Geometry” prepared by Shamoth,
Translation is performed in accordance with the procedure shown on page 108 of Soken Publishing, translated by Takao Asano and Tetsuo Asano (step 105).

The processing of steps 101 to 105 is
The processing is performed on two images captured by the cameras 11 and 12, and a convex hull of the package 7 area captured from above and the package 7 area captured from the side is obtained.

Next, the polygon of the convex hull is subjected to the vertex reduction processing described below to determine the shape of the package 7.

First, the method of vertex reduction will be described.

As shown in FIG. 4, a vertex P which is a part of a polygon is
It is assumed that there are 1, P2, P3, P4, and P5 and are arranged counterclockwise from the center of gravity of the area.

At this time, the method for reducing the vertex P3 is as follows.
There is a street. One method is to simply reduce the vertex P3 to reduce the polygon. The other is a method using a polygon having a side circumscribing at a vertex P3.

The intersection of the straight line passing through the vertex P3 and the straight line passing through P1 P2 is P'2, and the intersection of the straight line passing through the vertex P3 and the straight line passing through P5 P4 is P'4. A straight line passing through the vertex P3 that minimizes the sum of the areas of the triangles P2 P3 P'2 and P4 P3 P'4 is selected, the intersection of this straight line and the straight line P1 P2 is P "2, and the intersection of the straight line P5 P4 is P "4. Triangle P2 at this time
Let S be the sum of the areas of P3 P "2 and the triangle P4 P3 P" 4.
In the triangle P2, P3, and P4, the number (area) of pixels that do not belong to the package 7 area is S '. The smaller value of S and S 'is defined as the evaluation value S ^* of the vertex P3.

When actually reducing the vertices, if S <S ', P1 P2 P4 P5 is set as a part of a new polygon, and S>
If it is S ', P1 P "2P" 4P5 is set as a part of a new polygon.

Assuming that the number of vertices of the convex hull is n, an evaluation value S ^* when all the vertices of the polygon having n points are reduced is calculated, and the minimum value of S ^{* and} the minimum value of S ^* are calculated. A vertex to be taken is obtained (step 106).

By reducing vertices taking the minimum S ^* , n-
A polygon consisting of one point is created (step 107).

The processing from step 106 to step 107 is repeated until the number of vertices becomes four (step 1
08).

By this vertex reduction processing, for example, FIG.
Is removed from the luggage 7 in the bag 100 as shown in FIG. Further, even when a non-box-shaped baggage 7 such as a golf bag is photographed, it is possible to measure a shape in which a quadrilateral surrounding the baggage 7 can be calculated.

Based on the calculated positions of the four vertices, information on the positional relationship between the cameras, the focal length of the cameras, and the like, when the baggage 7 is assumed to be a rectangular parallelepiped, whether the four vertices can be seen in the image taken by the horizontal camera 12 or not? Calculate whether vertices are visible. A coordinate system that places the origin on the plane of the belt conveyor 1
Is at coordinates (0, 0, Z1), the focal length is f1, the optical axis is oriented at (0, 0, -1), and the y-axis of the imaging surface is (0, 1,.
0), the camera 12 moves the coordinates (0, Y2,
0), the focal length is f2, and the optical axis is (0, -1, 0)
, And the y-axis of the imaging surface faces (0, 0, 1). In the image taken by the camera 11, the coordinates (hi, vi
) Is on the straight line represented by the following equation 1.

[0062]

(Equation 1) For each vertex, the x coordinate of the projection on the camera 12 imaging plane is calculated, and the projection of the foremost vertex A (the tentative coordinate y becomes maximum) as viewed from the camera 12 on the camera 12 imaging plane is calculated. Compared to the projections of the other vertices, if the x coordinate of the projection of vertex A is the minimum or maximum of the four vertices,
In the subsequent step 110, the number of vertices to be reduced is set to 4, otherwise, it is set to 6 (step 109).

According to the result determined in the above step, the vertex reduction processing is repeatedly performed on the image photographed by the horizontal camera 12 until the number of vertices becomes 4 or 6 (step 110).

Next, the coordinates of the vertices photographed by the camera 11 and the coordinates of the vertices photographed by the camera 12 are correlated, and the three-dimensional coordinates of the vertices are measured by stereo measurement. If the number of vertices obtained in step 109 is 4, the projection of the four vertices photographed by the camera 11 onto the imaging surface of the camera 12 is calculated,
The point where the x coordinate of the projection is the maximum is defined as the corresponding point in the image of the camera 11. For vertices taken by camera 12, x
The vertex having the largest coordinate and the vertex having the second largest x coordinate are selected, and the one having the larger y coordinate is selected as the corresponding point.

If the result of step 109 is 6, the vertex in the image photographed by the camera 11 having the maximum y of the provisional coordinates is the corresponding point, and the y coordinate of the 6 vertices photographed by the camera 12 is The point that becomes the maximum is the corresponding point.

Assume that the position of the associated vertex in the image taken by the camera 11 is (hi, vi), and the position in the image taken by the camera 12 is (hj, vj). The point at which the position in the image taken by the camera 12 is (hj, vj) is on the straight line represented by the following equation (2).

[0067]

(Equation 2) The three-dimensional position of the corresponding vertex can be obtained from the intersection of the straight line represented by Expression 2 and Expression 1 (Step 111).

As for the remaining three vertices of the vertices photographed by the camera 11, the three-dimensional position is obtained from Expression 1 on the assumption that the Z coordinate is the same as the vertex obtained in step 111. From the three-dimensional position of the four vertices, the length, width,
The height value is obtained and output (step 112).

(Camera Control Unit 3) The camera control unit 3 controls the camera using the information input from the shape measuring unit 2 so that an image focused on the baggage 7 can be taken.

The line scan camera 6 of the in-focus image input unit 4 adjusts the focus of the line scan camera 6 because the line scan camera 6 is located right above the belt conveyor 1, that is, right above the baggage 7, as shown in FIG. Focus information
Distance M between the top surface of package 7 and line scan camera 6
Becomes

For this purpose, the camera control unit 3 first extracts height information h from the three-dimensional shape information of the package 7 measured by the shape measuring unit 2. Next, the height information h is converted into focus information with reference to a table storing calibration data registered in advance. Specifically, since the distance L between the line scan camera 6 and the belt conveyor 1 is stored in the table, the focus information M is obtained by subtracting the height information h from the distance L.

Then, according to the focus information M,
The motor attached to the focus adjustment ring of the line scan camera 6 is driven to adjust the focus.

(Focused Image Input Unit 4) The focused image input unit 4 inputs an image from the line scan camera 6 using the lens whose focus has been adjusted by the camera control unit 3, and reads the input image into the tag reading unit 5. Transfer to At this time, a high-precision image can be input by photographing using the line scan camera 6.

(Package Tag Reading Unit 5) Next, the package tag reading unit 5 will be described. FIG. 7 is a configuration diagram of the tag reading unit 5.

The tag reading unit 5 reads a tag detection unit 31 for detecting the position of the tag from the input image, and reads information such as destination information necessary for operations of the transportation business such as sorting from the tag at the detected position. It is composed of a tag information processing unit 32.

The tag detecting section 31 detects a tag by comparing the input image with dictionary information recorded in advance.

Hereinafter, the tag detection processing will be described with reference to the flowchart of FIG.

First, characteristic points (corner points) are extracted from the input image. The extraction of corner points is described in, for example, "Active Vision" by Andrew Blake, Alan Yuille, MIT
The method described on page 265 of Press Publication is used (step 201).

Next, the direction of the extracted corner is determined.
For each corner point, a convolution operation of a Gaussian derivative centered on the corner point is performed, and from the result of the convolution, the corner direction θ is obtained by the following equation (step 20).
2).

[0080]

(Equation 3) A template near the detected corner point is cut out from the image. The image is rotated around the extracted corner point so that the corner direction matches the x-axis. A small area within a predetermined distance r from the corner point is extracted from the rotated image as a template (step 20).
3).

All the templates extracted from the image are collated with the dictionary information (step 204).

In the image processing apparatus 10 of the present embodiment, the number of tags to be processed is determined to be a plurality of finite types, and the image information of all the tags is obtained in advance.
It is unknown which tag is attached to which surface of the package 7 and the shape and size of the package 7 are undefined. Therefore, the position and orientation of the tag in the captured image is unknown. The size of the tag in the image also changes according to the size and position of the package 7 (according to the distance between the camera and the tag). If the tag is attached to a surface that is obliquely viewed from the camera, the tag in the image is deformed. When the tag is affixed to a plane, the above-mentioned changes in position, orientation, size, etc. can be approximated by affine transformation of the image shown below. (X, y) is the coordinates of the pixel before the conversion, and (x ', y') are the coordinates of the pixel after the conversion.

[0083]

(Equation 4) In the image processing apparatus 10 of the present embodiment, the affine transformation parameters of the tag image are unknown, but a change range of each parameter can be limited.

For example, the maximum size of the baggage 7 to be processed by the image processing apparatus 10 of the present embodiment is determined in advance by the regulations of a transportation company or the like. You can ask for it. The tag attached to the rectangular parallelepiped luggage 7 of the maximum size is the largest tag attached to the camera when the camera optical axis and the center of the tag attached to the plane are orthogonal to each other. A photograph of a package 7 having almost no thickness such as an envelope is a minimum-sized package tag.

Here, the dictionary information used for matching will be described first.

In the dictionary information, an image of a tag of the type to be processed is taken in advance, corner points are extracted, and several points are selected from the corner points. The template is cut out with the selected point as the center. As described above, the tag photographed by the camera is transformed from an image photographed for creating dictionary information by affine transformation. Each parameter of the affine transformation at this time is within the parameter change range. Therefore, when creating a dictionary, each parameter is changed at regular intervals within the change range, and a template is cut out from the affine-transformed image using the parameter. Thus, in the dictionary information, a plurality of templates are prepared for one corner point.

In the collation between the template extracted from the input image and the dictionary information, the cross-correlation between the template of the input image and the template of the dictionary information is calculated, and the maximum value of the correlation value is used as the similarity of the template. In addition, the dictionary information template having the highest correlation value determines which corner point of the dictionary information is affine-transformed with which parameter.

Next, the similarity of each template is compared with a predetermined threshold value, and only templates having a similarity value equal to or greater than the threshold value are selected (step 205).

The number of templates selected in step 205 is assumed to be n and the i-th template. The three combinations are sequentially examined from the selected template, and the following processing is performed.

First, an affine transformation parameter for converting the coordinate system of the dictionary to the coordinate system of the image is obtained from the center coordinates of the selected three templates and the center coordinates of the template of the dictionary information collated by each template. The center coordinates of the template extracted from the image are represented by (Xi, Yi
), The center coordinates of the dictionary template are (xi, yi). The translation component of the affine transformation (e, f in Equation 4) can be obtained by the following Equation 5 from the movement of the center of gravity.

[0091]

(Equation 5) For the remaining parameters a to d,

(Equation 6) If the obtained affine transformation parameters are within the parameter change range, if not, the subsequent combination is selected without performing the subsequent processing (step 2).
06).

For the combination selected in step 206, the affine transformation parameters calculated in step 206 are compared with the affine transformation parameters of the collated dictionary template.

For each of the parameters a to d other than the translation, the sum of the absolute values of the differences between the parameters calculated in step 206 and the parameters of the three sets of dictionary templates is calculated and used as the evaluation value of the combination of the templates. (Step 207).

The evaluation values in step 207 are calculated for all the template combinations selected in step 206, and the combination having the highest evaluation value is determined to be the template combination representing the detected tag. If the highest evaluation value is lower than a predetermined threshold, it is determined that there is no tag.

In the tag information processing section 32, the tag detecting section 31
The information described on the tag detected in step (1) is read, converted into information that can be processed by a computer, and recorded.

Calculation of the affine transformation parameter for mapping the template in the image obtained from the tag detection unit 31 and the template in the image from the template of the dictionary information to the position of the template of the dictionary information as an inverse matrix of Equation 6. Can be. When this affine transformation is applied to the input image, the tag in the input image is copied to the position of the tag in the image used in the dictionary information, and the position, orientation, size change, and shape of the tag in the input image are changed. Can be normalized.

After normalizing the image,
An area in which character information and code information such as a destination, a telephone number, a slip number of a tag, and a store arrival code are written is cut out. The position of the character or code is specified in advance in the position of the tag image used in the dictionary when creating the dictionary information, and this information is stored and used together with the dictionary information.

After cutting out areas such as characters and codes,
A character recognition process and a barcode recognition process are performed on the image in this area to obtain character information and code information.

The tag information processing section 32 displays the obtained character information and code information on the display device of the terminal.
Communication of the information through a wired or wireless communication medium, recording on a recording medium, and the like are performed. Also, information is recorded or rewritten on a semiconductor recording medium attached to the package 7 and readable and writable a plurality of times.

Second Embodiment Next, a second embodiment according to the present invention will be described.

The system configuration of the image processing apparatus 10 of this embodiment has the same configuration as that of the first embodiment shown in FIG.
The internal configurations of the shape measurement unit 2 and the camera control unit 3 are different.

(Shape Measuring Unit 2) First, the configuration of the shape measuring unit 2 according to the present embodiment will be described with reference to FIG.

The camera 11 is located above and beside the belt conveyor 1.
And the camera 12. The photo switch 13 is provided at a fixed position on the belt conveyor 1.

The load 7 is conveyed on the belt conveyor 1 at an appropriate interval. The photo switch 13 detects that the load 7 has reached a predetermined position on the belt conveyor 1, and
A signal is sent to the stereo camera control unit 340.

Upon receiving the signal from the photo switch 13, the stereo camera controller 340 immediately captures an image with the cameras 11 and 12, digitizes the image, and inputs the digitized image.

The shape measurement control section 41 inputs a digital image of the package 7 and measures the shape of the package 7 from this image. The shape measurement control unit 41 detects the baggage 7 area from an image of the baggage 7 captured from above by the camera 11 and uses the same method as the tag detection unit 31 in the same manner as the shape measurement control unit 21. The position of the tag is detected from the image taken by the camera 11.

Since the shape and position of the tag in the image used to create the dictionary information are known, the position of the tag in the input image can be calculated from the position of the tag used in the dictionary information.

The positions of the four corners of the tag in the image are represented by (xi, y
i), the length of the long side of the tag is l1, and the length of the short side is l2
And The focal length of the camera 11 is f3. The camera coordinate system and the three-dimensional coordinate system Pi = (Xi, Yi, Zi) are shown in FIG.
0, and the three-dimensional coordinates correspond to the camera coordinates (xi, y
i) is projected to

(Equation 7) Holds. P1 P2 and P3 P4 are P1 P
4, if P2 P3 is the short side,

(Equation 8) Holds. Furthermore, since each side is orthogonal,

(Equation 9) Holds. By solving equations 7, 8, and 9, the three-dimensional coordinates PI = (Xi, Yi, Zi) at the four corners of the tag can be calculated. The shape measurement control unit 41 outputs the three-dimensional coordinates of the four corners of the tag and the coordinates of the four corners of the package 7.

(Camera Control Unit 3) Next, the camera control unit 3 according to the second embodiment will be described.

The camera controller 3 inputs the three-dimensional coordinates of the four corners of the tag and the three-dimensional coordinates of the four corners of the package 7.

With the traveling direction of the belt conveyor 1 as the axis S (the traveling direction of the package 7 is the + direction), the three-dimensional coordinates of the four corners of the package 7 and the three-dimensional coordinates of the four corners of the tag are perpendicular to the axis S. Lower.

Among the coordinates of the legs of the perpendiculars descending from the four corners of the package 7, the largest one is Sf and the smallest one is Sl.

[0113] Of the coordinates of the legs of the perpendiculars descending from the four corners of the tag, the largest one is S'f and the smallest one is S'l.

The Z coordinate of the corner taking S'f is Zf, S'l
Let Zl be the Z coordinate of the corner that takes

Further, the amount of progress of the belt conveyor 1 per unit time is dS.

In the camera control section 3, the in-focus image input section 4
When the luggage 7 arrives below the line scan camera 6 of

(Equation 10) And according to the progress of the belt conveyor 1.

[Equation 11] Change the shooting step by step.

According to the method described above, even if the tag is attached to a plane oblique to the camera of the baggage 7, the focus can be continuously adjusted according to the movement of the baggage 7 to cover the entire tag. An in-focus image can be taken.

[0118]

According to the image processing apparatus and the image processing method of the present invention, an image taken by the image input means is processed to obtain information necessary for taking an image of the luggage in a focused state. Since the focal point is controlled by using the information and the focused image is input by the focused image input means, the information written on the tag can be accurately read from the tag attached to the package.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image processing apparatus 10.

FIG. 2 is a configuration diagram of a shape measuring unit 2.

FIG. 3 is a flowchart of the shape measuring unit 2.

FIG. 4 is a vertex of a convex hull.

FIG. 5 is an example of a package having a complicated shape.

FIG. 6 is a diagram showing a relationship between a package and a line scan camera.

FIG. 7 is a configuration diagram of the tag reading unit 5.

FIG. 8 is a flowchart of a tag detection process.

FIG. 9 is a configuration diagram of a shape measuring unit 2 according to a second embodiment.

FIG. 10 is a diagram showing a camera coordinate system and a three-dimensional coordinate system.

[Explanation of symbols]

 Reference Signs List 1 belt conveyor 2 shape measurement unit 3 camera control unit 4 focused image input unit 5 tag reading unit 6 line scan camera 7 luggage 11 camera 12 camera

Claims (6)

[Claims] 1. An image input means for inputting an image of a load, a focused image input means for inputting an image focused on the load, and a load from the image taken by the image input means. A shape measuring means for obtaining focus information necessary for capturing an image in focus by the focused image input means; and controlling a focus of the focused image input means using the focus information measured by the shape measuring means. An image processing apparatus comprising: a focus control unit that reads information on a tag of the package from an image input by the focused image input unit; 2. The shape measuring means measures a three-dimensional shape of the baggage, and sends the measured three-dimensional shape information to the focus control means as focus information. Calculating the distance between the package and the in-focus image input means based on the three-dimensional shape information of
2. The image processing apparatus according to claim 1, wherein the focus of the focused image input unit is adjusted based on the distance. 3. The belt is conveyed by a belt conveyor, wherein the image input means is disposed at a first point of the belt conveyor, and the focus is at a second point after the first point of the belt conveyor. 2. The image processing apparatus according to claim 1, further comprising a line scan camera serving as an image input unit. 4. An image inputting step of inputting an image of a package, a focused image inputting step of inputting an image adjusted in focus on the package by a tag reading camera, and the image captured in the image inputting step. From, a shape measuring step of obtaining the focus information necessary to capture the luggage in focus with the tag reading camera, and the tag reading camera using the focus information measured in the shape measuring step An image processing method, comprising: a focus control step of controlling a focus; and a tag reading step of reading information described on a tag of the package from the image input in the focused image inputting step. 5. The shape measuring step measures a three-dimensional shape of the package, and sends the measured three-dimensional shape information as focus information to the focus control step. The focus control step includes: 5. The image according to claim 4, wherein a distance between the package and the tag reading camera is calculated based on the three-dimensional shape information, and a focus of the tag reading camera is adjusted based on the distance. Processing method. 6. The baggage is conveyed by a belt conveyor, an image input camera used in the image input step is arranged at a first point on the belt conveyor, and a first point after the first point on the belt conveyor. 5. The image processing method according to claim 4, wherein a line scan camera serving as the tag reading camera is disposed at the second point.