RU2848167C1 - System and method for measuring the size of an object in a plane - Google Patents

Abstract

FIELD: measurement technology.

SUBSTANCE: invention relates to the field of measurement technology, in particular to the determination of the dimensions of objects using an optical system. The system includes: a lens (1), a light-sensitive matrix (2), a measurement area (3), wherein the lens (1) is equal to or larger than the measurement area (3), and the light-sensitive matrix (2) is located behind the lens (1) to capture strictly parallel light beams (5), illumination that highlights the object (6) against the background against which the object (6) is measured, a computing device connected to the light-sensitive matrix (2) and designed to perform the measurement process. The method includes the steps of placing the object (6) in the measurement zone (3), fixing the image of the parallel projection of the object (6) together with the background by means of the lens (1) and the light-sensitive matrix (2), wherein the parallel projection obtained is formed by focusing parallel rays of light behind the lens (1) on the light-sensitive matrix (2), in the obtained image, the image of the object (6) is separated from the background, the size of the object (6) in pixels on the image is determined, the sizes in pixels are converted into the actual sizes of the object (6) in the projection plane using pre-set calibration coefficients.

EFFECT: ability to quickly, accurately and easily measure the dimensions of an object on a plane, simplifying setup and maintenance.

8 cl, 8 dwg

Description

Field of technology

The invention relates to the field of determining the sizes of objects.

State of the art

Currently, a large number of different systems and methods for determining the sizes of various objects are known.

A prior art invention for measuring weight and size characteristics (CN 102506754 B, November 6, 2023) is known. This source of information discloses a method for using a confocal measuring device for both appearance and color measurements, which includes the following sequential steps: mounting a test sample on a three-dimensional sample scanning platform and supplying light from a standard white light source, which passes through a multimode light guide fiber and is focused by an optical system. The beams then converge, focus on the surface of the test sample, and are focused into a light spot and reflected; the reflected light returns to its original path, hits the core diameter at the head end of the fiber, and is separated by the multimode light guide. Input to the spectral color recognition unit and confocal determination of the white light position, a unit for obtaining color and microheight information at a point in the sample, respectively. The three-dimensional sample scanning platform sets the sample in three-dimensional scanning motion, which can then obtain the shape of the sample surface and color information at the corresponding point.

This source also discloses a device used in the method, which includes a standard white light source, a multimode light-guiding fiber, an optical fiber, a head end, an optical head holder, a focusing optical system, a 3D sample scanning platform, a confocal position measurement unit, a spectral color recognition unit, and a computer. The standard white light source, the confocal white light position measurement unit, and the spectral color recognition unit are connected to the head end of the optical fiber via the optical fiber. The head end is held by the fiber head holder; the head end of the fiber, the focusing optical system, and the 3D sample scanning platform are connected to a common optical path.

The disadvantage of this solution is the need for lengthy scanning to determine the size of objects.

Also known in the prior art is a device for measuring the shape of objects (US 11002535 B2, 11.05.2021), which comprises a platform on which the measurement object is placed; a light projection section configured to emit patterned light onto the measurement object located on the stage; a light receiving section configured to receive the patterned light emitted from the light projection section and reflected from the measurement object, and to output light receiving data; an optical axis direction control section configured to relative move the stage in the direction of the optical axis relative to the light receiving section in order to thereby adjust the focal position of the light receiving section; and a control section for performing: processing the collection of stereoscopic shape data for measurement based on the light receiving data output by the light receiving unit, using a pattern projection method, the shape of the measurement object present in the measurement range on the plane of the stage, orthogonal to the optical axis of the light receiving section and obtaining stereoscopic shape data.

A disadvantage of this solution is the long time it takes to determine the size of an object. Another disadvantage is the use of stereo vision for shape estimation, which often leads to measurement errors due to inaccurate matching of points from different cameras.

Disclosure of invention

The problem that the claimed solution is aimed at solving is to eliminate the shortcomings identified in the prior art.

The technical result of the claimed invention consists in:

- using all the advantages of parallel projection for measuring the size of an object: the size of the object on the parallel projection image is constant, regardless of the distance between the object and the lens (Fig. 2.1, Fig. 2.2, Fig. 3); there are no invisible parts of the object (in the dead invisible zone from the camera) that affect its size in a given plane (Fig. 2.2); the size of the object includes both its elements located close to the lens and those located far away (Fig. 4.1, Fig. 4.2),

- a high degree of measurement objectivity. The size in the plane is determined as is, without dead zones and assumptions (Fig. 2.2),

- constant measurement accuracy, independent of the geometric dimensions of the measured object,

- measurement of the object's size in one frame from the light-sensitive matrix (Fig. 4.1, Fig. 4.2). Or a linear dependence of the object's measurement time on its size, in the case where the object is larger than the measurement zone and it is necessary to collect its full parallel projection in parts,

- a simple and predictable way of measuring, simplifying setup and maintenance.

The claimed technical result is achieved in that the system for measuring the size of an object (6) in a plane includes: a lens (1), a light-sensitive matrix (2), a measurement zone (3), a lens (1) of a size equal to or larger than the measurement zone (3), a light-sensitive matrix (2), located behind the lens (1) for recording strictly parallel beams of light (5), a computing device connected to the light-sensitive matrix, and configured to carry out the measurement process.

The system additionally contains a backlight (4) located behind the measurement zone (3) and serving as a background for the object (6).

The system additionally contains a fixed background, against which the measured object (6) is illuminated by front (7) or side illumination.

The system additionally contains enclosing structures designed to limit access of extraneous light to the measurement zone (3) and the lens (1).

The internal walls of the enclosing structures are made of a matte black color, which absorbs light from backlighting or external light sources.

The lens (1) is a diverging lens, and the measurement area (3) is smaller than the light-sensitive matrix (2)

The lens (1) is a collecting lens, and the measurement zone (3) is larger than the light-sensitive matrix (2) and the matrix is at a focal distance from the lens (1).

A method for measuring the size of an object (6) in a plane, by means of a system according to any of the preceding paragraphs, including the stages in which the object (6) enters the measurement zone (3), an image of a parallel projection of the object (6) is recorded together with the background, by means of a lens (1) and a light-sensitive matrix (2), wherein the resulting parallel projection is formed by focusing parallel rays of light behind the lens (1) on the light-sensitive matrix (2), an image of the object (6) is separated from the background in the resulting image, the size of the object (6) in pixels in the image is determined, the dimensions in pixels are converted into the actual dimensions of the object (6) in the projection plane using pre-fixed calibration coefficients.

Brief description of the drawings

Fig. 1.1 shows the visibility of an object in diverging beams of light. It can be seen that a nearby small object has the same apparent size as a large but distant one.

Fig. 1.2 shows the visibility of an object in diverging beams of light. It's clear that the portion of the object that influences its size is in the shadow (dead zone) of the camera. Therefore, with this method of observing the object, its size will be determined incorrectly.

In Fig. 1.3 the camera sees only the upper part of the truncated pyramid, and the shape and size of the invisible part can only be guessed at.

Fig. 2.1 shows the visibility of an object in parallel beams of light. It is clear that its apparent size does not change with distance.

Fig. 2.2 shows the visibility of an object in parallel beams of light and the correspondence between the apparent and actual size of the object. Parts of the object that do not affect its size in a given plane are in the shadow (dead zone) of the camera. Thus, the object's size in a plane perpendicular to the beams of light will be recorded correctly.

Fig. 2.3 schematically shows operation with a measurement zone smaller than the light-sensitive matrix. A diverging lens is used.

Fig. 3 shows a diagram of how parallel light rays strike a lens and are then focused on a light-sensitive matrix during object size measurement. Front-side illumination is used. It is clear that the object's size in the image should not change with distance.

Fig. 4.1 schematically shows a variant of a system for measuring the size of an object in the horizontal plane (XY).

Fig. 4.2 schematically shows a variant of the system for measuring the height of an object relative to a platform in the vertical plane (Z).

Fig. 4.3 schematically shows a measurement system that simultaneously measures both the size of an object in the horizontal plane (XY) and its height relative to the platform in the vertical plane (Z). To prevent the two backlights from interfering with each other, different colors are used, easily filtered from each other.

Fig. 5 shows an example of an image obtained using the claimed invention of identical objects located at different distances. It is evident that the size of the objects in the image is the same, regardless of their distance.

Fig. 6 shows an example of the operation of a system with simultaneous measurement of both the size in the horizontal plane and the height above the platform in the vertical plane, using the measurement method of the present invention.

Fig. 7 shows another example of the operation of another system using the measurement method of the present invention. Objects enter the measurement zone in a continuous mode on a conveyor belt.

In Fig. 8, the system is similar to Fig. 7. An object significantly longer than the measurement zone passed through the measurement zone. Its image is obtained by stitching together the required number of frames.

Implementation of the invention .

When capturing an object using conventional video equipment, the image doesn't convey complete information about its size. Nearby objects appear large in the image, while distant objects appear small (Fig. 1.1). Therefore, to determine the size of an object, it's necessary to know its distance. Also, if the object is very close, the camera can't see its edges; they appear in the blind spot, and one must guess what's hidden by the visible part of the object (Fig. 1.2). The size of the object shouldn't be measured, but rather assumed. If a box is used as the object being measured, its visible surface corresponds to its dimensions. However, if the object being measured is, for example, a truncated pyramid, its size and shape can only be guessed at (Fig. 1.3).

In the proposed measurement method, a lens (1) with a size greater than or equal to the measurement zone (3) is used to record the measured object (6) on video. Specifically, for a converging lens, if video equipment is placed at a focal length behind the lens (1), an image will be recorded in parallel beams of light (Fig. 2.1, Fig. 3). This results in a parallel projection of the object onto a plane perpendicular to the beams of light. A distinctive feature of parallel projection is that the size of the object's projection does not change with distance to the object. This means that knowing the distance to the object is no longer necessary to measure its size. A real photograph of two identical objects (6) located at different distances from the lens is shown in Fig. 5.

Another advantage of parallel projection is that there cannot be any hidden parts of the object in the shadow of the front part of the object that affect its size in the projection plane (Fig. 2.2).

Using this principle, one can measure both the object's size in a plane (Fig. 4.1) and its height above the platform (Fig. 4.2). This can be done simultaneously (Fig. 4.3), immediately obtaining the object's full XYZ dimensions (6). Gravity presses the measured object to the platform, and its highest point in the lateral parallel projection becomes its natural height. To prevent the backlights from interfering with each other when used simultaneously, they can be made different colors for easy color filtering. Each camera sees only one—its own backlight—while filtering out the others.

The lens (1) can be created using various physical principles. Depending on the lens type, the arrangement of its elements must change accordingly. For example, when using a mirror lens, it must be positioned at an angle, with the measurement zone (3), the light-sensitive matrix (2), and the background all on the same side. Regardless of the lens type (1), the light-sensitive matrix (2) must be positioned at the focal point of the lens (1) to capture parallel beams of light passing through the measurement zone (3), creating a parallel projection of the object (6).

A converging lens (1) is used if the measurement zone (3) is larger than the light-sensitive matrix (2) (the lens of the video camera). If the measurement zone (3) is smaller, a diverging lens is used. In any case, it is necessary to ensure that the light rays recorded in the measurement zone (3) are parallel to capture a parallel projection.

To speed up the process of measuring objects (6), it's possible to measure objects passing through the measurement zone (3), for example, on a conveyor belt. With high illumination, the required camera exposure is very low, the measurement occurs almost instantly, and there's no need to hold the object in place.

This method can also be used for objects larger than the measurement zone (3). However, in this case, it is necessary to capture the object in sections, passing it through the measurement zone (3) (or moving the measurement zone (3) over a stationary object). Then, the resulting raster images are combined into a single image of the entire object and its size is determined (Fig. 8).

The object's size in the image is first determined as the pixel size of the minimum bounding rectangle (e.g., Intel RealSense, Graham's algorithm). Then, using a predetermined calibration coefficient, the pixel size is converted to the actual size in the required units of measurement (mm, m, etc.).

This measurement method can be integrated into weight and size characteristics (WSC) measurement systems, both for a fixed object in the measurement zone (Fig. 6) and for continuous measurement of objects moving through the measurement zone (Fig. 7, Fig. 8).

Measurement accuracy depends on the camera resolution used, lens quality, and assembly precision. For example, using a 640x360 resolution with a measurement area size of 300x300x200, the measurement error is less than 1 mm.

The system for measuring the size of an object (6) in a plane includes a lens (1), a light-sensitive matrix (2), a measurement zone (3) and a computing device (Fig. 1, Fig. 2.1, Fig. 2.2, Fig. 2.3). The lens (1) is equal in size to or larger than the measurement zone (3). The light-sensitive matrix (2) is located behind the lens (1) for recording strictly parallel beams of light (5). The computing device is connected to the light-sensitive matrix and is configured to perform the measurement process.

The lens (1) can be any type that focuses parallel rays of light onto the photosensitive matrix (2). For example, a Fresnel lens, a mirror lens, a regular convex lens, etc.

To facilitate object identification in the image, the system includes a backlight (4) located behind the measurement zone (3). The interior walls of the enclosing structures are finished in matte black, absorbing light from the backlight (4) or other external light sources.

The backlight (4) can be of a fixed color, against the background of which the measured object (6) is illuminated by the front (7) or side lighting.

A system in which the measurement zone (3) is smaller than the light-sensitive matrix (2) and a diverging lens (1) is used to capture an image of a parallel projection of the objects being measured.

Additionally, the system includes enclosing structures designed to restrict ambient light from reaching the measurement zone (3) and the collecting lens (1). To enhance efficiency, the inner walls of the enclosing structures are finished in matte black, absorbing light from backlights or external light sources.

The system operates as follows: an object (6) enters the measurement zone (3), and using parallel beams of light, the lens (1) and photosensitive matrix (2) capture an image of the parallel projection of the object (6) along with the background in a plane perpendicular to the beams of light. The resulting image is then separated from the background, the size of the object (6) in pixels in the image is determined, and the pixel dimensions are converted to the actual dimensions of the object in the projection plane using pre-set calibration coefficients. In a particular case, the vector contour of the object (6) in pixels in the image is determined, simplified to a convex polygon, and then the minimum rectangle in which the vector contour of the object is located is determined by iterating over the sides of the convex polygon.

The calibration factor can be calculated as follows: the program defines the object's size as 200 x 200 mm, but the physical size is actually 150 x 150 mm. To get 150, multiply 200 by 0.75 (150/200 = 0.75). Therefore, the calibration factor is 0.75.

An example of an image of two identical objects at different distances photographed in a parallel optical system (Fig. 5). It is clear that focusing deteriorates with distance, but the size of the object remains the same regardless of distance.

An example of the proposed system's operation (Fig. 6). The object is manually placed in the measurement zone. A transparent cup is installed. It is clearly visible against the backlight (4). The cup's projection and size in the horizontal plane (XY) and projection and height in the lateral plane (XZ) are visible. The measurement time for one object is approximately 4 ms (limited by the camera frame rate) with an accuracy of approximately 1 mm.

Another example of the proposed system's operation (Fig. 7, Fig. 8). The objects to be measured (6) pass through the measurement zone (3) on a conveyor. The time, dimensions, and black-and-white XY and XZ projection images of the passing objects are recorded. Objects can be significantly longer than the measurement zone (Fig. 8), as their images are composed of the required number of frames. The measurement time for one object depends on the conveyor belt speed, but is no faster than 4 ms (limited by the camera frame rate), with an accuracy of approximately 1 mm.

Claims (18)

1. A system for measuring the size of an object (6) in a plane, including: lens (1), light-sensitive matrix (2), lens (1) of size equal to or larger than the measurement zone (3), the light-sensitive matrix (2) is located behind the lens (1) to record strictly parallel beams of light (5), backlighting that ensures the selection of the object (6) against the background of which the object (6) is measured, a computing device connected to a light-sensitive matrix (2) and configured to carry out a measurement process. 2. The system according to claim 1, further comprising a backlight (4) located behind the measurement zone (3) and serving as a background for the object (6). 3. The system according to claim 1, further comprising a fixed background, against which the measured object (6) is illuminated by front (7) or side illumination. 4. The system according to paragraphs 1-3, additionally containing enclosing structures designed to limit access of extraneous light to the measurement zone (3) and the lens (1). 5. The system according to paragraph 4, in which the internal walls of the enclosing structures are made in a matte black color that absorbs light from backlights or external light sources. 6. The system according to claim 1, in which the lens (1) is a diverging lens, and the measurement zone (3) is smaller than the light-sensitive matrix (2). 7. The system according to claim 1, in which the lens (1) is a collecting lens, and the measurement zone (3) is larger than the light-sensitive matrix (2) and the matrix is at a focal distance from the lens (1). 8. A method for measuring the size of an object (6) in a plane by means of a system according to any of the preceding claims, comprising the steps of, object (6) enters the measurement zone (3), the image of the parallel projection of the object (6) is captured together with the background by means of a lens (1) and a light-sensitive matrix (2), wherein the resulting parallel projection is formed by focusing parallel rays of light behind the lens (1) on the light-sensitive matrix (2), on the resulting image, the image of the object (6) is separated from the background, determine the size of the object (6) in pixels in the image, The pixel dimensions are converted into the actual dimensions of the object (6) in the projection plane using pre-fixed calibration coefficients.