DE102013106345A1 - Object detection system and examination device for visual disturbances - Google Patents

Abstract

The invention relates to an object detection system (1) for detecting the position and / or the orientation of an object (2) in a room (3) with a computer unit, a camera (4) connected to this for the transmission of image data and at least one from the camera (4) spaced reference element (5), the computer unit being designed such that it automatically recognizes the reference element (5) in the camera image (8) of the camera (4) and uses an algorithm to calculate the position of the object (2) in a spatial coordinate system . According to the invention, the camera (4) is fixed to the object and the reference element (5) is fixed in space. The invention further relates to an examination arrangement for visual disorders with such an object detection system.

Description

The present invention relates to an object detection system for detecting the position and / or orientation of an object in a room with a computer unit, a camera connected thereto for transmitting image data and at least one reference element spaced from the camera, wherein the computer unit is designed such that it automatically recognizes the reference element in the camera image of the camera and calculates the position and / or orientation of the object in a spatial coordinate system by means of an algorithm. Furthermore, the present invention relates to an examination arrangement for visual disturbances, in particular strabismus, with a tangent board and a means for measuring and automatic evaluation of the eye position and / or orientation relative to the tangent board.

The automatic detection of objects in a defined space is an important prerequisite for reality-based information presentation, especially in the case of virtual reality and augmented reality applications. In this case, the position and / or the orientation of the object in space is automatically determined via an object detection system, which is referred to in technical jargon as a tracking system. In this case, the movement or even only the static position and orientation of the object, in particular the head or other body parts, such as the hand, are automatically detected by the tracking system. Current tracking systems have a number of disadvantages. Thus, these are often not precise enough, allow only a small radius of interaction and / or are very expensive to purchase, so they are denied the mass use.

An object detection system for detecting the position and orientation of a person in a room is, for example, from US 2011/0025852 A1 known. Here, a plurality of motion detection cameras are spaced from each other arranged so that they define a movement space in which the person to be detected can move freely. The human being whose movement is to be detected is provided with several markers. The markers are thus firmly connected to the object. In contrast, the motion detection cameras are static or space fixed. The markers permanently connected to the object are automatically detected by the cameras in the respective camera image and their position calculated by a computer unit in a spatial coordinate system. Such object detection systems have the disadvantage that they require many, in the acquisition very expensive motion detection cameras for a sufficiently high detection accuracy. As a result, these object detection systems are very expensive. Another disadvantage is that in these systems, the interaction radius, ie the space in which the object can move freely, can only be chosen relatively small. This is due in particular to the fact that several cameras would also be required for a larger interaction space, which is only possible to a small extent due to the high acquisition costs of the motion detection cameras. Furthermore, the precision of the object detection system with the greater distance of the cameras from the object to be detected deteriorates so much that these systems can be used only in a relatively small interaction space.

From the EP 0 940 117 B1 there is known a method and apparatus for performing visual examinations, especially strabismus. In this case, projection devices, visual impression encoders, fixation and pointing objects, Tangententafeln and electronic and computer-aided measuring and evaluation means, with monitoring of the head and / or eyeball position used. Thus, a virtual Tangententafel can be generated on a display area computer-controlled to the coordinate-related visual impressions are freely presented. However, the systems used in this context for detecting the head position do not meet the high requirements in terms of recognition accuracy and low acquisition costs.

The object of the present invention is therefore to provide an object detection system for detecting the position and / or orientation of an object in a room, by means of which the above-mentioned disadvantages can be eliminated. Furthermore, the object is to provide an examination arrangement for visual disturbances, in particular strabismus, with such an object detection system for eliminating the aforementioned disadvantages.

The object is achieved by an object detection system and an examination arrangement having the features of the independent patent claims.

According to the invention, the object detection system for detecting the position and / or the orientation of an object in a room has a computer unit, a camera connected to it for the transmission of image data, and at least one reference element spaced from the camera. The object to be detected in this context is preferably the head and / or the hand of a human being, which play an essential role in common object detection applications. Of course, the invention is not on limited to these objects. Thus, the present object detection system is suitable for any type of object, in particular for detecting additional or other parts of the body, of the entire body or else also of remotely controlled objects, such as, for example, flying, floating, and / or mobile model vehicles. The camera is preferably wireless, in particular via radio or Bluetooth, and / or wired or wire connected to the computer unit. The computer unit can be attached to the object at a distance from the camera and thus move with it. Alternatively, this can also be arranged statically at a fixed location, in particular in the room. In order to increase the object detection speed, it is also advantageous if the distance between the camera and the computer unit is kept as low as possible. For this purpose, it is advantageous if the camera and computer unit are designed as a component, in particular form a unit formed as a chip.

The computer unit is designed such that it automatically recognizes the reference element in the camera image of the camera and calculates the position and / or the orientation of the object in a spatial coordinate system by means of an algorithm. According to the invention, the camera is fixed to the object and the reference element is designed to be fixed in space. As a result, it is not like the one from the US 2011/0025852 A1 known system, a variety of expensive cameras necessary to determine the exact position and / or orientation of the object in a defined range of motion, but it can be reduced to a small compared to the number of cameras. As a result, the object detection system can be made very cost-effective, since the reference elements are much cheaper to purchase compared to the camera. Preferably, the object detection system has more reference elements than cameras, although it can also manage with a single camera.

A major disadvantage of existing systems is their positional inaccuracies, which increases with greater distance between the reference element and the camera. This is due in particular to the fact that the basic principle of such object detection systems is based on the recognition and evaluation of the perspective distortion of the reference elements, in particular the angular markers, in the camera image. This perspective distortion can no longer reliably be detected by the camera and / or the computer unit over greater distances, as a result of which the accuracy of the object detection system greatly deteriorates. These problems can be solved if the computer unit calculates the position of the object in the spatial coordinate system according to the invention additionally or alternatively to the perspective distortion by means of at least two intersecting barrel circle arcs.

In this case, the computer unit preferably calculates the points of intersection of at least two, in particular at least three, barrel arcs that intersect. These at least two barrel arcs may preferably lie in the same spatial plane or in different, mutually inclined spatial planes.

The geometric term "barrel arc" refers to a circular arc over a stretch from which this stretch always appears at the same angle. The line thus forms a chord of the barrel circle arc completed to a closed circle. The angle at which this string is viewed from all points of the barrel arc is called the circumference angle. A special case of the barrel arc is the Thales semicircle, in which the circumferential angle is 90 ° and the chord forms the diameter of the completed circle. As a result, when determining the position of an object by means of at least one barrel arc, the object to be detected or the camera arranged thereon is located at one of the points forming the barrel arc, from which a distance extending from a first to a second reference element is below a viewing angle of the camera is detected. Preferably, the object is thus on the use of at least one barrel arc on the barrel arc, so that the possibility of the positions in question are limited in space.

It is advantageous if the object detection system comprises a plurality of spatially fixed reference elements spaced apart from one another, in particular in the x, y and / or z direction of the spatial coordinate system. This allows a particularly large space in which the object can move freely, are formed. Furthermore, the recognition accuracy of the object detection system can be improved.

For detecting the position of the object in space, it is advantageous if the position of at least one of the reference elements in the spatial coordinate system and / or its real distance to at least one other reference element is stored in a memory of the computer unit. Additionally or alternatively, it is also advantageous if the position of at least one of these reference elements in the spatial coordinate system and / or its real distance to one or more other reference elements, in particular to an adjacent, independently by the computer unit can be measured and / or calculated. Advantageously, the reference elements can thus very quickly and with little effort, ie without the user having to pay attention to an exact positioning of the reference elements in the room, they are placed in the room.

In order to establish a relationship between the real environment and the virtual environment of the computer unit, it is advantageous if the computer unit is designed such that it creates a virtual, in particular three-dimensional space model by means of a plurality of reference elements and / or stores in the memory.

The computer unit calculates the respective barrel arc on the basis of known input data, namely on the basis of the known real distance of a first and a second reference element and / or a visual angle, under which this distance is seen from the object. In this case, the real distance of the two reference elements located on the barrel arc forms a chord of the barrel arc or of the barrel circle. The visual angle under which this tendon is seen from the object position to be calculated forms the circumferential angle of the barrel arc. As a result, the position of the object to be detected is advantageously reduced to the positions forming the barrel arc.

In a further advantageous embodiment of the invention, the computer unit is designed such that it is able to calculate the visual angle independently. Advantageously, the computer unit determines the shortest, in particular the horizontal and / or vertical, pixel spacing between the two reference elements from the camera image and / or calculates the visual angle by means of the camera lens angle known to the computer unit, preferably stored in memory. The viewing angle at which the tendon is seen from the object position to be calculated can thus be determined very quickly and simply from the measured pixel pitch and the camera lens angle known to the computer unit.

For position determination in three-dimensional space, it is advantageous if at least two barrel arcs lie in different and mutually inclined spatial planes of the spatial coordinate system.

In this regard, it is also advantageous if the computer unit forms at least two intersecting tori, in particular spindle tori, which are each formed from a plurality of barrel arcs. A torus is a geometric body that is formed by rotating a circle around a rotation axis that is in the same plane as the circle. The spindle torus represents a special case. In this case, the axis of rotation lies within the circle to be rotated. The axis of rotation thus represents a chord of the circle. Thus, to form the torus, the barrel circle or the barrel arc forms the circle or arc to be rotated and the chord of the barrel circle or the barrel arc the axis of rotation. The surface of the torus, in particular spindle torus, thus forms the sum of all points from which the chord or the real distance between the two reference elements in three-dimensional space appears at the same visual angle.

Advantageously, therefore, the barrel arches forming the respective spindle arches are formed by means of the two same reference elements, so that all of these - the respective torus forming - barrel arcs have the same real distance and the same angle of vision, or include points, from which viewed the distance between the two reference elements appears under the same angle of vision.

In order to determine the object position or the camera position in space, it is advantageous if the computer unit determines at least some of the intersections of the intersecting barrel arcs and / or tori and / or calculates a single common intersection point from these intersection points whose position in the spatial coordinate system matches the position of the interpolation point Camera in the room corresponds. Preferably, these have a single common point of intersection all those Fassbogenbögen which lie in a space having the object of the spatial coordinate system. As a result, it suffices to determine the position in two-dimensional space, to use at least two, preferably at least three barrel arcs lying in this spatial plane. Additionally or alternatively, it is advantageous if these have a single common point of intersection all tori formed, preferably at least two, more preferably at least three.

It is advantageous if the barrel arcs have the common point of intersection whose position in the spatial coordinate system corresponds to the position of the camera in space. The computer unit is designed such that it can determine this common point of intersection. The computer unit can then infer the known relative position of the object to the camera to the exact object position in the room.

For further restricting the exact object position in the spatial coordinate system, it is advantageous if the computer unit, in particular for determining the first and second coordinates of the object in the spatial coordinate system, determines the common intersection of at least two, preferably at least three, drum arcs. Preferably, these Fassbogenbögen lie in one first the objet having and / or common room level.

It is advantageous if at least three reference elements are arranged relative to one another in space so that the chords of the drum arches formed thereby are aligned coaxially with each other. Additionally or alternatively, the chords of the barrel arcs may be spaced or at least partially overlapped.

To further restrict the possible object positions in the spatial coordinate system, it is advantageous if the computer unit, in particular for determining the third coordinate of the object in the spatial coordinate system, a single common intersection between the Fassbogenbögen lying in the first space level with at least one in a second object having the spatial plane determined by the spatial coordinate system Fassbogenbogen determined. Additionally or alternatively, however, the computer unit can also determine the common point of intersection of at least two, in particular at least three, barrel arcs lying in the second spatial plane. The second spatial plane is preferably inclined to the first spatial plane, in particular to this orthogonal aligned.

The accuracy of the object detection system can be further improved if the computer unit for determining the object position compares a plurality of determined intersections and / or determines the object position in space or a single common intersection between all Fassbogenbögen based on a frequency distribution, preferably within a tolerance range.

For quick and uncomplicated commissioning of the object detection system, it is advantageous if the camera by means of a holder releasably with the object, in particular a head or a hand of a human being, in a fixed position to the object is fastened.

It is advantageous if the object detection system has a plurality of cameras, in particular four, aligned in different directions, in particular offset by 90 ° relative to each other. As a result, the accuracy of the object detection system can be improved, as several reference elements can be detected simultaneously by the cameras and, as a result, the computer unit can create several barrel arcs and determine their points of intersection.

Additionally or alternatively, it is advantageous if the object detection system has an acceleration sensor connected to the holder. With this minimal movements of the object can be sensed, which are not detectable by the camera, so that the accuracy of the object detection system can be improved.

In an advantageous development of the invention, the holder is a head-based vision system for determining the position of at least one eye and / or the viewing direction of a person in relation to its surroundings. As a result, there is a hybrid system including the object detection system and the gaze detection system. In this case, the camera of the object detection system is preferably designed as a field camera of the gaze detection system. The gaze detection system comprises an eye camera by means of which the eye of a user can be detected directly and / or indirectly, in particular via a semitransparent mirror. The field camera is in the direction of the object, i. of the user, and therefore records the field of view of the person. The video images captured by the eye and field cameras are billed together to determine the exact location the person is looking at.

The computational complexity of the computer unit can be reduced if at least three reference elements are arranged relative to one another in a row, preferably horizontal and / or vertical, and / or have a standardized size. Furthermore, this can reduce the error rate of the object detection system.

Advantageously, the reference element is a markable in space, in particular on a room wall and / or a space object, attachable. The marker preferably has a pattern, so that it can be clearly identified by the computer unit and / or its orientation relative to the camera can be detected. Advantageously, it is thus possible to deduce the orientation of the object in the spatial coordinate system if the computer unit knows the orientation of the marker in space.

Additionally or alternatively, it is advantageous if the reference element is a spatial object already in the room, in particular a wall, a door, a window, a piece of furniture and / or a strip advertising.

Alternatively or additionally, it is also advantageous if the reference element is designed as a part of these attachable markers and / or these space objects located in space. Preferably, the reference element is a reference point or has a plurality of reference points. This reference point can be, for example, the corner and / or the center of a marker or spatial object. A square marker would thus have, for example, five reference points.

In a particularly advantageous embodiment of the invention, the computer unit determines the, preferably substantially vertically and / or horizontally aligned, edges of the marker or the space object. The term "edge" is not necessarily the edged end of a marker or spatial object to understand, but rather defined by the camera in the camera image detectable edge, which is preferably identifiable as a rectilinear contour of the computer unit. Subsequently, the computer unit can determine the corner points via the detected edges, which ultimately form the reference points for determining the object position.

Advantageously, a marker and / or spatial object has a plurality of reference elements spaced apart from one another, in particular horizontally and / or vertically. In order to be able to create at least three barrel arcs in one spatial plane, the computer unit advantageously thus requires only a minimum of two markers, in particular rectangles or squares.

In order to be able to reduce the computational outlay for the computer unit, it is advantageous if the computer unit knows the size and / or the real distance between the reference elements. For this purpose, it is advantageous if these real distances and / or real variables are standardized and stored in a memory of the object detection system.

It is advantageous if a plurality of reference elements, in particular a plurality of markers, are arranged on a wallpaper which can be applied in the room and / or on a detachable and / or roll-up strip which is preferably mutually longitudinally of the wallpaper and / or the strip, in particular equidistant and / or or are spaced to the respective adjacent reference element in one of the computer unit known real distance.

For quick attachment of reference elements in space, it is advantageous if the tape is designed as an adhesive tape. Alternatively, it is advantageous if the tape is rotatably mounted in the housing for unwinding and rolling about a longitudinal axis of a housing.

In order to form a reference element enclosure which defines the space in which the object can move freely, it is advantageous if the band can be detachably connected in the region of its free end to a further band and / or housing.

Furthermore, it is advantageous if the housing preferably comprises at least two bands arranged parallel to one another and spaced apart from one another in the longitudinal direction of the housing, since this can improve the accuracy of the object detection system.

Additionally or alternatively, it is advantageous if a plurality of housings, in particular in the region of their end faces, can be connected to one another to form a column. Thus, the forming the Referenzelementgehege bands can be transported very easily and space-saving.

It is advantageous if a plurality of bands can be connected to a reference element enclosure defining the space, in particular an n-cornered one.

Advantageously, the object detection system is used in a method for detecting the position and / or orientation of an object in a room. The method in this case has the method features of the preceding description, which may be present individually or in any combination.

The examination arrangement according to the invention for visual disturbances, in particular strabismus, comprises a, in particular virtual, tangent board and a means for the measurement and automatic evaluation of the eye position and / or orientation with respect to the tangent board. A tangent board has a coordinate system according to ophthalmological calculations. In the following, the term "tangent board" will refer to the different ophthalmic devices, such as the ophthalmoscope. Harms wall, Helmholtz screen, Hess screen, Lee screen, Maddox cross etc, subsumed. For the purposes of the invention, the term "virtual tangent board" means that the mathematical-medical basis of the tangent board is computer-bound and can be retrieved as needed in digital or analog form and presented visually, in particular projected.

The examination arrangement for detecting the head position and / or head orientation comprises an object detection system, which is designed according to the preceding description, wherein said features may be present individually or in any desired combination. Advantageously, therefore, the patient need not assume a fixed forced position and / or orientation for the purpose of examining his present visual disturbance, but can freely position and move in a predetermined space or reference element enclosure. Furthermore, by means of the object detection system, changes in the position of the patient can be detected and preferably by a suitably designed one mathematical model can be compensated in the computer unit.

It is advantageous if the tangent board has at least one reference element which can be recognized by the camera of the object detection system in the camera image. Thus, no separate reference elements must be installed in the room. Advantageously, the corner points and / or the center of the tangent board form the reference elements or reference points. Furthermore, it is advantageous if the reference elements are virtually formed together with the tangent board.

It is also advantageous if the reference elements are infrared light emitting and / or reflecting markers that can not be perceived by the human eye. As a result, a distraction of the patient can be avoided by means of corresponding reference elements or markers.

Further advantages of the invention are described in the following exemplary embodiments. It shows:

1 a two-dimensional camera image of a room with reference elements,

2a to 2c a plan view of a virtual space model of the room with trained in a first spatial plane of a spatial coordinate system barrel arcs for determining the object position,

3 a top view of the virtual space model with three additional or alternative barrel arches,

4 a side view of the virtual space model with lying in a second space level of the spatial coordinate system barrel arcs,

5 an alternative embodiment of the object detection system with a plurality of object-fixed cameras,

6 a perspective view of an adhesive tape designed as a band with reference elements,

7 a perspective view of a housing with it off and / or rolled up arranged bands and

8th a perspective view of a formed from a plurality of housings column to form a Referenzelementgeheges.

The 1 to 4 illustrate a possible operation of an object detection system 1 for detecting the position and / or orientation of an object 2 in a room 3 , where the object is 2 in the real space 3 can move freely. In the present embodiment, the object 2 the head of a person, but also other body parts with the object detection system 1 could be detected.

The object detection system 1 for detecting the position and / or the orientation of the object 2 in the room 3 includes a computer unit, not shown here, a memory, not shown here, a camera 4 and / or several reference elements 5 , The computer unit not shown here is with the camera 4 so connected, that the camera 4 Transfer image data to the computer unit. The image data transmission can be between the camera 4 and the computer unit by means of a cable. Additionally or alternatively, this data transmission can also be carried out wirelessly, in particular via radio or Bluetooth. The camera 4 has a camera lens angle 6 on, the coverage of the camera 4 defined (cf. 5 ). This camera lens angle 6 is in the memory of the object detection system, not shown here 1 stored. The computer unit can store data in the memory and / or access data stored in the memory.

The camera 4 is according to 2a trained object. This means that the camera 4 on the object 2 at one of the computer unit known, ie preferably stored in memory, position and / or orientation, in particular releasably secured. By determining the position and / or orientation of the camera 4 in the room 3 can from the computer unit thus also the position of the object 2 and / or its orientation in space 3 be determined.

Unlike the camera 4 are the reference elements 5 spatially fixed trained. You are thus always at a fixed position in the room 3 , The position data of the reference elements 5 and / or their real distances 12 to each other, the computer unit not shown here are known. The position data and / or the real distances 12 can from the object detection system 1 independently measured, calculated and / or stored in the memory and / or already at least partially stored in the memory.

With a variety of in the room 3 arranged reference elements 5 the computer unit can connect to the real room 3 adapted virtual space model 7 create (cf. 2a ). This virtual space model 7 is stored by the computer unit in the memory of the computer unit. Alternatively, this can already be deposited in the memory.

In the present embodiment recognizes the computer unit for determining the head position of a person or the object 2 according to 1 first automatically from the camera 4 detected, solid-trained reference elements 5a . 5b . 5c . 5d . 5e . 5f in the camera picture 8th that of the object-fixed camera 4 is transmitted to the computer unit. Subsequently, the computer unit calculates the position of the object by means of an algorithm or method described in more detail below using this image information 2 in a spatial coordinate system. The origin of the space coordinate system can be at any point in the virtual space model 7 be arranged. Preferably, however, the origin is in a reference element 5 arranged, in the present embodiment, a corner of the room 3 or virtual room model 7 is.

1 shows a snapshot of the camera 4 recorded videos. This camera picture 8th that of the object-fixed camera 4 recorded and forwarded to the computer unit for further processing, shows the room 3 with its spatially trained reference elements 5a . 5b . 5c . 5d . 5e . 5f , but not all of these reference elements for clarity 5 are provided with reference numerals.

The reference elements 5a . 5b . 5c . 5d . 5e . 5f are preferable to the room 3 aligned vertically or horizontally. In the present embodiment, most reference elements 5a . 5b . 5c . 5d . 5e . 5f through in the room 3 attached markers 9 educated. The markers 9 have a pattern 10 on that every marker 9 individually marked. The computer unit can do this pattern 10 recognize through an image processing algorithm and thus each of these markers 9 clearly identify. About the pattern 10 the computing unit can change the orientation of the markers 9 in the camera picture 8th determine and hereby the orientation of the object 2 in the room 3 to calculate. The markers 9 have one of the computer unit known contour, shape and / or size. Preferably, all markers 9 the same basic geometric shape. Furthermore, all distances between the markers arranged in a horizontal or vertical row are standardized and known to the computer unit, ie stored in the memory.

The reference elements 5a . 5b . 5c . 5d . 5e . 5f are in the present embodiment, the upper vertices of the marker 9 , These are determined by the marker edges, which in reality, ie in the physical space 3 , vertically and / or horizontally aligned. Each of these markers 9 thus has several reference elements 5a . 5b . 5c . 5d . 5e . 5f or reference points, which may ultimately be formed by the four marker corners and the marker center.

To determine the position of the object 2 in a first spatial level of the spatial coordinate system, the computer unit determines according to 1 initially at least three, in particular arranged in a row, reference elements 5a . 5b . 5c , Subsequently, the computer unit determines from the camera image 8th the respective, in particular shortest, pixel spacing 11a . 11b . 11c these reference elements 5a . 5b . 5c to each other. The object detection system 1 thus does not determine the optical distortion of the markers to determine the object position, as in known methods 9 in the camera picture 8th because they can be detected at very large distances only very inaccurate and unreliable. Instead, we have the shortest pixel pitch 11a . 11b . 11c two mutually parallel reference elements 5a . 5b . 5c measured. Because the reference elements 5a . 5b . 5c from the computer unit in the camera image 8th can be reliably detected even at long distances, with the present method, a very high accuracy for determining the position of the object 2 in the room 3 be ensured.

In a second step calculates the computer unit based on this from the camera image 8th determined pixel distances 11a . 11b . 11c and the computer unit known camera lens angle 6 for each of these pixel distances 11a . 11b . 11c a real viewing angle α, β, γ, below which this by the respective pixel pitch 11a . 11b . 11c defined route is seen from the object position (cf. 2a to 2c ).

The computer unit are thus, for example, for the three reference elements 5a . 5b . 5c , the respective real distances 12a . 12b . 12c and the respective viewing angles α, β, γ, among which the real distances 12a . 12b . 12c from the camera position to be determined or object position are known. Based on this information, calculates the computer unit, as shown below, using Fassfassbögen 13 the unique object position in the room 3 by giving the one common point of intersection 15 of these barrel arcs 13 determined (cf. 2c ).

The real distances 12a . 12b . 12c each form a chord and the angle of view α, β, γ respectively the circumferential angle of the associated barrel circle arc 13a . 13b . 13c , The circumferential angle is the angle at which the respective chord of each point of each barrel arc 13a . 13b . 13c is seen from.

According to 2a becomes the first barrel circle arch 13a from the computer unit with the first real distance 12a and the first viewing angle α. The candidate positions of the object 2 or the camera 4 in the virtual space model 7 are thus on the first barrel arc 13a limited.

To further limit the choices created the computer unit according to 2 B over the second real distance 12b by means of the corresponding second viewing angle β a second barrel arc 13b , This cuts the first barrel arc 13a in at least one point of intersection 14 , Since the first and the second barrel arc 13a . 13b in the present embodiment in the same room level, they always have two intersections 14a . 14b on. The choices of the object positions of the object in question 2 in the virtual space model 7 can thus affect these two intersections 14a . 14b be reduced.

In the present embodiment, these two Fassbogenbögen would already 13a . 13b suffice for the position of the object 2 in the first spatial level can be clearly determined, as in accordance with 2 B the second intersection 14b the position of the second reference element 5b corresponds and could therefore be excluded.

Alternatively or additionally, but according to 2c the first and second barrel arc 13a . 13b with a third drum arc 13c get cut. These three barrel arcs 13a . 13b . 13c have several intersections with each other 14 but only a single common point of intersection 15 , This thus clearly determines the position of the object 2 in a first spatial plane, namely in the xz plane, of the spatial coordinate system. With the in 2a to 2c Illustrated methods may be two coordinates of the object 2 in the room 3 or in the virtual space model 7 , namely the x and z coordinates, are determined.

To increase the positioning accuracy, it is advantageous if the computer unit according to 3 more barrel arches 13d . 13e . 13f created and their intersections 14 for determining the one common point of intersection 15 calculated.

For the determination of the third coordinate, namely in the present case the y-coordinate, the computer unit generates according to 4 at least one further drum arc 13g . 13h . 13i , which is arranged in a second space plane inclined to the first spatial plane. For this purpose, the computer unit uses according to 1 the vertically arranged in series reference elements 5e . 5f and calculates the measured pixel distances analogously to the first spatial level 11d and the respective visual angles δ the points of intersection 14 the barrel arches 13g . 13h . 13i for determining the one common point of intersection 15 , In the present embodiment, the second spatial plane is the yz plane and thus aligned orthogonal to the first xz plane. The common point of intersection 15 corresponds to the common point of intersection found in the first spatial level 15 and thus defines the exact position of the object 2 in the room 3 or in the virtual space model 7 ,

Basically, the position of the object 2 in the room 3 by means of only two Fassbogenbögen 13 be determined if they are in mutually inclined spatial planes and only in a single common intersection 15 to cut. However, the determination of further barrel arcs can increase the accuracy and reduce the error rate of the object detection system.

The above with reference to the 1 to 4 explained method for determining the object position is particularly suitable for those applications where the distance of the object 2 essentially not or only slightly changes from the ground and the reference elements 5 or reference points are arranged substantially on camera height.

For use cases where the distance of the object 2 varies greatly to the ground, it is advantageous in an alternative embodiment not shown here, if in the 2a to 4 illustrated Fassbogenbögen 13 to three-dimensionally extending Tori in space, in particular Spindeltori, are extended. For this purpose, the tori by a rotation of the two-dimensional Fassfassbögen 13 about their respective chord or their respective real distance 12 educated. The Tori also have a single common point of intersection 15 which determines the position of the camera in three-dimensional space. The object position can then be determined via simple translation.

5 shows a plan view of the object detection system 1 , wherein the object detection system 1 in contrast to the previous embodiment, several with the object 2 firmly connected cameras 4a . 4b . 4c . 4d having. The orientation of the cameras 4a . 4b . 4c . 4d and their respective camera lens angle 6 , is chosen such that the detection areas of two adjacent cameras 4a . 4b . 4c . 4d in an overlap area 16 overlap (for clarity, only one of the camera lens angle and overlapping areas provided with a reference numeral). This can be a gapless 360 ° detection area of the object detection system 1 be formed.

5 further shows a room item 17 the one in the room 3 or virtual space model 7 is arranged. Especially if this space object 17 has a square or rectangular shape, its horizontal and / or vertical edges can be used to determine the reference points or vertices that come from the camera 4 can be detected. Alternatively or additionally, however, the room walls, in particular the room corners as reference elements 5 or reference points serve.

So the object detection system 1 the position of the object 2 in the room 3 can compute its computer unit, as already described above, the real distances 12 between the reference elements arranged in series 5 know. Because of this, it is advantageous if the reference elements have a defined size and / or a defined distance from each other and these parameters in the memory of the object detection system 1 are stored so that the computer unit can access this data.

To quickly and easily attach the reference elements 5 in the room 3 It is beneficial if these, as in 6 shown on a tape 18 are arranged. The ribbon 18 is presently designed as an adhesive tape and has several markers 9 on, which have the same size and the same distance to the adjacent marker.

According to 7 can the band 18a . 18b but also retractable and rollable in a housing 19 be stored. Next to the markers 9 show the bands 18a . 18b a dimension 23 on, leaving the total length of the unrolled tape 18a . 18b can be determined. According to the in 7 illustrated embodiment, two such mutually parallel bands 18a . 18b in the case 19 rotatably mounted. In the area of their ends, these have a fastener 20 on, by means of which the bands 18a . 18b on a second housing, not shown here, in its fastening region 21 can be releasably attached. With several connected bands 18 it is possible to form a reference element enclosure, not shown here, which surrounds that space 3 defines, in which the object 2 can move freely.

According to 8th can have several such housing 19a . 19b . 19c to a pillar 22 be releasably connected to each other. The pillar 22 has a stand at one of its two ends 24 on.

The present invention is not limited to the illustrated and described embodiments. Variations within the scope of the claims are also possible as a combination of features, even if they are shown and described in different embodiments.

LIST OF REFERENCE NUMBERS

1

Object detection system

2

object

3

room

4

camera

5

reference element

6

Camera lens angle

7

Virtual room model

8th

camera image

9

marker

10

template

11

pixel pitch

12

Real distance

13

Barrel arc

14

intersection

15

common point of intersection

16

overlap area

17

Space object

18

tape

19

casing

20

fastener

21

fastening area

22

pillar

23

dimensions

24

stand

α, β, γ, δ

viewing angle

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 20110025852 A1 [0003, 0008]
• EP 0940117 B1 [0004]

Claims (19)

Object detection system ( 1 ) for detecting the position and / or the orientation of an object ( 2 ) in a room ( 3 ) with a computer unit, a camera connected to it for the transmission of image data ( 4 ) and at least one of the camera ( 4 ) spaced reference element ( 5 ), wherein the computer unit is designed such that it the reference element ( 5 ) automatically in the camera image ( 8th ) the camera ( 4 ) recognizes and by means of an algorithm the position of the object ( 2 ) in a spatial coordinate system, characterized in that the camera ( 4 ) object-fixed and the reference element ( 5 ) is fixed in space and that the computer unit determines the position of the object ( 2 ) by means of at least two intersecting Fassfassbögen ( 13 ). Object detection system according to the preceding claim, characterized in that the object detection system ( 1 ) several spatially fixed and to each other, in particular in the x, y and / or z direction of the spatial coordinate system ( 3 ), spaced reference elements ( 5 ) and the position of at least one of these reference elements ( 5 ) in the spatial coordinate system ( 3 ) and / or its real distance ( 12 ) to at least one other reference element ( 5 ) stored in a memory of the computer unit and / or independently by the computer unit can be measured and / or calculated. Object detection system according to one or more of the preceding claims, characterized in that the computer unit the respective barrel circle arc ( 13a ) by means of a first and a second reference element ( 5a ; 5b ) and / or by means of a visual angle (α), wherein the real distance ( 12a ) of the two on the Fassbogenbogen ( 13a ) reference elements ( 5a ; 5b ) a chord of the barrel arc ( 13a ) and / or the visual angle (α), under which the tendon is seen from the object position to be calculated, the circumferential angle of the barrel arc ( 13a ). Object detection system according to one or more of the preceding claims, characterized in that the computer unit for determining the visual angle (α) the, in particular the shortest, pixel spacing ( 11a ) between the two reference elements ( 5a ; 5b ) from the camera image ( 8th ) and / or the visual angle (α) by means of the camera lens angle known to the computer unit ( 6 ). Object detection system according to one or more of the preceding claims, characterized in that at least two barrel arcs ( 13 ) lie in different and mutually inclined spatial planes of the spatial coordinate system. Object detection system according to one or more of the preceding claims, characterized in that the computer unit forms at least two intersecting tori, in particular spindle tori, each of which comprises a plurality of barrel arcs ( 13 ) are formed. Object detection system according to one or more of the preceding claims, characterized in that the barrel arcs forming the respective torus, in particular spindle torus, ( 13 ) by means of the two same reference elements ( 5a ; 5b ) are formed so that all these Fassbogenbögen ( 13 ) the same real distance ( 12a ) and equal angles of view (.alpha.), or include points from which, viewed from the point of view, the distance between the two reference elements appears at the same viewing angle. Object detection system according to one or more of the preceding claims, characterized in that the computer unit at least some of the intersection points ( 14 ) of the intersecting barrel arches ( 13 ) and / or Tori, in particular Spindeltori, determined and / or from these intersections ( 14 ) the one common point of intersection ( 15 ) whose position in the spatial coordinate system determines the position of the camera ( 4 ) in the room ( 3 ) corresponds. Object detection system according to one or more of the preceding claims, characterized in that the one common point of intersection ( 15 ) all those Fassbogenbögen ( 13 ), which in one the object ( 2 ) having spatial plane of the spatial coordinate system, and / or all formed tori, in particular spindle torii, have. Object detection system according to one or more of the preceding claims, characterized in that the computer unit for determining the first and second coordinates of the object ( 2 ) in the Spatial coordinate system the common intersection ( 15 ) at least two, in particular three, in a first the object ( 2 ) containing spatial plane of the spatial coordinate system lying Fassfassbögen ( 13a ; 13b ; 13c ; 13d ; 13e ; 13f ). Object detection system according to one or more of the preceding claims, characterized in that the computer unit for determining the third coordinate of the object ( 2 ) in the spatial coordinate system the common point of intersection ( 15 ) at least two, in particular three, in an inclined to the first space plane, in particular to this orthogonal aligned, the second object ( 2 ) containing spatial plane of the spatial coordinate system lying Fassfassbögen ( 13g ; 13h ; 13i ). Object detection system according to one or more of the preceding claims, characterized in that the computer unit has a multiplicity of determined points of intersection ( 14 ) and the one common point of intersection ( 15 ) within a tolerance range. Object detection system according to one or more of the preceding claims, characterized in that the camera ( 4 ) by means of a holder detachably on the object ( 2 ), in particular a human head or hand, can be fastened in a fixed position. Object detection system according to one or more of the preceding claims, characterized in that the object detection system ( 1 ) several cameras oriented in different directions ( 4a ; 4b ; 4c ; 4d ) and / or has an acceleration sensor connected to the holder. Object detection system according to one or more of the preceding claims, characterized in that the holder is a head-based vision system for detecting the position of at least one eye, wherein the camera ( 4 ) is preferably designed as a field camera of the gaze detection system. Object detection system according to one or more of the preceding claims, characterized in that the reference element ( 5 ) in the room ( 3 ), in particular on a room wall and / or a room item ( 17 ), attachable markers ( 9 ) and / or already in the room ( 3 ) space object ( 17 ) is and / or as a part of this, in particular as a defined reference point, preferably vertex and / or center is formed. Object detection system according to one or more of the preceding claims, characterized in that a plurality of markers ( 9 ) on one in the room ( 3 ) attachable wallpaper and / or a removable and / or roll-up tape ( 18 ) are arranged, which preferably from one another in the longitudinal direction of the wallpaper and / or the band ( 18 ), in particular equidistant and / or to the respectively adjacent marker ( 9 ) in one of the computer unit known real distance ( 12 ), are spaced. Object detection system according to one or more of the preceding claims, characterized in that a plurality of strips ( 18 ) to a room ( 3 ), in particular n-angular, reference element enclosure ( 5 ) are connectable. Examination arrangement for visual disturbances, in particular strabismus, with a, in particular virtual, Tangententafel and a means for measuring and automatic evaluation of the eye position and / or orientation relative to the Tangententafel, characterized in that the examination arrangement for detecting the head position and / or orientation of an object detection system ( 1 ) according to one or more of the preceding claims.

Images