DE202010017889U1 - Arrangement for recording geometric and photometric object data in space - Google Patents

Abstract

Recording arrangement for obtaining geometric and photometric object data with motor-driven and on target angle coordinates rotatable pan-tilt unit and a mechanically connected with this optoelectronic distance measuring device, characterized in that the pan-tilt unit is connected to a line scan camera so that these at least in one of whose axes of rotation is moved.

Description

Field of application and technical background

The invention relates to the adoption of features of real objects and scenes, such as their image and their spatial dimensions, in a three-dimensional computer model. The necessity of such computer models arises, among other things, in the production of deformations that are suitable for deformation as the basis for a subsequent configuration of reconstructions or expansions based thereon, but also in various simulations of reality for a variety of purposes or in securing volatile situations, such as. B. after accidents or crimes. The aim is to obtain the object features in an object-related form as quickly and accurately as possible, using both surveying techniques and photographic images.

State of the art

Photogrammetric recording methods are known that acquire geometric object information point by point from photographic images with the aid of further information, such as inner and outer orientation, control points or distances, or reference photographs. Characteristic of these methods is a high time-personnel effort for post-processing, which results mainly from the necessary Maßbestimmung or referencing all object points of interest in the figure.

An alternative to photogrammetry in many applications are laser scanners that record a dot cloud around the observation point with its polar coordinates in rapid succession. Each recorded point forms a data set of horizontal and vertical angles, as well as its distance to the observation point. Occasionally, the reflection behavior of the surface for the wavelength of the laser as non-panchromatic photography, or a superimposed photographic image is also stored. Each of the points measured in this way are immediately assigned the complete coordinates in the room. Nevertheless, the effort for the post-processing is not insignificant, because an assignment of the numerous points to individual objects is difficult to automate. Also, the effort is largely independent of the respective requirements on the complexity and accuracy of the result, so that the methodology is hardly scalable. Thus, compared to the large amount of information provided by the results, it has so far found only relatively little circulation and is only used by a few specialized working groups and engineering firms.

Tachymeters and total stations are known from the measurement technique, the individual points - mostly reflective measuring marks - aim by horizontal and vertical positioning of a riflescope and determine the distance to these by laser distance measurement. Some products are also equipped with a camera in the beam path of the optical viewfinder. The use of such systems is very effective in land surveying with individual points and with similar measurement tasks with high accuracy. But they offer few advantages for complex object modeling.

In particular, for various applications in construction and reconstruction and historic preservation specialized oversize systems are known, comprising a pan-tilt unit with precise adjustment or measurement of the pan and tilt angle and an attached optoelectronic distance measuring device for determining the distance between observation - and measuring point have. All known motorized Aufmaß systems are using an operating device such. B. a joystick (see GM DE202007010475 ), moved by manual control in the space points to be measured. Compared to a direct operation of the unit by hand, this avoids the fact that changes in position of the measuring system are caused and lead to measurement errors. However, the basic method of working manually, point by point in the object space, remains. The distance measuring device of known motor-driven oversize systems is a laser distance measuring device which generates a visible light spot in the object. This is used in addition to the distance measurement also to the bearing of the measured point to be measured. In addition to the simple visually controlled bearing (see GM DE202007010475 ) also come telescopic sight (see OS DE19648626 ) or a camera with a telephoto lens. Thus, the previously known arrangements thus follow the methods already used in conventional tachymeters without any new methodological approaches which would result in a more effective recording or measurement.

An additional possible use of the visible light spot is the marking of geometrically known spatial points, for example, to mark holes (see GM DE202007010475 ). For this purpose, it is necessary to previously refer to the location and orientation of the measuring system by targeting at least three known spatial points. The OS DE10308525 describes a combination of motorized survey system and digital array camera. The camera is arranged to the laser distance measuring device, that the effective beam paths are coaxial, parallel, lying in common plane or arbitrary. The imaging of the Survey scene with the camera follows a survey of relevant points in space with a laser distance meter. All measuring points are manually targeted in a known manner and the measured values obtained are linked to selected pixels. As a result, in addition to the positional relationship between the camera and the distance measuring device, one obtains a positional relationship between the geometric model and the camera image-but unfortunately only through a complex assignment by the operator.

Known motorized measuring systems are connected to powerful computing and operating devices, which are often designed as a laptop PC. The data connection is either wired or radio-based and realized using common interfaces and data protocols, such as USB or Bluetooth (see OS DE19648626 or GM DE202007010475 ) Due to the high computing power and good rendering of graphics, CAD functionality is increasingly being transferred to such operator consoles in order to interpret the acquired measurement data directly at the point of recording (see OS DE19648626 ).

The GM DE202007010475 indicates a way of working to simplify the drawing work. Here, before the measurement of a point in space, its membership in a straight, curved or broken line can be specified. This should be particularly advantageous for body edges. However, this is only very limited use because body edges are usually associated with spatial discontinuities that make a distance measurement using laser distance meter very difficult.

In summary, it should be noted that, despite the known variety of solution variants of the effort to take over real objects in a computer model is still very high.

Technical task

The invention has for its object to design a recording arrangement that allows a time-saving and accurate acquisition of real objects in a computer model. After leaving the location, no significant reworking for modeling is required. The total workload required to mount a model project will depend on the scope, complexity, and accuracy requirements of the project, so lower model complexity requirements will also result in lower costs.

Problem solving, description of the invention

The problem is solved with the invention characterized in the main claims. Advantageous embodiments are specified in the further claims.

The following terms are used in the following description:
The pan / tilt unit is a mechanical assembly that allows the pivoting and tilting of attached instruments. In the narrower sense, this assembly allows either the motor movement in a specified by horizontal angle and vertical angle target position, or the manual movement in such a position and subsequent determination of the two angles - in the strict sense by electronic scanning and transmission. A rangefinder is technically advantageous as an opto-electronic distance measuring device, for. B. as a laser distance meter, realized and has in the strict sense, a data interface for transmitting the measurement results. If this is mechanically connected to a pan-tilt unit, the result is a coordinate measuring machine according to the known prior art, which has already been described here as an oversize system. If one speaks here of the 'approaching' of a position, then it means the motoric movement into predetermined angular positions, in order subsequently to carry out a distance measurement in this position. As a camera, an electronic imaging device is referred to, which consists of a lens and a photosensitive transducer matrix, for. B. as a CCD or CMOS array, and generates an electronic image signal. The term nodal point is understood to mean the assumed projection center on the optical axis of the camera through which approximately all projection beams pass. This point is located in many optical constructions within the lens, but can also be in front of or behind it. A panoramic image is the wide-angle image of the surroundings of an observation point, which is not limited to the imaging limits of the 180 ° planar central projection, but the entire definition range of the polar coordinate system of horizontal 0 ... 360 ° and vertical -90 ... +90 ° can reach as an imaginary projection on a full sphere. A computer is understood to be an electronic computing device that consists of at least one control unit, an arithmetic logic unit, a program and data memory, and data input and output units, both as a data interface and as a user interface. Furthermore, the term computer also implies its operating system and suitable application software. The computer may be part of the control unit of the recording system, this may include completely or be connected to an external control unit via data interface. As a display, any type of electronic image display device, e.g. As an LCD screen, understood as a pointer, a graphical pointing device, eg. Legs Computer mouse, is called, which allows the pinpoint selection of a pair of coordinates in the image area. The term joystick describes an input device that, like a pointer, acts on two independent coordinates, but does not select target coordinates, but acts on movement speeds or accelerations in these directions. As a keyboard, any type of input device for alphanumeric data is called a user interface. So z. As well as a keyboard shown on the display, which can be operated with a pointer, take over the role of the keyboard. The term object is used in two different ways: First, as a designation of the original in the actual space. In this context, the terms 'object space', 'surface points in object space' etc. are used. On the other hand, this term is taken from the object-oriented classification and refers here to model objects as elements of the model space with fixed properties, eg. For example, surfaces, cylinders, etc. Therefore, the term 'model object' is used in this context. Other terms introduced in the specialist literature for object-oriented classification or programming, such as class, instance and property, are used here in the same conventional sense.

Basis of the invention is the idea to construct the computer model already at the recording location, similar to a CAD drawing program, and to use all available information that can be obtained automatically by technical means, ie without direct user input from the object space. This information is the polar coordinates of selected surface points in the object space and, as far as their inclusion is provided, the image information. All of this information should be able to be recorded from a single point of observation. However, this does not exclude that for larger models, several sub-models thus obtained can be connected to each other via connection points. In contrast to known methods, in which either only individual points or lines are collected, but here first spatial model objects, z. For example, plane or curved surfaces, cylinders or cuboids, are declared abstract and then assigned to these model objects measuring points in the object space whose coordinates are then determined by the measurement. If it is possible and meaningful also differentiated to which object areas the respective point belongs, so z. B. to the circular or lateral surface in a cylinder. In the simplest case, a pivot-tilt unit with attached rangefinder is required to determine the point coordinates of interest in the object space. Selecting the angular coordinates in which the distance is to be measured, is done in the simplest case by direct manual control of the movement of the pan-tilt unit while observing the position of the dot of the rangefinder. More convenient options will be described later. Depending on the degree of completeness of the object points determined in this way, the computer either determines entire model objects or parts thereof with their dimensions and their position in space. It may also happen that the information is not yet sufficient for a complete determination. In this case, the model object is first defined as far as possible for the current level of information. In order to provide the best possible support to the operator in his work, the model objects are always displayed as far as is known from the point of view of the viewing point on the display. It can thus quickly recognize which information gaps still exist and should be closed by further point coordinate measurements. Moreover, to obtain a better overview, it is also advantageous if the observer can disengage his perspective from the actual observation point and move freely in the virtual representation of the object space.

By the relation of different objects to each other in the room additional information is gained. If z. For example, if a wall surface is first measured and described using the methods of analytical geometry from at least three measuring points as an infinite surface, then it is already bounded on one side with the next adjacent wall surface obtained in the same way, whereby an edge is defined as a further model object. Already here, the advantage over the attempt to determine this edge directly as a line, visible: The model objects arise immediately from the measured coordinates in the three-dimensional object space, the distance measurements are basically not for vertices or edges, but for surface points. In this case, greater uncertainties in the exact selection of the measuring point coordinates can be tolerated, as long as they are only within the declared surface areas. Objects that are more closely related to each other can be grouped into a more complex object. To facilitate the work here, a structured object tree can be created in the application program, which also allows a hierarchical grouping of model objects. Furthermore, it is advantageous to be able to individually turn on or off objects and groups in the presentation. What is important is that an interactive and largely intuitive workflow is created, in which alternately new measurement points are determined and measured and immediately the new information obtained for the extension of the three-dimensional model are evaluated and the current result is displayed, after which again Further measuring points are strategically selected, etc. In addition to the manual selection of the measuring points described so far, the determination of further automatically defined points is also advantageous. Thus, by automatically small variation of the angle coordinates, measuring points in the immediate vicinity of the selected point can additionally be determined in order to achieve a higher measuring reliability or to determine their normal vector in small surface areas. For large areas, intermediate coordinates can be checked automatically. B. to determine curvatures, and so on. All these additional measurements can advantageously be carried out automatically whenever the system has waiting times until the next user input.

• 1. Der zu erfassende Objektbereich ist rund um den Beobachtungspunkt fotografisch abzubilden. Am universellsten ist hierfür eine lückenlose Aufnahme.
• 2. Die Bildkoordinaten der fotografischen Abbildung müssen den Horizontal- und Vertikalwinkeln eindeutig zugeordnet sein.
• 3. Die Koordinaten-Messeinrichtung mit Entfernungsmesser muss im gleichen Koordinatensystem betrieben werden, wie die Bildaufnahmevorrichtung, also auch mit identischem Koordinaten-Ursprung, und für jedes vorgegebene Horizontal- und Vertikalwinkel-Paar den Abstand zwischen Beobachtungspunkt und Objektpunkt messen können.

A significant gain in information is achieved if the recording system is suitable in addition to image acquisition, in such a way that the following conditions are met:

• 1. The object area to be detected is to be photographically photographed around the observation point. The most universal for this is a complete recording.
• 2. The image coordinates of the photographic image must be uniquely assigned to the horizontal and vertical angles.
• 3. The coordinate measuring device with rangefinder must be operated in the same coordinate system as the image pickup device, so also with identical coordinate origin, and for each given horizontal and vertical angle pair can measure the distance between the observation point and object point.

With these assumptions, the two-dimensional photographic image around an observation point can be used in a particularly comfortable manner. Because the image coordinates have a direct relation to the angular coordinates in the object space, a selection of the desired measurement coordinates in the object space can be made via the selection of image coordinates by means of pointers against the background of the image display. By using a variable magnification of the image display, the measurement points can be placed very precisely in the image details without the operator having to leave his seat. Because the image display and the display of the resulting model space refer to the same observation point, both can be superimposed and displayed simultaneously, which further improves the clarity of the spatial representation. Semitransparent false color display and selective display of individual elements allow a very concentrated and productive work with low risk of error. The above-described procedure for object-related data acquisition is also practiced with the aid of the camera image in the same way, except that instead of the light spot in the object space, the pointer on the display serves for a precise specification of the measurement coordinates. If necessary and in special cases, manual positioning by means of a light spot is still possible.

According to the invention, all of these requirements are met by a recording system which, on the one hand, contains a calibrated subsystem for panoramic image recording and, on the other hand, a rangefinder which can be positioned in polar coordinates.

The recording of a gapless panorama image is carried out by segmenting the spatial angle to be detected around the observation point and individually photographing these segments. For this purpose, a recording camera in the observation point to rotate stepwise to produce a series of individual shots. It may be sufficient, the camera in only one axis, z. B. horizontally, and to obtain a sequence of images in the form of a band which is annularly closed in the case of a 360 ° recording. If the camera does not capture a sufficiently large picture angle, the tape is too narrow and the sequence of pictures must be repeated with different angles of inclination in order to place several such bands on top of each other. This is achieved when the camera can be tilted and tilted not only in one, but in two mutually perpendicular axes of rotation. These motion functions are achieved with a pan-tilt unit on which the camera is mounted. In order to control these and to be able to travel in any angular positions, they must be motor-driven and sufficiently precisely positioned with electrically transmitted control instructions.

The following is based on the widespread arrangement in which is pivoted about a horizontal angle and then tilted in this position by a vertical angle. This does not mean a chronological sequence - the two independent movements can run in any order or at the same time. Rather, it is about the fact that the position of the one axis of rotation is changed by the rotation of the other, while a reverse causality does not exist. This arrangement is consistent with the commonly used polar coordinate system as well as with widely used astronomical and geodetic instruments, and is most useful in most cases. The vertical axis about which the horizontal rotation takes place is referred to as the horizontal axis and the axis of rotation for the vertical rotation as the vertical axis. This general situation assignment of the pan-tilt system described is not fixed but arbitrary. Thus, the subsequent references to horizontal or vertical layers also transferable to differently oriented layers of the device mutatis mutandis.

An important prerequisite for the later merging of the individual images into an overall panorama is that the axes of rotation and the optical axis of the camera have a common point of intersection and that the plane of the objective nodal point is likewise located at this point of intersection. Only then is a correct transition from one frame to the next without parallax error possible. The design of the pan-tilt unit must therefore meet these requirements. The merging of the individual shots into the overall panorama is a well-known process under the term "stitching" in panoramic photography, for which numerous software products are customary in the market.

As has been shown, accurate calibration of the photographic image in horizontal and vertical angles is an important condition for the process. In addition to the calibration and equalization of the camera image by known photogrammetric methods, there is a particularly interesting method in cooperation with the pan-tilt unit provided in the arrangement. For this purpose, optical targets are imaged in a series of different precisely set pan-tilt angles and thus an exact association between the pixel positions of the camera image and the associated angles of incidence is established. Because of the extraordinarily high number of pixel positions of the area sensor of such a camera - according to current practice, this can already be several tens of millions - but in reality only a rough survey will be possible, with the help then only an approximated two-dimensional assignment or equalization done can. Therefore, the geometric deviations will generally exceed the physical pixel resolution of the image sensor, resulting in overlap errors between the frames. This problem is well known from the practice of panoramic photography and is solved by automatic adjustments of the transitions, which are hardly visible visually, but are not correct in the geometric sense. Depending on the project requirements, the accuracy may nevertheless be sufficient.

As a more precise solution variant, a line scan camera is used according to the invention. This term is understood to mean an electronic camera with an imaging objective which, instead of the otherwise two-dimensional matrix-type image recording sensor, has a one-dimensional line or line image acquisition sensor, also called a line sensor. This line camera is also to be fastened with its nodal point at the intersection of the axes of rotation of the pan-tilt unit. First let the line sensor be aligned parallel to the horizontal axis of rotation. Now, if the horizontal drive is rotated stepwise or continuously and at the same angular distances read out the signal of the line sensor and assembled row by row to form a picture, the result is the cylindrical image or development of the environment around the observation point. This imaging method is known by specialized electronic panorama cameras, also called panorama scanners. By means of the vertical drive of the pan-tilt unit or a similar construction and a brand, for. As a simple black and white transition, it is now possible to assign each of the linearly arranged sensor pixels to a precise angle of incidence and permanently store this assignment as a one-dimensional calibration data. Technical line sensors have a number of pixels up to about 10,000, so that this task can be realized in practice with a manageable small calibration file. If the pixel positions with the assignment known therewith are assigned to the vertical angle positions, the cylindrical image becomes the image of a spherical ring segment. This ensures that the coordinate assignment in the vertical direction with an accuracy, which is predetermined by the number of pixels per angular range or the positioning accuracy of the vertical drive of the pan-tilt unit. In technical systems both accuracies have comparable orders of magnitude and should be brought into agreement in the proposed arrangement advantageously.

In principle, it is possible to produce a line scan camera with a recording angle range of 180 ° and thus to be able to image a complete sphere solely by horizontal rotation. Due to a variety of technical reasons, this should be avoided. Instead, it is useful to provide a smaller angular range and tilt the line camera with the pan-tilt unit not only horizontally but also to tilt vertically to record several ball ring segments on top of each other and then combine to form a complete ball. Since all pixels of the line scan camera are calibrated in angles of incidence, only the angle of inclination of the pan-tilt unit is added as an offset, whereby the processing effort remains small.

The closer you get into the vertex area of the polar coordinate system when you increase the vertical angle, the closer the horizontal angle steps - comparable to the longitudes in the geographic grid - are closer to each other. Therefore, it is advantageous without a reduction of the object-related image resolution, in this area the Increase angle increments between the line captures, which reduces redundancy and shortens recording time.

Line sensors for the reproduction of colors are generally designed so that a plurality of similar lines with usually the same number of pixels, but different spectral sensitivity, are arranged in a small distance parallel to each other, resulting in multiple color channels, eg. B. for red, green and blue, build. Because of the non-identical locations of the different color lines, the angle assignment - not only horizontally but also vertically - must be done individually for each of the individual lines, which multiplies the effort but only with the number of channels.

The use of line cameras for panorama image acquisition generally has two disadvantages compared to the recording and merging of flat individual images: On the one hand, any movements of objects in the object space during the scan movement lead to image artifacts. For the method provided here, however, it is anyway, at least temporarily, to start from immutable positions of the elements in the object space, because otherwise also the coordinate measurements would have no meaning. On the other hand, there is a longer recording time, because in the same time in which a matrix camera exposes thousands of lines simultaneously, only one line can be exposed here. However, the practical time required for recording is far from this ratio, because more time for saving the image and for the movement of the camera is to be considered in the next position and because the computational effort for low-fusing merging of the images is not insignificant. In practical use, the total time required to produce the panoramic image for both recording methods is comparable.

A special consideration deserves the illumination of the image recording area, if an artificial lighting is required. While in the simplest case, a well-known photo flash light can be used for single-frame recording, there is no energy-adapted solution for shooting with line scan cameras. Here is a synchronized with the line recording frequency strobe light with line-shaped focusing on the recording area of advantage, which also gets by with a relatively low flash output because of their small projection area. Especially light-emitting diodes with approximately point-shaped light exit surface, which are arranged linearly and are focused with a trained in only one direction parabolic reflector or a cylindrical lens system, are suitable for this purpose as a light source. In this case, one or more light sources can be equipped with a focusing system or several such systems, preferably on both sides of the camera, complement each other.

• – Beide Funktionseinheiten sind auf der Schwenk-Neige-Einheit gegeneinander austauschbar – entweder durch vollständigen manuellen Austausch, indem die eine Einheit entfernt und durch die andere ersetzt wird, oder durch eine mechanische oder elektromechanische Umschaltvorrichtung, die zu einem Austausch der Positionen führt, z. B. in der Art eines Revolvers.
• – Beide Funktionseinheiten befinden sich gleichzeitig auf der Schwenk-Neige-Einheit. In einem oder beiden Objektivstrahlengängen befindet sich eine optische Umlenkvorrichtung in Form eines oder mehrerer Spiegel oder Umlenkprismen, die beide Objektivstrahlengänge in Objektrichtung betrachtet in eine gemeinsame optische Achse umlenken. Diese Umlenkvorrichtungen können z. T. halbdurchlässig sein, um ständig für beide Funktionseinheiten verfügbar zu sein. Dies hat den Nachteil, dass die Lichtempfindlichkeit der Kamera und die Reichweite der Entfernungsmesseinrichtung reduziert werden. Weil eine ständige Verfügbarkeit nicht nötig ist, kann die Umlenkvorrichtung auch in zeitlicher Folge umgeschaltet werden, so dass alternierend entweder die eine oder die andere Funktionseinheit, dafür aber mit voller Empfindlichkeit, betrieben wird. Als umschaltende Umlenkvorrichtung kommen beispielsweise Klapp- oder Drehspiegel in Frage. Eine besonders interessante Ausführungsmöglichkeit stellt die Verwendung eines Spiegels mit einem Diaphragma in dessen Mitte dar: Hierdurch ist es möglich, den ausgesendeten Lichtstrahl der Entfernungsmesseinrichtung durch dieses Diaphragma zu senden, dessen Lichtpunkt auf der Objektoberfläche aber über die übrige Spiegelfläche in den Strahlengang der Kamera umzulenken und diesen gleichzeitig mit dem Objekt mit abzubilden, um die korrekte Koordinatenbeziehung zwischen Messpunkt und Bildpunkt zu kontrollieren.
• – Die Funktionseinheiten sind so auf der Schwenk-Neige-Einheit befestigt, dass sie zwar einen gemeinsamen Koordinaten-Ursprung im Schnittpunkt der Drehachsen, jedoch unterschiedliche Drehwinkel aufweisen. Die Aufnahmekamera muss, wie bereits erläutert, nicht nur einen gemeinsamen Schnittpunkt ihrer optischen Achse mit den Drehachsen besitzen, sondern gleichzeitig auch eine vollkommene Übereinstimmung des Nodalpunktes mit diesem Schnittpunkt. Hierdurch ist die Lage der Kamera auf der Schwenk-Neige-Einheit eindeutig festgelegt. Da der Nodalpunkt in den meisten Fällen innerhalb der Objektive oder der Kamera liegt bedeutet das auch, dass der Raum um den Koordinaten-Ursprung mit der Kamera oder deren Objektiv belegt ist. Für die Entfernungsmesseinrichtung gelten hingegen einfachere Regeln für die Ausrichtung: Hier ist lediglich zu gewährleisten, dass die optische Achse durch den Schnittpunkt der Drehachsen verläuft und dass der Abstand zwischen Bezugspunkt der Entfernungsmessung und Koordinaten-Ursprung bekannt ist. Dadurch ist es möglich, die Entfernungsmesseinrichtung hinter oder neben der Kamera anzuordnen und deren Aufnahmerichtung abweichend von der Aufnahmerichtung der Kamera zu wählen, z. B. in entgegengesetzter Richtung oder rechtwinklig dazu. Durch die Schwenk-Neige-Einheit lassen sich die Winkelunterschiede einfach durch Wahl eines Offsets, z. B. von 180° oder 90°, kompensieren.

In order to operate the distance measuring device in the same coordinate origin as the recording camera, different approaches are possible:

• - Both functional units are interchangeable on the pan-tilt unit - either by complete manual replacement, by removing one unit and replacing it with the other, or by a mechanical or electromechanical switching device resulting in an exchange of positions, e.g. B. in the manner of a revolver.
• - Both functional units are located on the pan-tilt unit at the same time. In one or both lens beam paths there is an optical deflection device in the form of one or more mirrors or deflecting prisms, which deflect both lens beam paths in the direction of the object in a common optical axis. These deflecting devices can, for. T. semipermeable to be constantly available for both functional units. This has the disadvantage that the photosensitivity of the camera and the range of the distance measuring device are reduced. Because a constant availability is not necessary, the deflection device can also be switched in chronological order, so that alternately either the one or the other functional unit, but with full sensitivity, is operated. As a switching deflector, for example, folding or rotating mirror come into question. A particularly interesting possibility of implementation is the use of a mirror with a diaphragm in the middle of: This makes it possible to send the emitted light beam of the distance measuring device through this diaphragm, but redirect the light spot on the object surface on the remaining mirror surface in the beam path of the camera and Imagine this simultaneously with the object in order to control the correct coordinate relationship between the measuring point and the pixel.
• - The functional units are mounted on the pan-tilt unit so that they have a common coordinate origin at the intersection of the axes of rotation, but different angles of rotation. The recording camera must, as already explained, not only have a common point of intersection of its optical axis with the axes of rotation, but at the same time also a perfect match of the nodal point with this point of intersection. As a result, the position of the camera on the pan-tilt unit is clearly defined. Because the nodal point is in most cases within the lens or camera also means that the space around the origin of the coordinates is occupied by the camera or its lens. For the distance measuring device, however, simpler rules for the orientation apply: Here, it is only necessary to ensure that the optical axis passes through the intersection of the axes of rotation and that the distance between the reference point of the distance measurement and coordinate origin is known. This makes it possible to arrange the distance measuring device behind or next to the camera and to choose the direction of their recording deviating from the direction of the camera, for. B. in the opposite direction or at right angles to it. Due to the pan-tilt unit, the angle differences can be easily selected by selecting an offset, z. B. of 180 ° or 90 °, compensate.

Thus, the image acquisition can be sensibly coordinated with the distance measurement, the entire assembly described so far, consisting of pan-tilt unit, recording camera and distance measuring device, as well as the switching elements of the beam path and the light source, if any, must be functionally connected to a common control unit , This control unit must control the angular orientation of the pan-tilt unit in both axes of rotation, the camera and the distance measuring device and receive both the image data and the distance measurement results. For carrying out the method according to the invention, furthermore, a computer with display, pointer and keyboard is required. For this purpose, in particular a personal computer, z. B. suitable as a laptop or notebook. It may be advantageous if the task of the control unit is also taken over by the same computer, which also serves to process the image and measurement data. On the other hand, it may prove advantageous to switch further microcontrollers or microcomputers between the computers and the modules mentioned, which then undertake subtasks, such as the interpretation and execution of instructions, a preprocessing of data or the bundling of the data flow for a common effective data interface. In addition, such subsystems realize without problems the required real-time capability of sub-functions such. As the motion control and synchronized data acquisition.

• 1. Festlegung der Bereichsgrenzen für die Messung durch Nutzereingabe. Neben der direkten numerischen Eingabe der Grenzwinkel kann der Lichtpunkt des Entfernungsmessers als Zeiger im Objektraum dienen oder ein Kamerabild als Sucherbild verwendet werden. Hierzu soll die Schwenk-Neige-Einheit, z. B. über Joystick, frei positionierbar sein, während die ausgewählten Positionen direkt als Winkelmaße übernommen werden können. Erfolgt grundsätzlich eine Panorama-Bildaufnahme ohne Winkelbegrenzung, was am universellsten ist, so kann dieser Schritt entfallen.
• 2. Aufnahme eines Panoramabildes: Die Steuerungseinheit fährt die vorgesehenen Winkelkoordinaten der Schwenk-Neige-Einheit nacheinander ab und löst in den notwendigen Positionen die Bild- oder Bildzeilenaufnahme aus. Hierbei ist es nicht in jedem Fall erforderlich, dass die Bewegung während der Aufnahme unterbrochen wird – gerade beim Einsatz einer Zeilenkamera können die Einzelzeilen auch während der Bewegung belichtet werden.
• 3. Winkelrichtige Montage des Panoramabildes aus den einzelnen Bild- oder Bildzeilenaufnahmen. Wegen der vorhandenen Kalibrierung der Kamera in Winkelkoordinaten ist dieser Prozess automatisch, je nach Leistungsfähigkeit des Rechners auch bereits während der Aufnahme, durchführbar.
• 4. Anzeige von wählbaren Ausschnitten des aufgezeichneten Panoramas in wählbarem Anzeigemaßstab auf dem Display. Hierbei ist eine zentralperspektivische Entzerrung des dargestellten Bildausschnitts von Vorteil, um eine wirklichkeitsgetreue Bilddarstellung zu erreichen.
• 5. Auswahl von Modelltypen zur Zuordnung der Messpunkte, z. B. aus einer Werkzeugpalette, und Instanziierung neuer Modellobjekte.
• 6. Markieren einer zweidimensionalen Bildkoordinate mit dem Pointer, hiernach automatisches Positionieren der Entfernungsmesseinrichtung auf die zugehörige Winkelkoordinate im Polarkoordinatensystem, Durchführen einer Entfernungsmessung und Speichern des Datensatzes aus Horizontalwinkel, Vertikalwinkel und Abstand als Objektpunkt gemäß Vorauswahl.

Essential functions of recording devices according to the invention consist of the following and must be fully or partially ensured depending on the scope provided:

• 1. Definition of the range limits for the measurement by user input. In addition to the direct numerical input of the critical angle, the light spot of the rangefinder can serve as a pointer in the object space or a camera image can be used as a viewfinder image. For this purpose, the pan-tilt unit, z. B. via joystick, be freely positionable, while the selected positions can be taken directly as angular dimensions. If, in principle, a panorama image recording without angle limitation, which is the most universal, so this step can be omitted.
• 2. Recording a panoramic image: The control unit moves the intended angular coordinates of the pan-tilt unit one after the other and triggers the image or image line recording in the necessary positions. In this case, it is not always necessary for the movement to be interrupted during recording - especially when using a line camera, the individual lines can also be exposed during the movement.
• 3. Correct mounting of the panorama picture from the individual picture or picture line recordings. Because of the existing calibration of the camera in angular coordinates, this process is automatically, depending on the performance of the computer during the recording, feasible.
• 4. Display of selectable sections of the recorded panorama in selectable display scale on the display. In this case, a central perspective equalization of the displayed image section is advantageous in order to achieve a true-to-life image representation.
• 5. Selection of model types for the assignment of the measuring points, eg. From a tool palette, and instantiating new model objects.
• 6. Marking a two-dimensional image coordinate with the pointer, thereafter automatically positioning the distance measuring device on the associated angular coordinate in the polar coordinate system, performing a distance measurement and storing the data set of horizontal angle, vertical angle and distance as object point according to preselection.

The panoramic image on the display thus advantageously serves as a kind of "copy master" for comfortably "constructing" the spatial model by selecting the visible object points with the pointing device and then automatically determining the associated spatial coordinates by controlling the pan-tilt. Unit in the associated angular position and distance measurement takes place in this position. Once objects or parts of them are constructed, they are displayed in the same view projection, which allows the user to easily control the congruence of both representations in the image.

An automatic recording mode is provided, in which the angle coordinates for the distance measurement need not be specified by the operator, but at regular intervals over a line, curve or area is made, or generally a dot grid is recorded in the vicinity of the observation point. Especially for simpler object spaces, eg. As Rohbauten or empty spaces after de-coring, it is thus possible to perform quick shots without qualified user input on site and only later in the manner described by assigning points to make objects modeling.

It is a special working mode provided in which the rangefinder is driven by a motor on a pointer marked in the display image angular position and then marks this position in the object space with its luminous point, so that this for processing, for. B. for attaching a hole, can be touched. The panorama image or the projected object representation with marks can be used as orientation. Similarly, numerically predetermined coordinates can be controlled and marked in the object space.

A display mode is provided in which the center of the displayed image section is always synchronized to the current angular position of the rangefinder, so that the point just measured in the center of the image is.

It is intended not only to combine individual objects into groups, but also to subtract from each other, so it is very beneficial to breakthroughs by surfaces such. As windows and doors in wall surfaces, or cavities, such as holes, channels u. Like. Insert into such surface or volume elements after the object has been described first simplified.

There is provided a mode for automatic identification of target marks in the image by known methods, e.g. B. to find reference or registration marks automatically and to take their coordinates, or to measure. In this way, an automatic outer orientation of the measuring arrangement is given. Furthermore, such marks serve for the orientation of neighboring locations during connection measurements.

For fluent work, keyboard shortcuts are provided for quickly switching display scales and for quickly fading in and out of the image and the individual objects or object groups.

In addition to the described use of the panorama image as a "design template", it is furthermore provided to automatically evaluate image features according to known methods of image processing and to match these alternately with the objects created in the model space in order to independently, ie without direct intervention by the operator, obscurities or inaccuracies to clarify, to readjust object edges or to improve the richness of detail of the model. For the execution of such automatic functions preferably waiting times of the system are used for user input.

If objects are modeled in the work process that can also be reused in other projects, they can be stored in an object database and called up immediately by the application program as another object type.

It is intended to be able to copy objects that have already been made dimensionally or complex object groups and to be able to move, rotate or mirror them in space. These functions are for faster modeling of repetitive elements of the original. Belong z. For example, if a building to be modeled contains several similar columns with pedestals and stucco, only one of them will be completely modeled, while the others will be duplicated from the first one and with some additional coordinates to be measured as snap points. Here, too, image features can be used to additionally automate this process. This multiplication of object features is used even if a number of similar objects are present but not completely visible. As an example, a plurality of similar windows in a facade may be mentioned, which are shadowed in different perspective for a central observation point. Here, the characteristics of each single instance of similar objects are collected and then applied to all.

It is intended to specify globally or individually valid tolerance limits for the deviation of idealized objects by user input, which then check an automatically running plausibility check the membership of surface points to the respective object. If the shape and position of objects are overdetermined on the basis of the number of measured surface points, then depending on the test result, the object parameters are determined according to the smallest error square or the object is decomposed into elements which are then further specified individually using measured surface points. For this purpose, it is also provided to automatically approach and measure further surface points after resulting additional information need.

Whenever the rangefinder is approaching and measuring several measuring points at the same time are measured, the order is optimized according to the principle of the shortest paths.

Based on embodiments, the invention will be explained below.

They show by way of example:

1 the overall structure of a recording arrangement according to a first example,

2 the overall structure of a recording arrangement according to a second example,

3 the functional structure of a complete recording system,

4 the workflow on site when using a camera and

5 the process of the background process for the acquisition of the respective distance information for predetermined angle coordinates. Same position numbers in different drawings denote the same elements.

A first embodiment according to 1 shows a recording device without showing the computer connected to it, which is a personal computer in a known design. On a main frame 1 There is a precision drive which can be positioned at a specified angle of rotation 2 for the horizontal angle and a similar precision drive 3 for the vertical angle, the instrument panel 4 drives. In the mounting foot 5 , which serves to fix the entire assembly on a measuring stand, there is additionally a brake for locking the horizontal angle during the measurement. The entire structure is thus pivoted horizontally relative to a base plate, while the instrument carrier is inclined relative to the main frame. On the instrument panel are a line camera 6 , as well as an optoelectronic rangefinder 7 attached. The optical axes of both instruments are parallel to each other. About a mirror 8th and a half-transparent mirror 9 Both beam paths are in a common optical axis 10 passing through the intersection of the two axes of rotation of the drives 2 and 3 runs, united. In addition, the nodal point of the line camera is also exactly in this intersection - not shown here. The advantage of this arrangement over the second exemplary embodiment described below is that the image acquisition can take place simultaneously with the distance measurement, so that the light spot of the rangefinder can be imaged by the camera and selected by means of differential image reproduction with the laser switched on and off. In addition, this arrangement can lead to a more compact construction of the mobile apparatus part. In one embodiment of this embodiment, the semipermeable mirror 9 replaced by a normal mirror, but with a motorized folding or pivoting device, so that the beam path can alternatively be switched to the camera or the rangefinder. In addition to the line scan are LED flashing lights 11 shown with a cylinder optics, which serve in insufficient light conditions for strip lighting of the object space in the receiving area of the camera. The flashing lights are activated synchronously with the line recording of the camera. On the main frame are still a control unit 12 and a removable storage battery 13 attached to the central power supply of the device. The control unit provides the central interface to the computer not shown here and communicates via a local data bus with the drives, the rangefinder, the camera and the flashing lights. Alternatively, a wired communication with the computer via a serial data interface or wireless communication via the antenna 14 selectable. Each of the modules described has its own microcontroller system that receives commands from the control unit 12 automatically executes over the local data bus and returns information. In a second embodiment according to 2 is dispensed with deflecting mirror. Instead, the camera and rangefinder are mounted at a right angle to each other. As a result, fewer structural elements are needed than in the first example and no sensitivity or range limitation due to the mirror must be accepted. Due to an offset of 90 ° on the vertical angle, it is possible to drive the rangefinder exactly in the same angular position, from the previously performed the line scan camera. A similar arrangement is also possible if a matrix camera is used instead of the line scan camera. 3 describes the structure and the functional interaction of a recording system as an embodiment in accordance with the invention. It consists of a recording and measuring unit, which mechanically according to the embodiment 1 or 2 or a similar embodiment of the invention may correspond, and a control unit, for. B. in the form of a conventional personal computer, or laptop. For the high data volume of the camera, a separate data channel is provided here. 4 describes by way of example the workflow on-site during the recording using an inventive arrangement. First, in 21 the survey field photographed around the location of the surveying device and converted into a panoramic image. Each pixel position of this image corresponds to the polar angle coordinates of the mathematical geometric coordinate system, which is based on the total measurement, and in its position and alignment with the pan-tilt device coincides. In 22 For example, the user manually selects an object type that best describes geometrically the next real object to be captured in the measurement field from a set of predefined types, and thus adds an instance of that type to the virtual model. Subsequently, in 23 Defines points in the panorama image that belong to the surface of the new object instance. These points can both partially automatically due to causal, the algorithm already known, relationships or known methods of image processing, as well as by user input, z. B. by means of pointer defined. The newly generated points, which belong to the object instance, can be directly assigned to their respective polar coordinates - initially without radius - and are stored in 24 Added to a list of the space points to be measured by the range finder. This list is governed by another asynchronous background process 5 already processed during the selection of further points. If the user does not need to add further measurement points or object instances to the model, he must enter 25 Wait for the background process to complete the survey of all the coordinates listed in the list. If the survey is provisionally completed, in 26 the calculated model parameters compared with the tolerance specifications of the user. If these are not sufficiently fulfilled, further measuring points are added to the relevant object instances and measured. Were all real objects of the environment that are of interest to the user geometrically recorded ( 27 ), and meet all the user - specified tolerance limits, so in 28 the recorded photographic image broken down into segments by known methods according to the model geometry. These segments are assigned to the respective model surfaces as so-called textures which describe the color of individual surface points of the respective geometrically defined model surface.

5 describes by way of example the sequence of the background process for the acquisition of the respective distance information for predetermined angle coordinates. This process runs constantly and independently of the user. The predefined angle coordinates originate from the in 4 shown method, and are expediently in this embodiment in a data structure of the type queue, technically referred to as First-In-First-Out (FIFO) out.

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

• DE 202007010475 U [0005, 0005, 0006, 0007, 0008]
• DE 19648626 A [0005, 0007, 0007]
• DE 10308525 A [0006]

Claims (9)

Recording arrangement for obtaining geometric and photometric object data with motor-driven and on target angle coordinates rotatable pan-tilt unit and a mechanically connected with this optoelectronic distance measuring device, characterized in that the pan-tilt unit is connected to a line scan camera so that these at least in one of whose axes of rotation is moved. Recording system according to claim 1, characterized in that the optical axis of the distance measuring device can be brought into coincidence with the optical axis of the line scan camera simultaneously or with a time offset by - First, both optical axes cut the two axes of rotation of the pan-tilt unit in their intersection and - Second, either the distance measuring device is aligned with your lens at a known angle to the optical axis of the line scan, or both lenses are permanently or temporally offset by optical deflection devices aligned on a common direction of incidence. Recording system according to claim 1, characterized in that a light source with projection optics, which generates a linear and parallel to the line sensor of the line camera aligned light strip is connected to the line scan that it performs all pan and tilt movements of the camera together with this and their light projected into the same object area that is being imaged by the line scan camera. Recording system according to claim 1, characterized in that the axis drives of the pan-tilt unit, the distance measuring device and the line scan camera are directly or indirectly connected to a common computer, which further comprises a display, a pointer and a keyboard. Recording arrangement for obtaining geometric and photometric object data with motor-driven and on target angular coordinates rotatable pan-tilt unit, a mechanically connected to this electronic image recording camera and also mechanically mechanically connected with this optoelectronic distance measuring device, characterized in that a software integrated into these following processing steps realized, which can also run in chronologically or organisationally separate work stages or in a different order: 1) Declaration of model objects within a class hierarchy of object type, object instance and, if necessary, a point type to that object, equivalent to geometric elements in object space, by user input to a computer; 2) selection of a number of surface points on the elements in the object space by the user as a set of individual points or as a rule for such a selection; 3) Consecutive targeting of the selected surface points from an observation point with a calibrated in angular dimensions pan-tilt unit and thus mechanically connected distance measuring device and acquisition of the full polar coordinates, ie horizontal angle, vertical angle and distance in the computer while assigning these coordinates to the in 1) declared model objects, with the aim of classifying structured storage of the point coordinates; 4) calculation of dimensions and location of the model objects from their associated point coordinates as a representation of the surface points of the model objects; 5) Handling object correspondences, such as occlusions, penetrations, or cut edges, and modeling collections and scenes. Recording arrangement according to claim 5, characterized in that it is suitable in addition to the said processing steps for a true-angle image recording, in particular panoramic image recording, of the measuring range of interest from the observation point out as an entire image or partial image while maintaining a precise and unambiguous association between the Point coordinates of the image display and the angular coordinates in the object space is displayed on a display, so that the target angle for the coordinate determination in the object space can be done by selection with the pointer on the display, after which the motor-driven pan-tilt unit drove to this angular position and the distance can be measured from the observation point in this angular position. A recording arrangement according to claim 5, characterized in that photometric image information, such as gradation, discontinuities and edges, are used for automatically finding or controlling object features in order to achieve an automated data input, model correction and / or a reduced amount of required measurement data. Recording arrangement according to claims 5 and 6, characterized in that the image information obtained for true-to-scale texturing and photorealistic representation of the spatial model is used. A receiving arrangement according to claim 5, characterized in that mutually similar spatial Objects in the recording area can be duplicated and thus reused by copying, possibly also buffering, and adapting to the image template.

Images