DE202019003042U1 - Arrangement of modules for 3D-based optical inspection of pipelines and ducts - Google Patents

Abstract

Arrangement of modules for 3D-based inspection of pipes, channels or shafts in sewer networks by means of mobile detection unit of image data (3D image acquisition, 2D image acquisition) and evaluation via system control (consisting of 4, 6, 7, 8, 9, 10, 11 ) characterized,
for inspection, the detection unit (2), which is equipped with a 2D / 3D detection module with integrated lighting (5) and camera technology and placed on a driving module (3), data is optically detected and thus linking the image data with the spatial data he follows,
wherein the detection unit (2) placed on the driving module (3) interacts with the 2D / 3D detection module (5) spatially separated from the control and evaluation unit of the system controller (4) by being wirelessly or wirelessly coupled, the modules the arrangement performs a position feedback and position detection for location-accurate detection and installation of damage data via suitable sensor technology,
wherein the image data are evaluated separately or combined according to spatial data and images,
wherein the arrangement via the system controller (4) and the operating and visualization unit (9) is operated as an interface (interface) by the linkage of the object information of the damage is at least partially automated,
wherein data of geometrical conditions are determined from the spatial image data with the module height profiles (6),
wherein the 2D color data (2D image data processing) module (7) classifies the 2D data in terms of color and shape,
wherein the objects found by means of shape and color-based object detection (8) are provided with limiting markings and / or labels,
wherein, starting from the detected objects, features are extracted with which the objects are classified on the basis of the patterns of a pattern database (11),
wherein the pattern database (11) is referenced to the damage class database (10), and
wherein in the operating and visualization unit (9) the color image data, the object borders, with labeling and / or the height profile geometrically referenced to the classification of the pipe or channel state of damage, in particular based on deviations to the new pipe state, primarily in unwound representation are displayed.

Description

The invention relates to an arrangement of modules for 3D-based optical inspection of shafts and pipes, in particular of non-accessible lines and channels.

The invention described here can be assigned to the technical field of spatial image-based inspection of tubular objects, in particular of pipe systems or sewers.

State of the art

The inspection of pipelines and canals is an important task to maintain the function of infrastructures. With inspections, the structural condition of larger shafts or canals can be explored. This is not possible with smaller cross sections. In addition, visually it may not be possible or reasonable to test tubes in addition to the space problems mentioned. Here then camera technology is used for driving. This technique is on the one hand usually hand-guided and the usual visual evaluation of driving is complex. The experience and constant attention of the executing employee plays a crucial role in detecting all the problems and providing them with the right diagnoses. In any case, site-specific documentation is essential for assessing the condition of the canal and planning rehabilitation measures.

For sewers inspection techniques are known, which allow a visual assessment of the structural condition. Thus already existing partially automated devices and systems exist which can perform an evaluation on color or black-and-white images by means of image processing methods and the classification with nearest neighbor classifier. The channel interiors are made sufficiently clear and obstacle-free prior to inspection with a high-pressure flush, so that vehicle-based systems, such as, for example, inspection robots ( DE 10 2004 028 741 B4 ) or the so-called channel spider ( DE 10 2008 027 477 B4 ) can be used for locomotion. These inspection systems for non-accessible channels have different structures, have different controllable motion drives and are equipped with cameras.

In the published patent application ( DE 101 22 397 A1 ), a method for assessing a sewer network and a vehicle is presented. In this case, a camera is used for the measurement of parameters of a channel network formed from pipes also used for checking, for example, with the detection of the location and size of cracks in the pipes. The corresponding method allows a variety of parameters via a measuring device with subsequent input to a computer system, where for a section of the sewer network specific, independent of the measured values parameters are given, such as a level caused by precipitation, flow rate or pH, etc The resulting state changes can be calculated and the results visually displayed.

The conventional methods start with special illuminations and lead to improved preprocessing, such as noise suppression, brightness compensation and subsequent edge enhancement, to images that are transformed into the desired binary image via a global or local threshold or values. This is possible for both known and unknown objects. All of these inspection methods rely on the projection of the jacket projection of the shaft or pipe to be examined. No systems are known that use digital linking of images of the channel to spatial information regarding the channel interior surfaces.

Technical task

The invention has for its object to provide an array of modules for 3D-based optical inspection of pipes, channels or shafts, with which the detection of damage of any kind based on textures and their spatial structures can be performed. The key approach is to improve the interpretability of structures and damage by integrating 3D information and the possibility of automated detection. This is clearly possible by referencing both imaging spaces and the position and orientation feedback. The basis for accurate damage detection is the traceability of the position and the orientation provided by the driving module.

The problem is solved with the features listed in the protection claim 1.

The dependent claims give advantageous embodiments again. In claim 8, the essential components of the arrangement are listed.

The invention proposes, by additionally acquiring 3D information on the one hand, to enable an analysis by means of simple, automatically evaluable status information and, on the other hand, to reliably ensure visual follow-up. With the combination of the detection method goes the Need to be able to distinguish, for example, adhesions of corrosion. With the determined data the actual conditions can be clearly judged.

• 1 die Erfassungseinheit in einem Kanalabschnitt und
• 2 ein Blockschaltbild mit den Modulen und Komponenten.

Below is the arrangement of modules by way of example with reference to the accompanying figures, which also disclose features essential to the invention will be explained. It shows

• 1 the detection unit in a channel section and
• 2 a block diagram with the modules and components.

In 1 the remote detection unit is shown schematically in a channel piece with branch. The representation in 2 shows the structure of the arrangement for sewer inspection.

The registration unit ( 2 ) with driving module ( 3 ) should detect the 3D geometry of the channel in addition to a possible uninterrupted driving and the usual functions of a video-based system. The camera is fixed and allows the acquisition of the 3D basic data in addition to the 2D image. To avoid occlusion, the recording takes place both in the direction of travel, as well as against the direction of travel. This is achieved by a suitable lighting and projection unit, which allows both the capture of color images, as well as the generation of 3D data allows. An automated classification of damage according to standard is made possible by the additional 3D geometry information, which is also provided to the maintenance planner. The lighting or projection devices are located on the driving module ( 3 ), which via the system control ( 4 ) is served.

For inspection, the channel ( 1 ) with an autonomously movable detection unit ( 2 ), by a driving module ( 3 ), driven on the inside. The acquisition unit is equipped with a 2D / 3D module ( 5 ) equipped with camera technology and integrated lighting for optical data acquisition. When driving through a channel, its surface is optically detected. The data acquired by a camera, if necessary with a wide-angle lens etc., are stored in a database and then processed with the aid of a classification.

The inspection trip takes place with the provided system control ( 4 ) spatially separated from the detection unit ( 2 ) and the driving module ( 3 ).

It is important that a distinction is made between spatial data, in particular 3D information, spatial surveying information or spatial image data and images, in particular visual analysis data or color image information, and these are evaluated separately or in combination.

The spatial data lead to the evaluation of geometric conditions on the inner wall surfaces, such as heights and volumes of, for example, adhesions, eruptions or erosion. From the spatial image data, the height profile module ( 6 ), eg the surface profiles of pipes or shafts.

In 2D image processing ( 7 ), the data can be color-classified via color information or preclassified form-related via the object border. The with the 2D image processing ( 7 ) are detected by means of shape-based object detection ( 8th ). They are eg provided with limiting markings and / or labels. Thus, a classification based on the height information, the outlines or by means of labeling. Additional object levels of interest can be highlighted, uninteresting ones suppressed.

In combination of the available image data, ie images (2D data), and the spatial data (3D data), the detected damages are differentiated according to their damage patterns. The preclassifications from the 2D color images are verified with the surface profile data from the 3D data and a final classification follows. Characteristics are extracted and the classification of the objects based on the patterns of a sample database ( 11 ) carried out. The sample database ( 11 ) serves as a reference for the class database ( 10 ) of the damage classes.

The classification of 2D image data and the classification of the objects can also be done interactively by means of linking, whereby a simplified processing and completion of the sample database ( 11 ) is achieved.

In the operating and visualization unit ( 9 ), also referred to as a visual user interface, a summary of the color image data, further the object borders, the labeling and the recorded height profiles for the classification of the deviations to the new tube. Here, the display of the recorded image data is optically equalized and preferably in unwound form. The individually generated information, such as object border, labeling and height profile can be individually faded in or out for visual assessment.

The arrangement is made via the system control ( 4 ) and the operating and visualization unit ( 9 ) served as an interface by the Linkage of the object information of the damage is at least partially automated. The arrangement of modules provides interactive access to create new supplementary patterns based on real damage.

2 shows in a block diagram with acquisition modules and evaluation part of the principal processing chain for the sewer inspection. On the one hand, it is based on color images and on the other on the acquired spatial information, which is considered to be referenced to one another and can be classified according to normatively prescribed damage types.

The particular advantage of the solution according to the invention is that in determining the states of e.g. Sewer channels can be supported in addition to the evaluation of color image information on spatial survey information. Thus, it is possible to increase heights and volumes of e.g. To assess buildup or breakouts and abrasions. The arrangement thus allows a more robust automated channel state determination, coupled with an improved provision of the database for repair planning.

LIST OF REFERENCE NUMBERS

1

Channel with branch (section)

2

Detection unit (driving module ( 3 ) with illumination and 2D / 3D acquisition module ( 5 ))

3

Movement unit Driving module with position and position sensors

4

control Panel

5

2D / 3D acquisition module with integrated lighting

6

Module height profile (3D image processing)

7

Module color image processing (2D image processing)

8th

Mold and color based object detection

9

Operating and visualization unit

10

Database damage classes

11

Database pattern

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 102004028741 B4 [0004]
• DE 102008027477 B4 [0004]
• DE 10122397 A1 [0005]

Claims (8)

Arrangement of modules for 3D-based inspection of pipes, channels or shafts in sewer networks by means of mobile detection unit of image data (3D image acquisition, 2D image acquisition) and evaluation via system control (consisting of 4, 6, 7, 8, 9, 10, 11 ) characterized in that for the inspection, the detection unit (2), which is equipped with a 2D / 3D detection module with integrated lighting (5) and camera technology and placed on a driving module (3), data are optically detected and thus a linkage of the image data takes place with the spatial data, wherein on the driving module (3) placed detection unit (2) with the 2D / 3D detection module (5) locally separated from the control and evaluation unit of the system control (4) acts by wired or wireless is coupled, wherein the modules of the arrangement via suitable sensor technology performs a position feedback and location detection for accurate detection and installation of damage data, wherein the image data are respectively separated or combined according to spatial data and images are evaluated, the arrangement via the system control (4) and the control and visualization unit (9) is served as an interface (interface) by making the linkage of the object information of the damage at least partially automated is, are determined from the spatial image data with the module height profiles (6) data geometric conditions, with the module color image processing (2D image data processing) (7) the 2D data color and shape are classified, the means of shape and color-based Object detection (8) objects are provided with limiting labels and / or labels, based on the detected objects features are extracted, with which a classification of the objects based on the patterns of a pattern database (11), wherein the pattern database (11) on the Damage class database (10) referenced is and in the operating and visualization unit (9) the color image data, the object borders, with labeling and / or the height profile geometrically referenced for the classification of the pipe or channel state of damage, especially related to deviations to the new pipe state, displayed primarily in unwound representation become. Arrangement of modules for 3D-based inspection of pipes, channels or shafts Claim 1 characterized in that the 2D / 3D detection module with integrated lighting (5) is equipped with one or more lighting or projection devices and is integrated in the driving module (3). Arrangement of modules for 3D-based inspection of pipes, channels or shafts Claim 1 and 2 characterized in that the images obtained from the image data and 3D spatial data are preferably displayed visually equalized. Arrangement of modules for 3D-based inspection of pipes, channels or shafts according to at least one of the preceding claims, characterized in that the preparation of the shape-based and color-based object detection (8) additionally highlights interesting object levels and / or suppresses uninteresting. Arrangement of modules for 3D-based inspection of pipes, channels or shafts according to at least one of the preceding claims, characterized in that a spatial and optionally a measurement of the extent of defective areas of the pipe or channel is optional. Arrangement of modules for 3D-based inspection of pipes, channels or shafts according to at least one of the preceding claims, characterized in that the classification of the processed images on the basis of the height information, the outlines and / or the labeling of the detected objects. Arrangement of modules for 3D-based inspection of pipes, channels or shafts according to at least one of the preceding claims, characterized in that the construction of the classifier (classification of the features) or the classification of the objects is carried out interactively by means of linking, whereby a facilitated processing and / or completing the pattern database (11). Arrangement of modules for 3D-supported inspection of pipes, channels or shafts, comprising a detection unit (2), a movement unit driving module with location and position sensor (3) on which the detection unit is mounted, a 2D / 3D detection module with integrated lighting (5), a system controller (4) comprising a PC-based control and evaluation unit, a height profile module (3D image processing) (6), a color image processing module (2D image processing) (7), a shape and color based Object detection (8), a control and visualization unit (9), a damage class database (10) and a pattern database (11).

Images