Elevator track detection system based on projection grating

Abstract

The invention discloses an elevator track detection system based on a projection grating. The system mainly comprises a curved rail support, a grating projector, a camera, a photosensitive sensor, an LED lamp strip and a fragrant citrus pie processor. The curved rail support is installed on the top of an elevator car and right faces a rail. The grating projector is located in the center of the track support, and cameras are arranged on the two sides to capture track images. And the photosensitive sensor and the LED lamp strip provide constant illumination, so that the imaging quality and the three-dimensional scanning precision are ensured. The orange pie processor is located on the rear side of the curved track support and is responsible for collecting images shot by the camera and information of the attitude sensor and conducting data transmission with an upper computer through a UDP communication protocol. The system can quickly and accurately detect the flatness of the elevator track, predict track faults, perform real-time early warning and perform remote communication. The device is suitable for detection of various elevator tracks and has a wide application prospect.

Description

Technical Field

[0001] The present application relates to the field of image processing, and in particular to an elevator track detection system based on projection grating. Background Art

[0002] Traditional elevator track inspection methods mainly rely on manual operation, usually using contact measurement tools for point-by-point inspection. This method is not only time-consuming and labor-intensive, but the measurement accuracy is easily affected by human factors, such as the operator's skill level and attention. In addition, contact inspection often makes it difficult to achieve high-precision inspection and real-time monitoring of the entire track, and due to direct contact with the track surface, it may cause wear or damage, thus affecting the normal operation and safety of the elevator. In recent years, non-contact inspection methods have gradually been introduced into the field of track inspection. Non-contact inspection technology, with its high measurement accuracy, fast response and non-destructive characteristics, can accurately detect the three-dimensional morphology of the track. This method uses lasers, grating projection or other advanced sensors to quickly obtain a large amount of data without contacting the track, reducing errors caused by human intervention.

[0003] However, although the existing non-contact detection technology can provide high-precision measurement results, it still has certain deficiencies in real-time monitoring and fault prediction. The existing technology cannot effectively guarantee the timely monitoring of track status changes and fault prediction during elevator operation. When a potential fault occurs in the track, maintenance measures cannot be taken in time, which affects the safety and maintenance efficiency of the elevator. Therefore, it is urgent to further improve the detection technology to achieve more efficient real-time monitoring and fault prediction capabilities, and provide more reliable guarantees for the safe operation of the elevator. Summary of the invention

[0004] An elevator track detection system based on projection grating, comprising:

[0005] A grating projector 5 is installed on the top of the elevator car;

[0006] There are two cameras 2, which are respectively arranged on both sides of the grating projector 5, and are used to collect the grating information projected onto the elevator track;

[0007] A photosensitive sensor 4 is installed between the adjacent grating projector 5 and the camera 2;

[0008] LED light strip 1, irradiating the photo taking position of camera 2;

[0009] The Orange Pi 6 is electrically connected to the camera 2, the light sensor 4, and the LED light strip 1 respectively; the Orange Pi 6 is also connected to the host computer;

[0010] Furthermore, it also includes a posture sensor 3 installed on the top of the elevator car, and the posture sensor 3 is electrically connected to the Orange Pi 6;

[0011] The photosensitive sensor 4, the camera 2 and the LED light strip 1 are installed on the top of the elevator car through the curved rail bracket 8;

[0012] Further, the grating projector 5 is installed in the middle of the curved rail bracket 8, and the LED light strip 1 is installed at both ends of the curved rail bracket 8;

[0013] The photosensitive sensor 4, the camera 2 and the LED light strip 1 are movably connected to the curved rail bracket 8 via a C-shaped movable base.

[0014] Elevator rail defect detection is carried out in the following steps:

[0015] S1, controlling the brightness of the LED light strip 1 through the photosensitive sensor 4 to obtain a constant light field;

[0016] S2, under the condition of constant light field, obtaining a reference surface fringe image and a deformation fringe image through a grating projector 5;

[0017] S3, collecting the reference surface fringe image and the deformation fringe image through camera 2, and transmitting the collected images to Orange Pi 6 for preprocessing; the details are as follows:

[0018] S3.1, compare the filtering effects of three different Gaussian filters, and select the Gaussian filter with the best effect for filtering;

[0019] 3 different Gaussian filters include: Gaussian filter with filter size 5×5 and standard deviation 0.1, Gaussian filter with filter size 5×5 and standard deviation 0.5, and Gaussian filter with filter size 5×5 and standard deviation 1;

[0020] After comparison, the selected filter size is 5×5 and the standard deviation is 1. At this time, the burr peaks of the stripes at different levels of crests and troughs are significantly improved;

[0021] S3.2, normalizing the filtered image to obtain the projection grating fringe period;

[0022] S4, calculating the phase principal values ​​of the preprocessed reference surface fringe image and the deformation fringe image and subtracting them to obtain a wrapping phase;

[0023] S5, performing phase unwrapping by diamond unwrapping;

[0024] The steps for diamond unwrapping are as follows:

[0025] S5.1, determining expandable points and non-expandable points in the phase diagram after subtraction;

[0026] Expandable points are those points that are not wrapped, have no phase jumps or errors, and the phase difference between adjacent points does not exceed 2π; there are no poles, noise or interference, and their true phase values ​​can be restored by phase difference calculation;

[0027] Unexpandable points refer to poles, noise areas, and edge areas; there are jumps in phase values;

[0028] The determination method is as follows:

[0029] S5.1.1. Detect the edge area of ​​the image by edge detection algorithm and mark it as unexpandable point;

[0030] S5.1.2, using the local variance calculation method to detect the noise area in the image and mark it as an undevelopable point;

[0031] S5.1.3, by calculating whether the absolute value of the phase difference between adjacent pixel points is greater than a preset value t, t∈(-π,π), if it is greater, it is an unexpandable point;

[0032] S5.1.4, by removing all the unexpandable points, the remaining points are all expandable points;

[0033] S5.2, take the expandable point as the initial point, and expand to the four neighbors in turn if the initial point satisfies the condition that all its four neighbors are expandable points;

[0034] S5.3, during the diamond expansion process, the phase of each point is unwrapped;

[0035] The unwrapping process is achieved through an integral path, which starts from the initial point and gradually expands along a diamond trajectory to the entire image to complete the unwrapping;

[0036] S6, converting the phase information after phase unwrapping into height information and performing three-dimensional reconstruction;

[0037] S7. Perform error analysis on the three-dimensional reconstruction image to obtain a final three-dimensional reconstruction image to complete the detection of the elevator track.

[0038] Furthermore, the elevator track detection system based on projection grating of the present application adopts different detection modes and performs corresponding detection according to the operating status of the elevator to ensure the accuracy and real-time nature of the detection results.

[0039] Elevator status detection:

[0040] The system first detects the running status of the elevator, which is divided into two situations: no-load or low-speed operation and high-speed operation. The identification of the elevator status is achieved through real-time data analysis of the attitude sensor 3 and other related sensors. If the elevator is in the no-load or low-speed state, it enters the high-precision mode; if the elevator is in the high-speed state, it enters the scanning mode.

[0041] High-precision mode: When the vehicle is unloaded or running at low speed, the system uses high-precision mode for track detection. In this mode, the grating projector 5 projects the grating pattern with higher precision, and the image captured by the camera 2 is also processed with high resolution to ensure that detailed information about the track is obtained. The system applies the illumination data collected by the photosensitive sensor 4 to the LED light strip 1 to ensure that the detection area always maintains uniform lighting conditions, thereby improving imaging quality and accuracy. In this mode, the system performs detailed track surface inspection, analyzes and records the flatness and wear of the track, and provides data support for subsequent maintenance.

[0042] Scanning mode: When the elevator is in high-speed operation, the system switches to scanning mode. In this mode, the grating projector 5 and camera 2 quickly collect data to adapt to the fast operation of the elevator. In scanning mode, the system focuses on the overall status monitoring of the track, captures the changes on the track surface in real time, and makes a preliminary judgment on the safety of the track. In this mode, Orange Pi 6 continuously monitors image data and transmits the detection results to the host computer in real time through the UDP protocol for further analysis and processing.

[0043] Track monitoring and real-time warning: Whether in high-precision mode or scanning mode, the system monitors the running status of the track in real time. When an abnormality of the track is detected, such as the flatness exceeding the preset threshold, the system will automatically issue an alarm signal and notify the maintenance personnel through the host computer.

[0044] Beneficial Effects

[0045] (1) This application uses a grating projector to project patterns, a camera to take pictures, and a photoresistor to control the brightness of the LED light to provide constant imaging lighting around the track, ensure the accuracy of three-dimensional scanning, obtain clear and high-quality images, and improve the accuracy of detection.

[0046] (2) This application uses Orange Pi to communicate the acquired images with the host computer through the UDP communication protocol. The host computer processes the data and reconstructs the track in three dimensions, thereby predicting track faults and providing real-time warnings.

[0047] (3) When shooting, the present application uses a posture sensor for horizontal calibration, which can effectively improve the shooting quality and ensure the stability and accuracy of the image. The system supports mobile three-dimensional measurement. By using the posture sensor to correct the image in real time, high-speed dynamic scanning and reconstruction can be achieved. This dynamic scanning technology not only improves the efficiency of detection, but also ensures the accuracy of the measurement results.

[0048] (4) The present application can also perform measurements while the elevator is working. In traditional contact measurement, the elevator must stop working for detection, which reduces the real-time and efficiency of the detection. The design of the present application enables accurate measurements to be performed while the elevator is running, greatly improving the ability of real-time monitoring and ensuring instant feedback of the track status.

[0049] (5) This application can switch detection modes according to different elevator operating states. When the elevator is unloaded and running at low speed, the system can turn on the high-precision mode to perform detailed track surface detection; when the elevator is running at high speed, the system switches to the fast scanning mode to ensure real-time monitoring of the track and fault warning. This mode switching function can provide comprehensive track detection services without affecting the elevator's operating efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the overall structure of this application;

[0051] Figure 2 is a schematic diagram of the back of this application;

[0052] Figure 3 It is the system framework diagram of this application;

[0053] Figure 4 It is the experimental parameter diagram of this application, where part (a) is the distance between the camera and the phase plane; part (b) is the horizontal distance between the camera and the projection; part (c) is the height of the elevator rail;

[0054] Figure 5 is a real image acquisition of the present application, wherein part (a) is a reference surface stripe of 0°; part (b) is a reference surface stripe of 90°; part (c) is a reference surface stripe of 180°; part (d) is a reference surface stripe of 270°; part (e) is a deformation stripe of 0°; part (f) is a deformation stripe of 90°; part (g) is a deformation stripe of 180°; part (h) is a deformation stripe of 180°;

[0055] Figure 6 This is a comparison chart of the filtering effects of different standard deviations of the present application. Part (a) is a filter size of 5×5 with a standard deviation of 0.1; Part (b) is a filter size of 5×5 with a standard deviation of 0.5; Part (c) is a filter size of 5×5 with a standard deviation of 1;

[0056] Figure 7 It is the calculation result of the track reference surface image of the present application, wherein part (a) is the main value of the reference surface phase; part (b) is the main value of the elevator track phase; part (c) is the phase difference;

[0057] Figure 8 It is a diamond decoding schematic diagram of the present application;

[0058] Fig. 9 is a reconstruction of the elevator track morphology of the present application;

[0059] Fig.10 is the mode switching flow chart of this application;

[0060] Figure numerals: 1-LED light strip, 2-camera, 3-gesture sensor, 4-photosensitive sensor, 5-grating projector, 6-Orange Pie, 7-mobile base, 8-curved rail bracket. DETAILED DESCRIPTION

[0061] The present application is further described in detail below in conjunction with specific embodiments.

[0062] like Figure 1 , Figure 2 As shown, an elevator track detection system based on projection grating includes:

[0063] A grating projector 5 is installed on the top of the elevator car;

[0064] There are two cameras 2, which are respectively arranged on both sides of the grating projector 5, and are used to collect the grating information projected onto the elevator track;

[0065] A photosensitive sensor 4 is installed between the adjacent grating projector 5 and the camera 2;

[0066] LED light strip 1, irradiating the photo taking position of camera 2;

[0067] The Orange Pi 6 is electrically connected to the camera 2, the light sensor 4, and the LED light strip 1 respectively; the Orange Pi 6 is also connected to the host computer;

[0068] Furthermore, it also includes a posture sensor 3 installed on the top of the elevator car, and the posture sensor 3 is electrically connected to the Orange Pi 6;

[0069] The photosensitive sensor 4, the camera 2 and the LED light strip 1 are installed on the top of the elevator car through the curved rail bracket 8;

[0070] Further, the grating projector 5 is installed in the middle of the curved rail bracket 8, and the LED light strip 1 is installed at both ends of the curved rail bracket 8;

[0071] The photosensitive sensor 4, the camera 2 and the LED light strip 1 are movably connected to the curved rail bracket 8 via a C-shaped movable base.

[0072] like Figure 3 As shown, elevator track defect detection is carried out in the following steps:

[0073] In this embodiment, Figure 4 As shown, the experimental parameters are measured and collected according to the preset conditions, as shown in Figure (5), where (a) the distance between the camera and the phase plane is d = 240 cm, (b) the horizontal distance between the camera and the projection is l = 10 cm, (c) the height of the elevator rail is about 5.5 cm, and (d) the height of the elevator deformation object is about 2.5 cm;

[0074] S1, controlling the brightness of the LED light strip 1 through the photosensitive sensor 4 to obtain a constant light field;

[0075] S2, such as Figure 5 As shown, under the condition of constant light field, the reference surface fringe image and the deformation fringe image are obtained by the grating projector 5;

[0076] S3, collecting the reference surface fringe image and the deformation fringe image through camera 2, and transmitting the collected images to Orange Pi 6 for preprocessing; the details are as follows:

[0077] The preprocessing is to smooth the image through a Gaussian filter. The steps are as follows:

[0078] S3.1, compare the filtering effects of three different Gaussian filters, and select the Gaussian filter with the best effect for filtering;

[0079] like Figure 6 As shown, after comparison, the filter size selected in this embodiment is 5×5, and the standard deviation is 1. At this time, the burr peaks of the stripes at different degrees of crest and trough are significantly improved;

[0080] S3.2, normalize the filtered image to obtain a projection grating fringe period; in this embodiment, one period is 32 pixels;

[0081] S4, such as Figure 7 As shown, the wrapped phase is obtained by subtracting the phase principal values ​​of the preprocessed reference surface fringe image and the deformed fringe image;

[0082] S5, performing phase unwrapping by diamond unwrapping;

[0083] The steps for diamond unwrapping are as follows:

[0084] S5.1, determining expandable points and non-expandable points in the phase diagram after subtraction;

[0085] Expandable points are those points that are not wrapped, have no phase jumps or errors, and the phase difference between adjacent points does not exceed 2π; there are no poles, noise or interference, and their true phase values ​​can be restored by phase difference calculation;

[0086] Unexpandable points refer to poles, noise areas, and edge areas; there are jumps in phase values;

[0087] The determination method is as follows:

[0088] S5.1.1. Detect the edge area of ​​the image by edge detection algorithm and mark it as unexpandable point;

[0089] S5.1.2, using the local variance calculation method, detect the noise area in the image and mark it as an undevelopable point;

[0090] S5.1.3, by calculating whether the absolute value of the phase difference between adjacent pixel points is greater than a preset value t, t∈(-π,π), if it is greater, it is an unexpandable point;

[0091] S5.1.4, by removing all the unexpandable points, the remaining points are all expandable points;

[0092] S5.2. Take the expandable point as the initial point, and the initial point satisfies the condition that all its four neighbors are expandable points, such as Figure 8 As shown: With point A as the center, expand to its four neighborhoods in the order of top-left-bottom-right (BCDE) in sequence, and then the starting point of the second batch of expansion is BCDE, and continue to traverse and expand.

[0093] S5.3, during the diamond expansion process, the phase of each point is unwrapped;

[0094] The unwrapping process is realized through an integral path, which starts from the initial point (i.e., the expandable point that meets the requirements), and gradually expands to the entire image along a diamond track (i.e., with A as the center, it expands to its four neighborhoods in the order of up-left-down-right, and then the starting point of the second batch of expansion is BCDE, and the expansion continues. The expansion of up-left-down-right (BCDE) is the diamond track) to complete the unwrapping; on the integral path, the phase difference of adjacent pixels is calculated and added to the phase of the previous pixel, so as to obtain the absolute phase value of the current pixel;

[0095] The absolute phase value is used to calculate the height information;

[0096] S6, converting the phase information after phase unwrapping into height information and performing three-dimensional reconstruction;

[0097] Specifically, Fig. 9As shown, the final reconstructed elevator track deformation height is 129.438 pixels.

[0098] S7. Perform error analysis on the three-dimensional reconstruction image to obtain a final three-dimensional reconstruction image to complete the reconstruction of the elevator track.

[0099] The average height of the reconstructed elevator track is about 252.408 pixels (since the camera size is calibrated to 0.2148 mm), and the corresponding actual height is 5.42 cm, which is about 0.8 mm different from the actual elevator track height. This can meet the standard requirements that when the track span is within 3.5 m, the height of each 1000 mm should not exceed ±2.0 mm; when the track span is above 3.5 m, the height of each 1000 mm should not exceed ±3.0 mm.

[0100] Furthermore, the period of the projection grating stripes was adjusted to 21 pixels and the test was repeated, and the error was reduced; therefore, the measurement accuracy is higher when the period of the projection grating is reduced. In reality, the period of the projection grating can be adjusted according to the actual location and the running speed of the elevator to achieve sufficient accuracy.

[0101] Further, such as Fig.10 As shown, the elevator track detection system based on projection grating of the present application adopts different detection modes, performs corresponding detection according to the running status of the elevator, and ensures the accuracy and real-time nature of the detection results.

[0102] Elevator status detection:

[0103] The system first detects the running status of the elevator, which is divided into two situations: no-load or low-speed operation and high-speed operation. The identification of the elevator status is achieved through real-time data analysis of the attitude sensor 3 and other related sensors. If the elevator is in the no-load or low-speed state, it enters the high-precision mode; if the elevator is in the high-speed state, it enters the scanning mode.

[0104] High-precision mode: When the vehicle is unloaded or running at low speed, the system uses high-precision mode for track detection. In this mode, the grating projector 5 projects the grating pattern with higher precision, and the image captured by the camera 2 is also processed with high resolution to ensure that detailed information about the track is obtained. The system applies the illumination data collected by the photosensitive sensor 4 to the LED light strip 1 to ensure that the detection area always maintains uniform lighting conditions, thereby improving imaging quality and accuracy. In this mode, the system performs detailed track surface inspection, analyzes and records the flatness and wear of the track, and provides data support for subsequent maintenance.

[0105] Scanning mode: When the elevator is in high-speed operation, the system switches to scanning mode. In this mode, the grating projector 5 and camera 2 quickly collect data to adapt to the fast operation of the elevator. In scanning mode, the system focuses on the overall status monitoring of the track, captures the changes on the track surface in real time, and makes a preliminary judgment on the safety of the track. In this mode, Orange Pi 6 continuously monitors image data and transmits the detection results to the host computer in real time through the UDP protocol for further analysis and processing.

[0106] Track monitoring and real-time warning: Whether in high-precision mode or scanning mode, the system monitors the running status of the track in real time. When an abnormality of the track is detected, such as the flatness exceeding the preset threshold, the system will automatically issue an alarm signal and notify the maintenance personnel through the host computer.

Claims

1. An elevator track detection system based on projection grating, characterized in that: include: A grating projector (5) is installed on the top of the elevator car; Cameras (2), two of which are respectively arranged on both sides of the grating projector (5) and are used to collect grating information projected onto the elevator track; A photosensitive sensor (4) is installed between the adjacent grating projector (5) and the camera (2); An LED light strip (1) is used to illuminate the photographing position of the camera (2); Orange Pi (6), the Orange Pi (6) is electrically connected to the camera (2), the light sensor (4), and the LED light strip (1) respectively; the Orange Pi (6) is also connected to a host computer; Elevator rail defect detection is carried out in the following steps: S1, controlling the brightness of the LED light strip (1) through a photosensitive sensor (4) to obtain a constant light field; S2, under a constant light field condition, obtaining a reference surface fringe image and a deformation fringe image through a grating projector (5); S3, collecting the reference surface fringe image and the deformation fringe image through the camera (2), and transmitting the collected images to Orange Pie (6) for preprocessing; The pre-processing comprises the following steps: S3.1, compare the filtering effects of three different Gaussian filters, and select the Gaussian filter with the best effect for filtering; The three different Gaussian filters include: a Gaussian filter with a filter size of 5×5 and a standard deviation of 0.1, a Gaussian filter with a filter size of 5×5 and a standard deviation of 0.5, and a Gaussian filter with a filter size of 5×5 and a standard deviation of 1; S3.2, normalizing the filtered image to obtain the projection grating fringe period; S4, calculating the phase principal values ​​of the preprocessed reference surface fringe image and the deformation fringe image and subtracting them to obtain a wrapping phase; S5, performing phase unwrapping by diamond unwrapping; The steps for diamond unwrapping are as follows: S5.1, determining expandable points and non-expandable points in the phase diagram after subtraction; The determination method is as follows: S5.1.

1. Detect the edge area of ​​the image by edge detection algorithm and mark it as unexpandable point; S5.1.2, using the local variance calculation method to detect the noise area in the image and mark it as an undevelopable point; S5.1.3, by calculating whether the absolute value of the phase difference between adjacent pixel points is greater than a preset value t, t∈(-π,π), if it is greater, it is an unexpandable point; S5.1.4, by removing all the unexpandable points, the remaining points are all expandable points; S5.2, take the expandable point as the initial point, and expand to the four neighbors in turn if the initial point satisfies the condition that all its four neighbors are expandable points; S5.3, during the diamond expansion process, the phase of each point is unwrapped; S6, converting the phase information after phase unwrapping into height information and performing three-dimensional reconstruction; S7. Perform error analysis on the three-dimensional reconstruction image to obtain a final three-dimensional reconstruction image to complete the detection of the elevator track.

2. The elevator track detection system based on projection grating according to claim 1, characterized in that: It also includes a posture sensor (3) installed on the top of the elevator car, and the posture sensor (3) is electrically connected to the Orange Pie (6).

3. The elevator track detection system based on projection grating according to claim 1, characterized in that: The photosensitive sensor (4), the camera (2) and the LED light strip (1) are installed on the top of the elevator car via a curved rail bracket (8); The grating projector (5) is installed at the middle of the curved rail bracket (8), and the LED light strip (1) is installed at both ends of the curved rail bracket (8).

4. The elevator track detection system based on projection grating according to claim 1, characterized in that: The photosensitive sensor (4), the camera (2) and the LED light strip (1) are movably connected to the curved rail bracket (8) via a C-shaped movable base.

Images