A video recognition-based safety brake test detection method for elevator brakes

Automatically calculate the braking distance of the elevator brake through video recognition technology, solving the problem that traditional elevator brake detection methods rely on manual marking and subjective judgment, realizing accurate and automatic detection of elevator speed and braking distance, and improving detection efficiency and accuracy.

CN115849126BActive Publication Date: 2024-11-29TONGJI UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202211533931.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-01
Publication Date
2024-11-29
Estimated Expiration
2042-12-01

AI Technical Summary

Technical Problem

Traditional elevator brake safety detection methods lack quantitative performance measurement standards and rely on the subjective judgment of maintenance workers. The existing detection methods require manual marking and ranging, which has the problems of false detection and low efficiency.

Method used

Using video recognition method, vertical video shooting on the front, video image recognition algorithm is used to identify the center point and feature points of the speed limiter traction machine, calculate the speed of the traction machine, elevator speed and braking distance, and avoid manual marking and measurement. distance, realize automatic calculation and verification.

Benefits of technology

Automatic calculation and verification of elevator speed and braking distance is realized, preventing detection fraud, improving detection efficiency and accuracy, simplifying the detection process, and reducing the need for manual operation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115849126B_ABST
    Figure CN115849126B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for detecting a safe braking test of an elevator brake based on video recognition, comprising: S1, obtaining a complete braking test process video by shooting a front vertical video; S2, identifying the center point of the traction machine attached to the speed limiter and the feature point detection by a video image recognition algorithm, and matching the feature points; S3, using the center point of the traction machine as a reference point, calculating the traction machine speed, elevator speed and braking distance by analyzing video data; S4, fuzzy evaluation of the calculated stopping distance. According to the present invention, the automatic calculation and verification of the elevator speed and braking distance can be realized, and detection fraud can be prevented, and marking on the traction machine and the traction rope can be avoided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of elevator safety detection, and in particular to an elevator brake safety brake test detection method based on video recognition. Background Art

[0002] The elevator brake is also called the elevator holding brake or the brake holding brake. It is a device that ensures that the elevator can stop smoothly at each floor when it is stationary. Elevator brakes are divided into drum, disc, block and belt elevator brakes. The brake is an important safety component of the elevator. When the brake fails, it will cause the elevator to rush to the top or squat, which is very likely to cause personal injury.

[0003] Traditional elevator brake safety inspection only detects the gap between the elevator brake wheel and the brake shoe. The inspection of elevator brake performance lacks quantitative performance measurement standards. The judgment of whether the elevator stops steadily and whether the brake stops promptly is based on the subjective judgment of the maintenance workers.

[0004] The current brake test detection method of elevator brakes. Many studies related to elevator brakes have made recommendations on the stopping distance. At the same time, it is explained that the stopping distance of the elevator going up and down is also different. Usually, the stopping distance is the distance of the elevator car when it falls. The brake test detection car stopping test must be carried out by two certified elevator inspectors. (1) First stop the elevator car at the bottom floor, and make a conspicuous and clear mark on the traction machine base and the same position of the wire rope; (2) The elevator is unloaded, and the inspector issues a test command to run from the bottom floor to the top at the rated speed. When the car reaches the middle floor station and the mark on the traction wheel reaches the previous mark, the other inspector immediately disconnects the main power supply of the elevator. After the elevator stops, make a mark on the wire rope at the traction machine base; (3) Measure the distance between the two marks on the wire rope when the elevator just starts and stops, which is the braking distance. It is not enough to adjust the brake shoe clearance in the current maintenance project. For example, if the brake shoe clearance is too small, the elevator can stop smoothly, but it will affect the elevator running speed and fail to reach the rated running speed of the elevator. If the elevator stops too close, the braking acceleration will be very large, which may cause the passengers in the elevator car to be vacant or to be subjected to more than 2 times the gravity acceleration when the elevator brakes, which is harmful to the safety of the passengers. If the distance is too small, the wear of the brake shoe will be more serious, which will make it more likely to slip. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide an elevator brake safety brake test detection method based on video recognition, which can realize the automatic calculation and verification of elevator speed and braking distance, and can also prevent detection fraud and avoid marking on the traction machine and traction rope. In order to achieve the above-mentioned purpose and other advantages according to the present invention, a video recognition-based elevator brake safety brake test detection method is provided, comprising:

[0006] S1. Obtain a complete braking test process video by shooting a front vertical video;

[0007] S2. Using a video image recognition algorithm, identify the center point of the speed limiter attached to the traction machine and detect feature points, and match the feature points;

[0008] S3, taking the center point of the traction machine as the reference point, calculating the traction machine speed, elevator speed and braking distance by analyzing the video data;

[0009] S4. Perform fuzzy evaluation on the calculated stopping distance.

[0010] Preferably, the frame where the angular velocity of the traction wheel begins to change sharply is used as the starting braking point, wherein the traction wheel speed and the elevator braking distance are calculated according to the following formula:

[0011]

[0012] S Z ——Stop distance

[0013] d——Diameter of traction sheave

[0014] α——Braking angle.

[0015] Preferably, the angular velocity is converted into the linear velocity of the elevator traction wheel. The elevator speed obtained based on the video data is a basis for calibrating the actual speed of the elevator operation and a basis for determining the braking point based on the speed change. The calculation formula is as follows:

[0016]

[0017] V o ——Elevator traction wheel linear speed

[0018] n——Number of selected frames

[0019] F——Video frame rate

[0020] θ——traction wheel rotation angle.

[0021] Preferably, the information contained in the two feature points is screened, and similar feature points are matched. The feature points are matched based on the grayscale differences caused by changes in brightness, texture and color. When the shape of the picture changes or the shooting angle changes, feature points can still be generated for matching.

[0022] Preferably, the traction machine linear speed calculated in step S4 is compared with the elevator control speed given in the test detection, and the error of the detection speed can be obtained. The speed detection error is used to indirectly infer and determine the detection error of the stopping distance, thereby obtaining the error of the stopping distance.

[0023] Preferably, the middle 1 / 3 section of the braking distance between Smin+(Smax-Smin) / 3 and Smax-(Smax-Smin) / 3 is used as the braking distance, Smin is the minimum distance, and Smax is the maximum distance.

[0024] Compared with the prior art, the present invention has the following beneficial effects: by studying the non-contact video image feature recognition, the stopping distance of the elevator brake is detected as the basis for brake performance detection. The braking test process of the elevator brake is filmed with a camera, and the filmed video image is processed using an intelligent algorithm. The video filmed by the camera is used to replace the manual marking and distance measurement operations, eliminating the need for manual operations. This method can realize the automatic calculation and verification of the elevator speed and braking distance, and can also prevent detection fraud and avoid marking on the traction machine and traction rope.

[0025] The non-contact detection method obtains the elevator stopping distance by directly detecting the rotation angle of the elevator traction wheel during the stopping process. The detection principle is simple and no special equipment is required for data processing.

[0026] Non-contact detection of elevator brakes based on image recognition can realize remote detection of elevator brake performance. It can shoot videos of multiple elevator brakes in a short time and upload them to the maintenance company for processing, thereby improving detection efficiency. The detection images can also be retained as a basis for maintenance to prevent detection fraud. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flowchart of a video recognition-based elevator brake safety brake test detection method according to the present invention;

[0028] Figure 2 A brake circle center recognition diagram of the elevator brake safety brake test detection method based on video recognition according to the present invention;

[0029] Figure 3 A key point diagram for generating a safety brake test detection method for an elevator brake based on video recognition according to the present invention;

[0030] Figure 4 It is a feature point matching diagram of the elevator brake safety braking test detection method based on video recognition according to the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] Reference Figure 1-4 ,A method for detecting a safe brake test of an elevator brake based on video recognition, comprises the following steps: S1, obtaining a complete brake test process video by front vertical video shooting;

[0033] S2. The center point of the circle where the speed limiter is attached to the traction machine and the feature point detection are identified through the video image recognition algorithm. The center point recognition uses the Hough Circles algorithm to detect the center of the circle and match the feature points. To detect the linear speed of the traction wheel running coaxially with the elevator brake, it is necessary to determine the coordinates of the center of the circle where the wheel is located and the radius of the circle. Therefore, the coordinates and radius of the circle where the traction wheel of the traction machine is located must be identified through the code. The results are as follows: Figure 2 As shown in the figure, this detection method is very effective when the target is only an incomplete arc or point and a complete circle needs to be detected. The Hough circle transform principle can meet the requirements of detecting the position of the circle in two-dimensional space and output information such as the coordinates of the center of the circle and the radius. It has the characteristic that the circumference can be determined by detecting any three points on the circumference.

[0034] The feature point detection adopts the ORB characteristic detection algorithm, and the detection method adopts the FAST algorithm. The core idea of ​​the FAST algorithm is to detect all points in the fixed pixel range according to the gray value of the pixels near the detection point, and select the point that is most incompatible with the surrounding pixels. This point can be considered as a feature point. The open source ORB algorithm with the characteristics of fast feature point detection speed and stable feature point extraction is selected. It is suitable for target tracking application scenarios. From the full name of ORB, it can be seen that it also involves the BRIEF algorithm, which also has the advantages of high computing speed, strong robustness, and illumination invariance, but the performance is very poor when it comes to rotation detection. The ORB algorithm handles this problem very well, and it also handles the key points that are rotating. It can carry out effective recognition, build a framework for the function orb_featurres(img), first initialize the orb function and set its parameters. The parameters in brackets correspond to feature points, pyramid level scaling ratio, number of pyramid levels, edge threshold, number of layers of the original image in the pyramid, number of pixels needed to generate descriptors, use the ORB_HARRIS_SCORE function in OpenCV to evaluate feature points, the size of the neighborhood around the key point when generating descriptors, and the threshold of the pixel value difference when calculating FAST corner points. If no modification is required, the default values ​​of the orb function are used. You only need to modify the number of generated feature points to generate as many feature points as possible to make the detection result more accurate.

[0035] Feature matching filters the information contained in two feature points and matches similar feature points. Feature points are matched based on the grayscale differences caused by changes in brightness, texture, and color. When the shape of the image changes or the shooting angle changes, feature points can still be generated for matching. The Brute-Force matcher used is also called a brute force matcher. The feature of this matcher is that it extracts the key points in the previous target and matches them with all the key points in the next target, and outputs the points with the highest matching degree as the matching objects. Although this matching method has a large amount of calculation, it is more suitable for brake wheels that run at fast speeds. To determine whether two feature points match, it is necessary to introduce the concept of "distance" in the calculation sample, and use the distance between matching points to measure the similarity between the two samples;

[0036] S3. Taking the center point of the traction machine as the reference point, the traction machine rotation speed, elevator speed and braking distance are calculated by analyzing the video data; the traction wheel rotation angle and speed are calculated by analyzing the number of video data frames, and the traction wheel rotation angle is calculated. According to the traction wheel radius, the angular velocity is converted into the elevator traction wheel linear velocity. The elevator speed obtained based on the video data is a basis for calibrating the actual speed of the elevator operation, and it is also a basis for determining the braking point based on the speed change.

[0037]

[0038] V o ——Elevator traction wheel linear speed

[0039] n——Number of selected frames

[0040] F——Video frame rate

[0041] θ——traction sheave rotation angle;

[0042] Determination of brake stopping point and calculation of stopping distance. In the actual use of non-contact detection technology, the time point (key frame position) when the elevator starts to stop can be judged by receiving the power-on signal of the elevator brake. The distance the traction wheel rotates from the time when the elevator brake starts to be powered on to the complete stop of the elevator traction wheel is the stopping distance. The starting stopping point is screened by the rotation angle between each frame of the video. The speed of the elevator when it starts braking should be the rated speed or test speed of the operation process. In this video, the running or test speed of the elevator is unknown, so the frame before the angular velocity of the traction wheel decreases is used as the starting stopping point.

[0043] The traction sheave speed and elevator stopping distance shall be calculated according to the following formula:

[0044]

[0045] S Z ——Stop distance

[0046] d——Diameter of traction sheave

[0047] α——braking angle;

[0048] S4. Perform fuzzy evaluation on the calculated stopping distance, and compare the calculated traction machine linear speed (i.e. elevator speed) with the elevator control speed given in the test (or the actual detection speed of the encoder), so as to obtain the error of the detection speed of this project.

[0049] The speed detection error is used to indirectly infer and determine the detection error of the stopping distance, thereby obtaining the error of the stopping distance.

[0050] Finally, compare the stopping distance with the data in Table 1 to determine whether the stopping distance meets the detection requirements.

[0051] The present invention takes the middle 1 / 3 section of the stopping distance between Smin+(Smax-Smin) / 3 and Smax-(Smax-Smin) / 3 as the better stopping distance, and other stopping distances are qualified, and provides a practical fuzzy evaluation method.

[0052] Table 1 Reference range of braking distance for traction elevator upward movement

[0053]

[0054] Table 2 Fuzzy evaluation of elevator stopping distance

[0055]

[0056] The number of devices and processing scales described here are used to simplify the description of the present invention, and the application, modification and variation of the present invention will be obvious to those skilled in the art.

[0057] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A video recognition-based elevator brake safety brake test detection method, characterized in that: The following steps are involved: S1. Obtain a complete braking test process video by shooting a front vertical video; S2. Using a video image recognition algorithm, identify the center point of the speed limiter attached to the traction machine and detect feature points, and match the feature points; S3, taking the center point of the traction machine as the reference point, calculating the traction machine speed, elevator speed and braking distance by analyzing the video data; S4, performing fuzzy evaluation on the calculated stopping distance; The calculated traction machine linear speed is compared with the elevator control speed given in the test to obtain the error of the detection speed. The speed detection error is used to indirectly infer and determine the detection error of the stopping distance, thereby obtaining the error of the stopping distance. The frame where the angular velocity of the traction wheel begins to change sharply is taken as the starting braking point, where the traction wheel speed and the elevator braking distance are calculated according to the following formula: ,in , ; ; ; Convert the angular velocity to the linear velocity of the elevator traction wheel. The elevator speed obtained based on the video data is a basis for calibrating the actual speed of the elevator and determining the braking point based on the speed change. The calculation formula is as follows: (1), ; ; ; .

2. The elevator brake safety brake test detection method based on video recognition according to claim 1, characterized in that: The information contained in the two feature points is filtered, and the similar feature points are matched. The feature points are matched based on the different grayscales caused by changes in brightness, texture and color. When the shape of the picture changes or the shooting angle changes, feature points can still be generated for matching.

3. The elevator brake safety brake test detection method based on video recognition according to claim 1, characterized in that: The middle 1 / 3 of the braking distance between Smin + (Smax- Smin) / 3 and Smax - (Smax- Smin) / 3 is taken as the braking distance, where Smin is the minimum distance and Smax is the maximum distance.

Citation Information

Patent Citations

  • Method for realizing intelligent vision system for synchronous detection of elevator brake spring

    CN109238678A

  • Elevator traction wheel slipping amount detecting device and method

    CN110240038A

  • Device and method for detecting elevator traction wheel slippage

    CN110255324A

  • Braking torque self-monitoring method for traction machine of traction-driven elevator

    CN112299180A

  • Elevator device

    JP2007331901A