A vision-based method for monitoring escalator step speed

Abstract

The application relates to a kind of escalator step speed monitoring methods based on vision, including real-time acquisition several mobile step image pair image is preprocessed and binarization processing;Find starting point A1, find boundary corner point B1, point A2 and point B2 from point A1, with the column where point A2 is as center, find each profile on the left, record the boundary corner point of each profile as point Ai-1, point Ai, with the column where point B2 is as center, find each profile on the right, record the boundary corner point of each profile as point Bi-1, point Bi, with point Ai as starting point, find the profile section of point Ai to point Bi, connect point A1 to point B1, point B1 to point Bi, and point Ai to point Bi, A1 to point Ai and the pixel point with the value of KZ1, KZ2, KZ3 and KZ4, form the profile edge of entire yellow outer boundary, obtain the speed of escalator according to profile edge and time used.

Description

Technical Field

[0001] The invention relates to the technical field of escalator safety, in particular to a vision-based monitoring method for escalator step speed. Background Art

[0002] With rapid economic development, accelerated urbanization, and improved infrastructure, escalators, with their superior and stable transportation capabilities, have become a fixture of everyday life. This has been particularly evident in crowded public spaces such as train stations, airports, hospitals, and shopping malls, where they have greatly facilitated transportation. However, if the speed of the handrails on either side of an escalator differs significantly from the speed of the steps and treads, it can directly impact the safety of passengers and potentially cause serious casualties. To ensure passenger safety, regular inspections and daily monitoring of escalators are essential. Currently, contact measurement is the primary method for inspecting escalator handrail speed deviation. This involves attaching a tachometer roller to the left and right handrails, as well as the steps, treads, or tape surfaces, measuring the corresponding speed and then calculating the speed using a formula. However, this method has certain drawbacks, including low detection efficiency. In addition to human factors, slippage on the contact surface, surface deposits, and wear of the tachometer roller all affect measurement accuracy. Therefore, a vision-based method for monitoring escalator step speed is necessary. Summary of the Invention

[0003] The purpose of the present invention is to provide a vision-based monitoring method for escalator step speed, which can realize real-time monitoring of escalator step running speed.

[0004] To achieve the above object, the present invention provides the following technical solution: a vision-based escalator step speed monitoring method, comprising the following steps:

[0005] Step S1: A camera is set at the end station of the escalator to acquire a number of moving step images in real time within a certain viewing angle; the images are preprocessed using a median filter method to eliminate discrete noise points in the images, and each pixel in the image is checked to see if it is within a threshold range. If it is within the threshold range, the pixel is set to 1; if it is not within the threshold range, the pixel is set to 0, thereby performing a binarization process on the image;

[0006] Step S2: Starting from the upper left corner of the processed binary image, traverse the image from left to right and from top to bottom, find the first boundary point that meets the requirements as the starting point A1, search for boundary corner points from point A1, record the boundary corner points in clockwise order as points B1, A2, and B2, and calculate the average point of the contour between points A2, A1, B1, and B2 with five consecutive pixels, and record the value KZ1;

[0007] Step S3: Centering the column where point A2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the left, starting from the leftmost pixel in the row below the row where point A2 is located. Note the boundary corner points of each contour segment as points Ai-1 and Ai. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ2.

[0008] Step S4: Centering the column where point B2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the right, starting from the rightmost pixel in the row below the row where point B2 is located. Note the boundary corner points of each contour segment as points Bi-1 and Bi. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ3.

[0009] Step S5: Using point Ai as the starting point and the starting point as the center, find the contour segment from point Ai to point Bi, calculate the average point of the contour segment from point Ai to point Bi for 5 consecutive pixels, and record the value KZ4;

[0010] Step S6: Connect point A1 to point B1, point B1 to point Bi, point Ai to point Bi, A1 to point Ai, and the pixel points with values ​​of KZ1, KZ2, KZ3, and KZ4 to form the contour edge of the entire yellow outer boundary, and obtain the speed of the escalator based on the contour edge and the time taken.

[0011] Furthermore, the step S2 of "searching for boundary corner points from point A1, and recording the boundary corner points in a clockwise direction as point B1, point A2, and point B2" is further as follows: calculating the slope at intervals of 5 pixels to the right from point A1, and if the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated; the pixel points where the slope suddenly changes are boundary corner points, and are recorded in a clockwise direction as point B1, point A2, and point B2.

[0012] Furthermore, the step S3 of "searching for each left contour segment starting from the leftmost pixel point in the next row of the row where point A2 is located, and recording the boundary corner points of each contour segment as point Ai-1 and point Ai" is further as follows: searching for pixel points starting from the leftmost pixel point in the next row of the row where point A2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then starting the search from the leftmost pixel point in another row, and repeating this process until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the left contour pixel search is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, and the contour segments of each segment are found and the endpoints of each segment are recorded as point Ai-1 and point Ai.

[0013] Furthermore, the step S4 of "searching for each contour segment on the right starting from the rightmost pixel point in the next row of the row where point B2 is located, and recording the boundary corner points of each contour segment as point Bi-1 and point Bi" is further as follows: searching for pixel points starting from the rightmost pixel point in the next row of the row where point B2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then starting the search from the rightmost pixel point in another row, and repeating this process until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the search for the pixel points of the contour on the right is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, the contour segments of each segment are found, and the endpoints of each segment are recorded as point Bi-1 and point Bi.

[0014] Furthermore, starting from the pixel point directly below point Ai, search counterclockwise to see if the values ​​of the 5 neighboring pixels are 1. When the pixel value of a neighboring point is 1, this point is the new center point. Starting from the new center point, search counterclockwise for neighboring points with a pixel value of 1. Repeat this process until point Bi is found. Calculate the slope starting from point Ai at intervals of 5 pixels to the right. If the pixel points in the interval are not within the set slope deviation range, remove the pixel points and find the contour segment from point Ai to point Bi.

[0015] The step S6 of "obtaining the speed of the escalator based on the contour edge and the time taken" is further as follows: finding the center of mass position (X1, Y1) of the contour of the entire yellow outer boundary of the first image, and the center of mass position (Xn, Yn) of the contour of the entire yellow outer boundary of the nth image. When the center of mass positions of the two images coincide, that is, a step length S is traveled, and the time taken is known from n images T, the speed V can be calculated, and the speed V = S / T.

[0016] Beneficial effects of the present invention: The present invention provides a vision-based method for monitoring the speed of escalator steps. A camera installed at the end of the escalator obtains a number of moving step images within a certain viewing angle in real time, pre-processes the images, quickly searches for multi-region segmented contours, merges the entire outer boundary contour, determines the center of mass position through the outer contour, calculates the running speed of the steps according to the change in the position of the center of mass, and realizes real-time monitoring of the step running speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of the image outline of the present invention; DETAILED DESCRIPTION

[0018] The present invention will be further described below with reference to the accompanying drawings.

[0019] See also Figure 1The present invention provides an embodiment: a vision-based method for monitoring escalator step speed, comprising the following steps:

[0020] Step S1: A camera is set at the end station of the escalator to acquire a number of moving step images in real time within a certain viewing angle; the images are preprocessed using a median filter method to eliminate discrete noise points in the images, and each pixel in the image is checked to see if it is within a threshold range. If it is within the threshold range, the pixel is set to 1; if it is not within the threshold range, the pixel is set to 0, thereby performing a binarization process on the image;

[0021] Step S2: Starting from the upper left corner of the processed binary image, traverse the image from left to right and from top to bottom, find the first boundary point that meets the requirements as the starting point A1, search for boundary corner points from point A1, record the boundary corner points in clockwise order as points B1, A2, and B2, and calculate the average point of the contour between points A2, A1, B1, and B2 with five consecutive pixels, and record the value KZ1;

[0022] Step S3: Centering the column where point A2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the left, starting from the leftmost pixel in the row below the row where point A2 is located. Note the boundary corner points of each contour segment as points Ai-1 and Ai. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ2.

[0023] Step S4: Centering the column where point B2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the right, starting from the rightmost pixel in the row below the row where point B2 is located. Note the boundary corner points of each contour segment as points Bi-1 and Bi. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ3.

[0024] Step S5: Using point Ai as the starting point and the starting point as the center, find the contour segment from point Ai to point Bi, calculate the average point of the contour segment from point Ai to point Bi for 5 consecutive pixels, and record the value KZ4;

[0025] Step S6: Connect points A1 to B1, B1 to Bi, Ai to Bi, A1 to Ai, and the pixels with values ​​KZ1, KZ2, KZ3, and KZ4 to form the outline of the entire yellow outer boundary. The escalator speed is obtained based on the outline and the elapsed time. KZ1, KZ2, KZ3, and KZ4 do not refer to four specific values, but rather to the average of five consecutive pixels.

[0026] Please continue reading Figure 1As shown, in one embodiment of the present invention, the "search for boundary corner points from point A1, and record the boundary corner points in a clockwise direction as point B1, point A2, and point B2" in step S2 is further as follows: starting from point A1 and going to the right at intervals of 5 pixels, the slope is calculated; if the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated; the pixel points where the slope suddenly changes are boundary corner points, which are recorded in a clockwise direction as point B1, point A2, and point B2.

[0027] Please continue reading Figure 1 As shown, in one embodiment of the present invention, the step S3 of "searching for each contour segment on the left starting from the leftmost pixel point in the next row of the row where point A2 is located, and recording the boundary corner points of each contour segment as point Ai-1 and point Ai" is further as follows: searching for pixel points starting from the leftmost pixel point in the next row of the row where point A2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then searching from the leftmost pixel point in another row, and repeating this process until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the search for the pixel points of the left contour is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, and the contour segments of each segment are found and the endpoints of each segment are recorded as point Ai-1 and point Ai.

[0028] Please continue reading Figure 1 As shown, in one embodiment of the present invention, the step S4 of "searching for each contour segment on the right starting from the rightmost pixel point in the next row of the row where point B2 is located, and recording the boundary corner points of each contour segment as point Bi-1 and point Bi" is further as follows: searching for pixel points starting from the rightmost pixel point in the next row of the row where point B2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then searching from the rightmost pixel point in another row, and repeating this process until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the search for the pixel points of the right contour is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, and the contour segments of each segment are found and the endpoints of each segment are recorded as point Bi-1 and point Bi.

[0029] Please continue reading Figure 1 As shown, in one embodiment of the present invention, the pixel point directly below point Ai is used as the starting point to search counterclockwise to see if the values ​​of the five neighboring pixel points are 1. When the pixel value of a certain point in the neighborhood is 1, the point is the new center point. The new center point is used as the starting point and the neighboring points with a pixel value of 1 are searched counterclockwise, and this process is repeated until point Bi is found. The slope is calculated starting from point Ai at intervals of 5 pixels to the right. If the pixel points in the interval are not within the set slope deviation range, the pixel points are removed to find the contour segment from point Ai to point Bi.

[0030] Please continue reading Figure 1 As shown, in one embodiment of the present invention, the "obtaining the speed of the escalator based on the contour edge and the time taken" in step S6 is further as follows: finding the centroid position (X1, Y1) of the contour of the entire yellow outer boundary of the first image, and the centroid position (Xn, Yn) of the contour of the entire yellow outer boundary of the nth image. When the centroid positions of the two images coincide, that is, a step length S has been traveled, and the time taken is known from n images T, the speed V can be calculated, and the speed V = S / T.

[0031] The present invention has the following working principle: This system takes the yellow discontinuous edge area of ​​the escalator step under test as the object, and uses the extracted yellow area and multi-area contour fast search methods, etc. The specific method is as follows:

[0032] 1. Image preprocessing: Use median filtering method to preprocess the image and eliminate discrete noise points in the image;

[0033] 2. Image binarization: Check whether each pixel in the image is within the threshold range. If it is within the threshold range, the pixel is set to 1, otherwise it is set to 0.

[0034] 3. Traverse the image: Starting from the upper left corner of the processed binary image, traverse the image from left to right and from top to bottom to find and quickly locate the contour edge. The contour centroid position of the first image (X1, Y1) and the contour centroid position of the nth image (Xn, Yn) are used. When the centroid positions of the two images coincide, it means that a step length S has been traveled. The time T taken for n images is known, and the speed V = S / T can be calculated.

[0035] The above description is only a preferred embodiment of the present invention and should not be understood as limiting the present application. All equivalent changes and modifications made within the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A vision-based method for monitoring escalator step speed, characterized in that: The steps include: Step S1: A camera is set at the end station of the escalator to acquire a number of moving step images in real time within a certain viewing angle; the images are preprocessed using a median filter method to eliminate discrete noise points in the images, and each pixel in the image is checked to see if it is within a threshold range. If it is within the threshold range, the pixel is set to 1; if it is not within the threshold range, the pixel is set to 0, thereby performing a binarization process on the image; Step S2: Starting from the upper left corner of the processed binary image, traverse the image from left to right and from top to bottom, find the first boundary point that meets the requirements as the starting point A1, search for boundary corner points from point A1, record the boundary corner points in clockwise order as points B1, A2, and B2, and calculate the average point of the contour between points A2, A1, B1, and B2 with five consecutive pixels, and record the value KZ1; Step S3: Centering the column where point A2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the left, starting from the leftmost pixel in the row below the row where point A2 is located. Note the boundary corner points of each contour segment as points Ai-1 and Ai. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ2. Step S4: Centering the column where point B2 is located and taking the 5 pixel range to the left and right of the column as the range, search for each contour segment on the right, starting from the rightmost pixel in the row below the row where point B2 is located. Note the boundary corner points of each contour segment as points Bi-1 and Bi. Calculate the average point of 5 consecutive pixels in each contour segment and record the value KZ3. Step S5: Using point Ai as the starting point and the starting point as the center, find the contour segment from point Ai to point Bi, calculate the average point of the contour segment from point Ai to point Bi for 5 consecutive pixels, and record the value KZ4; Step S6: Connect point A1 to point B1, point B1 to point Bi, point Ai to point Bi, A1 to point Ai, and the pixel points with values ​​of KZ1, KZ2, KZ3, and KZ4 to form the contour edge of the entire yellow outer boundary, and obtain the speed of the escalator based on the contour edge and the time taken.

2. The method for monitoring escalator step speed based on vision according to claim 1, characterized in that: In step S2, the boundary corner points are searched from point A1, and the boundary corner points are recorded clockwise as point B1, point A2, and point B2: the slope is calculated every 5 pixels to the right from point A1, and if the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated; the pixel points where the slope suddenly changes are the boundary corner points, and are recorded clockwise as point B1, point A2, and point B2.

3. The method for monitoring escalator step speed based on vision according to claim 1, characterized in that: In step S3, each contour segment on the left is searched starting from the leftmost pixel point in the next row of the row where point A2 is located, and the boundary corner point of each contour segment is recorded as point Ai-1 and point Ai: starting from the leftmost pixel point in the next row of the row where point A2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then another row is searched starting from the leftmost pixel point, and this is repeated until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the left contour pixel search is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, and the contour segments of each segment are found and the endpoints of each segment are recorded as point Ai-1 and point Ai.

4. The method for monitoring escalator step speed based on vision according to claim 1, characterized in that: In step S4, each contour segment on the right is searched starting from the rightmost pixel point in the next row of the row where point B2 is located, and the boundary corner point of each contour segment is recorded as point Bi-1 and point Bi: starting from the rightmost pixel point in the next row of the row where point B2 is located, when the pixel value of a point in the pixel interval of a row is 1, or the pixel values ​​of all points in the pixel interval are 0, then a new row is started to search from the rightmost pixel point, and this is repeated until the pixel values ​​of all points in the pixel interval of a row are 0 for 5 times, then the right contour pixel point search is completed, and the slope is calculated for every 5 pixels in each contour segment. If the pixel points in the interval are not within the set slope deviation range, the pixel points are eliminated, and the contour segments of each segment are found and the endpoints of each segment are recorded as point Bi-1 and point Bi.

5. The method for monitoring escalator step speed based on vision according to claim 1, characterized in that: In step S5, the contour segment from point Ai to point Bi is found with the starting point as the center: starting from the pixel point directly below point Ai, search counterclockwise to see if the values ​​of the 5 neighboring pixels are 1. When the pixel value of a certain point in the neighborhood is 1, the point is the new center point. Starting from the new center point, search counterclockwise for neighboring points with a pixel value of 1. Repeat this process until point Bi is found. Calculate the slope every 5 pixels to the right from point Ai. If the pixel points in the interval are not within the set slope deviation range, remove the pixel points to find the contour segment from point Ai to point Bi.

6. The method for monitoring escalator step speed based on vision according to claim 1, characterized in that: The escalator speed is obtained based on the contour edge and the time taken in step S6 as follows: the center of mass position (X1, Y1) of the entire yellow outer boundary of the first image and the center of mass position (Xn, Yn) of the entire yellow outer boundary of the nth image are calculated. When the center of mass positions of the two images coincide, it means that the escalator has traveled a step length S. The time taken by n images is known, and the speed V can be calculated. Speed ​​V = S / T.

Images