Methods, measuring devices and storage media for measuring the speed of objects conveyed by belts

Abstract

This invention discloses a method, measuring device, and storage medium for measuring the speed of an object conveyed by a belt conveyor. The method includes: capturing images of a conveyor belt with laser lines using two cameras; matching the images captured by the two cameras to obtain the depth of the object containing the laser lines; calculating an evenly spaced object thickness distribution based on the obtained depth of the object containing the laser lines; obtaining the displacement of the object on the conveyor belt between two consecutive frames based on the evenly spaced object thickness distribution; and calculating the object's velocity based on the obtained displacement between the two consecutive frames. This invention obtains the evenly spaced object thickness distribution by interpolating the depth of the object containing the laser lines and accurately extracts the belt conveyor speed information using a cross-correlation algorithm, achieving direct measurement of the conveyor belt speed and ensuring high measurement accuracy.

Description

Technical Field

[0001] This invention relates to the field of measurement technology, specifically a method for measuring the speed of an object conveyed by a belt. Background Technology

[0002] A conveyor belt is a material handling machine that continuously transports materials along a fixed path; it is also known as a continuous conveyor. Conveyors can transport materials horizontally, inclined, and vertically, and can also form spatial conveying lines, which are generally fixed. Conveyors have a large conveying capacity, long conveying distance, and can complete several technological operations simultaneously during the conveying process, so they are widely used.

[0003] Belt conveyors are used in industries such as food, metallurgy, power, coal, chemicals, building materials, docks, and grain. Belt conveyor structures include horizontal straight conveying, inclined conveying, and turning conveying, among others. Accessories such as lifting baffles and side baffles can be added to the conveyor belt to meet various process requirements. Drive methods include geared motor drive. Speed ​​control methods include frequency conversion speed regulation and stepless speed regulation.

[0004] During the transmission motion of a conveyor belt mechanism, the transmission rate needs to be detected to control the conveying speed. In actual conveyor belt mechanisms, speed sensors are generally not installed on the power unit. Modifying the power unit to add speed sensors is difficult for some workshop personnel. Purchasing new, more advanced power units with speed measurement capabilities and installing them on every conveyor belt would be wasteful of resources (some conveyor belts do not require speed detection when transporting certain items) and significantly increase costs. Furthermore, current technologies primarily use stepper motors or frequency converters to achieve high speed measurement accuracy, but this method is costly. Also, due to conveyor belt aging and varying installation tightness, slippage can still occur, leading to errors in the calculated conveyor belt speed and affecting the accuracy of coal sample testing. Some methods use encoders for conveyor belt speed measurement; encoders are rotary sensors that convert angular displacement or angular velocity into a series of digital electrical pulses. The disadvantage of this method is that it does not account for errors caused by couplings, gearboxes, belt slippage, etc. Therefore, the speed measurement of conveyor belt equipment should be carried out using a technology that requires minimal modification to the equipment, causes minimal damage, is highly applicable, and provides accurate measurement.

[0005] Binocular stereo vision is a stereo vision system designed to directly mimic the human and animal eye systems. It works by using two cameras at different positions to observe the same scene, then matching the acquired image pairs to detect the position of a point in the scene on the image pair, calculating its pixel coordinates, and then using an imaging geometry model to obtain the scene's 3D coordinate information, thus acquiring the scene's 3D geometric information. The key to binocular stereo vision is stereo matching technology. Depending on the characteristics of the object, there are many stereo matching methods, with mainstream methods divided into stereo matching based on local features and stereo matching based on global features. As an important branch of computer vision, binocular stereo vision has been extensively studied by many scholars and research institutions both domestically and internationally, and has been applied to fields such as robot navigation and space exploration.

[0006] Belt speed measurement direction: Existing patent publication number CN 107499862 A, entitled "A Method, Device, and Conveyor Belt for Detecting the Speed ​​of a Coal Sample Testing Line Conveyor Belt." This invention discloses a method, device, and conveyor belt for detecting the speed of a coal sample testing line conveyor belt. The method includes installing at least one signal transmitting point on the driven wheel of the coal sample testing line conveyor belt, installing a sensor near the circumference of the rotating circumference of the signal transmitting point, the sensor generating a pulse signal when the signal transmitting point passes the sensor, and calculating the speed of the conveyor belt by acquiring the pulse signal through a processing unit.

[0007] Existing belt speed measurement methods, such as the aforementioned similar patent "A Method, Device, and Conveyor Belt for Detecting the Speed ​​of a Coal Sample Testing Line," require modification of the conveyor belt itself to add a signal transmitter and the placement of sensors in the accessories. This method has the following disadvantages: 1) Equipment installation: Modification of the equipment itself is required, which is difficult in some industrial settings; 2) Measurement signal: The signal transmitter is installed on the driven wheel. The sensor receives a pulse signal through the circumferential motion of the driven wheel. Two problems exist: first, the conveyor belt may slip, causing the driven wheel to rotate asynchronously, resulting in a signal pulse that does not match the actual situation; second, there may be obstructions, causing the signal emitted by the signal transmitter to be blocked, leading to inaccurate measurements; 3) Measurement method: The measurement is of the circumferential speed of the driven wheel. When the conveyor belt has been running for a long time, slippage and wear may occur, resulting in a significant difference between the circumferential speed of the driven wheel and the belt conveyor speed.

[0008] Binocular vision speed measurement direction: Existing patent publication number CN 110189377 A, entitled "A High-Precision Speed ​​Measurement Method Based on Binocular Stereo Vision," discloses a high-precision speed measurement method based on binocular stereo vision, comprising the following steps: (a) setting up a longitudinal speed measurement binocular camera and a lateral speed measurement binocular camera; (b) calibrating the longitudinal and lateral speed measurement binocular cameras to obtain the correspondence between camera pixels and actual dimensions; (c) acquiring images of moving objects using the longitudinal and lateral speed measurement binocular cameras, and obtaining feature points of the moving objects through SIFT corner detection; (d) calculating the vehicle distance directly corresponding to two adjacent frames using a binocular parallax algorithm, and calculating the longitudinal and lateral velocity components of the moving object based on the frame rate; (e) appropriately synthesizing the longitudinal and lateral velocity components to obtain the speed of the moving object. This method solves the problems of existing bullet screen distance measurement methods, which require identifying the measured object before distance estimation, neglecting the lateral movement speed of some vehicles, and having insufficient measurement accuracy. This method of speed measurement essentially achieves displacement measurement by measuring the change in distance between the vehicle and the binocular camera between two consecutive frames.

[0009] Existing binocular vision velocimetry technology suffers from low discriminative power when dealing with objects with weak or no texture. The images have uniform pixel values ​​and few feature points, leading to mismatches and reduced accuracy in disparity calculation. In the aforementioned patent, "A High-Precision Velocity Measurement Method Based on Binocular Stereo Vision," the moving object is a car, which exhibits clear texture features under natural light. However, for objects transported on conveyor belts in industrial settings, such as coal, ore, grain, and tobacco, the overall color and texture are not clearly distinguishable, making accurate matching impossible. This results in significant distortion in the reconstructed 3D image, ultimately hindering accurate displacement identification and speed calculation. Summary of the Invention

[0010] The technical problem to be solved by the present invention is to provide a method for measuring the speed of belt-conveyed objects based on binocular vision to achieve high-precision displacement and speed measurement.

[0011] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0012] This invention first provides a method for measuring the speed of objects conveyed by a belt based on binocular vision, comprising:

[0013] Two cameras were used to capture images of the conveyor belt with laser lines.

[0014] By matching images of the conveyor belt with laser lines captured by two cameras, the depth of the object containing the laser lines can be obtained.

[0015] Based on the depth of the object where the laser line is located, interpolation is performed to obtain the equally spaced object thickness distribution;

[0016] Based on the equally spaced object thickness distribution in two consecutive frames, the displacement of the object on the conveyor belt between the two frames is obtained.

[0017] The velocity of the object is determined based on the displacement of the object in the two frames before and after it is obtained.

[0018] The displacement of the object on the conveyor belt between two frames is obtained based on the equally spaced object thickness distribution in the two consecutive images, including:

[0019] To perform cross-correlation calculation on the object thickness distribution of two consecutive frames, and obtain the number of displacement pixels M corresponding to the maximum value of the cross-correlation function;

[0020] The number of displacement pixels M corresponding to the maximum value of the cross-correlation function is obtained, and the displacement of the object on the conveyor belt between the previous and next frames is obtained.

[0021] In the step of obtaining the displacement of the object in two consecutive frames, the cross-correlation function used for cross-correlation calculation is:

[0022]

[0023] Where n is the number of displacement pixels; N is the length of the thickness distribution sequence; g1 i Let g1 be the object thickness at the i-th displacement pixel in the object thickness distribution of the previous frame image, and g2 be the object thickness at the i-th displacement pixel. i Let g2 be the object thickness at the i displacement pixels in the object thickness distribution of the next frame image, where i = 1, 2, 3, ..., N.

[0024] The velocity of the object is to be determined as follows:

[0025]

[0026] Where v is the velocity of the object; M is the number of displacement pixels corresponding to the maximum value of the cross-correlation function; T is the sampling time interval between two consecutive frames; Z0 is the depth of the conveyor belt plane; Δx(z0) is the actual spatial distance of one pixel at a depth of z0.

[0027] The depth of the object containing the laser line is obtained by matching images of the conveyor belt captured by two cameras, including:

[0028] Perform row alignment matching on the images captured by the two cameras, searching for a unique pixel in a row of the images captured by the two cameras;

[0029] After searching for pixels in the images captured by the two cameras, the coordinates of the bright spots in each row are recorded to complete the parallax calculation:

[0030] d = x L -x R

[0031] Where d is the parallax; x L The coordinates of the bright spot in the left image; x R The brightness coordinates are shown in the right figure;

[0032] Calculate the depth of the object containing the laser line based on parallax:

[0033]

[0034] Where z is the depth of the object containing the laser line; f is the camera focal length; and B is the distance between the optical centers of the two cameras.

[0035] The present invention also provides a device for measuring the speed of objects conveyed by a belt based on binocular vision, comprising:

[0036] A linear laser emits a laser line parallel to the direction of the conveyor belt's speed.

[0037] A binocular camera, positioned on both sides of the conveyor belt, is used to acquire images of objects on the conveyor belt with laser lines.

[0038] The processor processes the image of the conveyor belt object with laser lines acquired by the binocular camera according to the binocular vision-based speed measurement method for conveyor belt objects provided above, and obtains the speed of the conveyor belt object.

[0039] The present invention also provides a computer-readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the above-described binocular vision-based belt conveyor speed measurement method.

[0040] Compared with the prior art, the present invention has the following advantages:

[0041] (1) This method enables precise measurement of conveyor belt speed. By actively adding laser lines, it can be applied to various objects, including those without texture, adding strong features to textureless slag objects and solving the problem of difficult matching of textureless objects in binocular vision stereo matching. By interpolating the depth of the object where the laser lines are located, the thickness distribution of the object at equal intervals is obtained, and the cross-correlation algorithm is used to accurately extract the belt conveyor speed information, realizing direct measurement of conveyor belt speed and ensuring high measurement accuracy. Instead of indirect measurement through measuring the drive wheel or transmission wheel, it avoids errors caused by belt wear and slippage.

[0042] (2) Real-time measurement is possible. This invention can achieve rapid processing of binocular vision images, processing 5-10 sets of images per second, far exceeding the real-time requirements of general industrial sites.

[0043] (3) Simple structure. The present invention consists of two cameras and laser software. The laser only needs to emit linear laser light, without the need for complex other structures such as galvanometers, structured light, etc.

[0044] (4) No modification is required to the equipment itself. This method only requires the equipment to be installed directly above the conveyor belt. No modification is required to the conveyor belt itself. It also adopts non-contact measurement, which has no impact or damage to the conveyor belt itself.

[0045] (5) High cost performance. The device of this invention can use similar general equipment on the market, with no special requirements, low cost, and accurate and fast measurement results, thus offering high cost performance.

[0046] (6) Wide range of applications. The present invention can be equipped with a protective cooling shell, making it highly adaptable to the environment and suitable for various industries such as coal, mining, grain, and tobacco that require materials to be transported by conveyor belt. Attached Figure Description

[0047] Figure 1 Schematic diagram of a binocular camera system;

[0048] Figure 2 Flowchart for measuring conveyor belt speed using an active binocular vision system;

[0049] Figure 3 A schematic diagram illustrating the principle of binocular vision.

[0050] Figure captions: 1: Binocular camera; 2: Laser; 3: Computer. Detailed Implementation

[0051] Example 1

[0052] This embodiment provides a method for measuring the speed of objects conveyed by a belt conveyor, which can be applied to measuring the ash discharge speed of a dry ash discharger in a power plant boiler. The measurement method includes the following steps:

[0053] S1. Capture images of the conveyor belt with laser lines using two cameras respectively;

[0054] S2. Match the images of the conveyor belt with laser lines captured by the two cameras to obtain the depth of the object where the laser lines are located;

[0055] S3. Based on the depth of the object where the laser line is located, perform interpolation calculations to obtain the equally spaced object thickness distribution;

[0056] S4. Based on the equally spaced object thickness distribution in the two consecutive frames, obtain the displacement of the object on the conveyor belt between the two consecutive frames.

[0057] S5. Calculate the object's velocity based on the displacement of the object in the two frames before and after the object.

[0058] Step S4, obtaining the displacement of the object on the conveyor belt between the two frames based on the equally spaced object thickness distribution in the two consecutive frames, includes:

[0059] S41. To perform cross-correlation calculation on the object thickness distribution of two consecutive frames of images, and obtain the number of displacement pixels M corresponding to the maximum value of the cross-correlation function;

[0060] S42. Obtain the number of displacement pixels M corresponding to the maximum value of the cross-correlation function, and obtain the displacement of the object on the conveyor belt between the previous and next frames:

[0061] L=MΔx(z0)

[0062] In step S41, the cross-correlation function used for cross-correlation calculation is:

[0063]

[0064] Where n is the number of displacement pixels; N is the length of the thickness distribution sequence; g1 i Let g1 be the object thickness at the i-th displacement pixel in the object thickness distribution of the previous frame image, and g2 be the object thickness at the i-th displacement pixel. i Let g2 be the object thickness at the i displacement pixels in the object thickness distribution of the next frame image, where i = 1, 2, 3, ..., N.

[0065] The velocity of the object obtained in step S5 is:

[0066]

[0067] Where v is the velocity of the object; M is the number of displacement pixels corresponding to the maximum value of the cross-correlation function; T is the sampling time interval between two consecutive frames; Z0 is the depth of the conveyor belt plane; Δx(z0) is the actual spatial distance of one pixel at a depth of z0.

[0068] Step S2: Match the images of the conveyor belt with the laser line captured by the two cameras to obtain the depth of the object containing the laser line, including:

[0069] Step S21: Perform row alignment matching on the images captured by the two cameras, and search for a unique pixel in a row of the images captured by the two cameras;

[0070] Step S22: After searching for pixels in the images captured by the two cameras, record the coordinates of the bright spots in each row to complete the parallax calculation.

[0071] d = x L -x R

[0072] Where d is the parallax; x L The coordinates of the bright spot in the left image; x R The brightness coordinates are shown in the right figure;

[0073] Step S23: Calculate the depth of the object containing the laser line based on parallax.

[0074]

[0075] Where z is the depth of the object containing the laser line; f is the camera focal length; and B is the distance between the optical centers of the two cameras.

[0076] Example 2

[0077] This embodiment provides a device for measuring the speed of objects conveyed by a belt conveyor, which can be applied to measuring the ash discharge speed of a dry ash discharge machine in a power plant boiler. The testing device is a binocular camera system, including a binocular camera 1 and a laser 2. The arrangement of the binocular camera system is described in [reference needed]. Figure 1 Two parallel cameras are used, with laser 2 and binocular camera 1 fixed on the same bracket. The line connecting the two cameras of binocular camera 1 is perpendicular to the direction of the conveyor belt speed being measured, and the emitted laser line is parallel to the direction of the conveyor belt speed. The images captured by the binocular cameras with the laser line are transmitted to the host computer 3 via network cable. The host computer 3 processes the images in conjunction with the belt speed to obtain the required volumetric flow rate. This system is versatile; there are no special requirements for the cameras and line laser light sources used. The specific camera and camera lens selection is determined based on the range of the measurement object. The line laser shape does not need to be complexly encoded into structured light, and the light source power is determined based on the absorptivity of the measurement object.

[0078] Cameras and lasers are equipped with cooling protection devices to protect them from damage caused by impacts, high temperatures, moisture, and other contaminants, depending on the environment in which they are used.

[0079] The measurement method based on the belt conveyor speed measuring device is described in the following document. Figure 2 It includes the following steps:

[0080] The first step is to complete the equipment installation, ensure that the laser line is installed as required, and complete the binocular camera calibration.

[0081] The essence of binocular vision measurement is a transformation process from the position information of an object in three-dimensional space to its position information in a two-dimensional plane. This involves coordinate system transformation, and the transformation relationship is reflected in the intrinsic and extrinsic parameters of the binocular camera, which need to be determined through camera calibration. For ease of calculation, four reference coordinate systems are typically established: the image coordinate system located on the camera's imaging plane, the camera coordinate system passing through the camera's optical center and optical axis, the world coordinate system located in actual space, and the object coordinate system established according to the object.

[0082] After obtaining the internal and external parameters such as the focal length f of the single camera, further binocular camera calibration is needed to obtain the relative positional relationship between the left and right cameras, i.e., to calculate the rotation matrix R and the translation matrix T. A point P in space has coordinates P0 in the coordinate systems of the left and right cameras.L and P R The rotation matrix R of the left and right cameras can be used. L and R R and translation vector T L and T R To indicate,

[0083]

[0084] Among them, P W It is the coordinate of point P in the world coordinate system. Equation (2) eliminates P W Afterwards, we can obtain:

[0085]

[0086] Based on the relationship between the left and right image points and the external parameter rotation and translation matrices, the required solution matrix for bi-target calibration can be obtained from equation (3):

[0087]

[0088] Currently, both Matlab and OpenCV have camera calibration toolkits that can be directly called for related processing. Here, R is the rotation matrix of the stereo system to be calibrated, and T is the translation vector of the stereo system to be calibrated. With the cameras arranged nearly parallel, R is approximately an identity matrix, and the vector T can be expressed as equation (4:

[0089] T = [B 00] (4)

[0090] Where B is the baseline length.

[0091] The second step is to calibrate the depth of the conveyor belt reference plane. The calibration process is shown in steps three through seven below, thereby obtaining the depth of the conveyor belt at the laser line.

[0092] The third step involves the left and right cameras capturing images with laser lines.

[0093] The fourth step involves preprocessing the captured images, including converting them to grayscale and histogram equalization to enhance image contrast and thus improve laser line features.

[0094] The fifth step is to binarize the grayscale image to separate the laser line from the background. Since the laser line occupies multiple pixels in the image, to further eliminate matching ambiguity, it is necessary to extract the laser line skeleton, so that each row of pixels is only a pixel-wide line shape as input, achieving unique matching.

[0095] In this embodiment, the mid-axis transformation method is used for skeleton extraction.

[0096] Let A be a planar region, and let MA be the central axis of A, defined as follows:

[0097]

[0098] Where p, q1, and q2 are points in region A, and d(p, q2) and d(p, q2) are Euclidean distances.

[0099] The median transformation of A refers to the ordered pairs consisting of a point in MA(A) and its distance to the boundary of A, which can be expressed by equation (6):

[0100]

[0101] Where MAT(A) is the skeleton of the central axis MA(A) of plane A, p is a point on the skeleton of plane A, q is an arbitrary point on the boundary of plane A, and r is the infimum of the Euclidean distance d(p,q) between p and q.

[0102] After extracting the skeleton based on the above central axis transformation, the laser line is only one pixel wide.

[0103] Step 6: Based on the calibrated skeleton image obtained from preprocessing, bright pixel searches can be performed on both left and right images for matching and depth calculation. Because the left and right images are row-aligned after calibration and correction, and the laser line after binarized skeleton extraction is only one pixel wide, accurate matching can be achieved by searching only a single pixel in one row when searching the left and right images. This solves the problem of mismatching or no matching for textureless objects in traditional binocular vision. After searching for pixels in the left and right images, the coordinates of the bright spots in each row are recorded, and then disparity calculation is performed. The coordinates x of the bright spot in the left image are obtained based on the maximum value. L And the brightness coordinate x in the right figure R The disparity d can be obtained as shown in equation (7):

[0104] d = x L -x R (7)

[0105] Step 7: Based on the binocular vision principle model diagram, as shown below. Figure 3 As shown in the figure, points pL and pR are the image points of point P in space on the imaging planes of the left and right cameras, respectively. The line segment x... L and x R denoted as , where are the distances from the left and right image points to the edge of the camera's imaging plane, f is the camera's focal length, and O is the distance from the left and right image points to the edge of the camera's imaging plane. L and O R Let P be the optical centers of the left and right cameras, respectively, and B be the distance between the optical centers of the left and right cameras, i.e., the system baseline length. Based on the triangle relationship, the distance z from a point P in space to the camera can be expressed as:

[0106]

[0107] Focal length and optical center distance are system parameters. Accurate values ​​are obtained through system calibration. Parallax is obtained through binocular stereo matching. Therefore, the depth of the spatial point is solved by equation (8), and then the depth of the object where the entire laser line is located is obtained.

[0108] Step 8: Solve for the thickness distribution along the laser line. The thickness distribution at each point along the laser line is determined based on the calculated object depth Z. X The object thickness *g* is obtained by subtracting the object's depth *Z* from the plane depth *Z*. Based on the measured object depth *Z*, and according to imaging principles, the actual spatial scale of each pixel differs at different depths. Further interpolation is then used to obtain the equally spaced object thickness distribution *g*. x .

[0109] Step 9: Cross-correlation calculation of displacement. The thickness distribution of an object moving with the conveyor belt is inconsistent between consecutive frames. Based on the thickness distribution of the object in the two frames, cross-correlation calculations are performed to obtain the displacement of the object between the two frames, and thus the object's velocity can be calculated. The principle of cross-correlation is as follows: For two consecutive frames with a sampling time interval of T, after thickness calculation, the thickness distribution g1 of the same object in the two frames is obtained. i With g2 i (i = 1, 2, 3, ..., N), their cross-correlation function R is:

[0110]

[0111] Where n is the number of displacement pixels, N is the length of the thickness distribution sequence, and the number of displacement pixels corresponding to the maximum value of the function is denoted as M. The velocity can be calculated as follows:

[0112]

[0113] Where v is the velocity of the object; M is the number of displacement pixels corresponding to the maximum value of the cross-correlation function; T is the sampling time interval between two consecutive frames; Z0 is the depth of the conveyor belt plane; Δx(z0) is the actual spatial distance of one pixel at a depth of z0.

Claims

1. A method for measuring the speed of objects conveyed by a belt based on binocular vision, applied to the measurement of the ash discharge speed of a dry ash discharger in a power plant boiler, characterized in that, include: Binocular camera and laser; The binocular camera is arranged in a parallel configuration, with two cameras positioned on either side of the conveyor belt. The laser is positioned between the binocular cameras and fixed on the same mounting bracket. The line connecting the two cameras is perpendicular to the direction of the conveyor belt speed being measured, and the emitted laser line is parallel to the direction of the conveyor belt speed. Two cameras capture images of a conveyor belt with laser lines attached; the laser lines are parallel to the speed direction of the conveyor belt. By matching images of the conveyor belt with laser lines captured by two cameras, the depth of the object containing the laser lines can be obtained. Based on the depth of the object where the laser line is located, interpolation is performed to obtain the equally spaced object thickness distribution; Based on the equally spaced object thickness distribution in two consecutive frames, the displacement of the object on the conveyor belt between the two frames is obtained. The velocity of the object is determined by the displacement of the object in the two frames before and after the object is obtained. The displacement of the object on the conveyor belt between two frames is obtained based on the equally spaced object thickness distribution in the two consecutive images, including: To perform cross-correlation calculation on the object thickness distribution of two consecutive frames, and obtain the number of displacement pixels M corresponding to the maximum value of the cross-correlation function; The number of displacement pixels M corresponding to the maximum value of the cross-correlation function is obtained, and the displacement L of the object on the conveyor belt between the previous and next frames is obtained: in, z 0 △ is the depth of the conveyor belt plane; x(z) 0 ） For depth z 0 One pixel corresponds to the actual spatial distance. In the step of obtaining the displacement of the object in two consecutive frames, the cross-correlation function used for cross-correlation calculation is: Where n is the number of displacement pixels; N is the length of the thickness distribution sequence; g1 i The thickness distribution of objects in the previous frame image g1 middle i The object thickness per displacement pixel g2 i For the thickness distribution of objects in the later frame image g2 middle i The object thickness per displacement pixel i = 1, 2, 3, ..., N ; The depth of the object containing the laser line is obtained by matching images of the conveyor belt captured by two cameras, including: Align and match the images captured by the two cameras, searching for a unique pixel in a row of the images captured by the two cameras; After searching for pixels in the images captured by the two cameras, the coordinates of the bright spots in each row are recorded to complete the parallax calculation: Where d is the parallax; x L The coordinates of the bright spot in the left image; x R The right figure shows the brightness coordinates; Calculate the depth of the object containing the laser line based on parallax: Where z is the depth of the object containing the laser line; f is the camera focal length; and B is the distance between the optical centers of the two cameras.

2. The method for measuring the speed of a belt-conveyed object based on binocular vision according to claim 1, characterized in that, The velocity of the object is to be determined as follows: in, v L is the velocity of the object; L is the displacement of the object between the two consecutive frames; T is the sampling time interval between the two consecutive frames.

3. A device for measuring the speed of objects conveyed by a belt based on binocular vision, characterized in that, include: A linear laser emits a laser line parallel to the direction of the conveyor belt's speed. A binocular camera, positioned on both sides of the conveyor belt, is used to acquire images of objects on the conveyor belt with laser lines. The processor, according to any one of the binocular vision-based belt conveyor speed measurement methods as described in claims 1-2, processes the image of the belt conveyor with laser lines acquired by the binocular camera to obtain the speed of the belt conveyor.

4. A computer-readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the binocular vision-based belt conveyor speed measurement method as described in any one of claims 1 to 2.

Images