A continuous loading system for rail scales based on point cloud recognition
By using a point cloud recognition-based continuous loading system for rail scales, sensors and a central controller are used to automatically adjust the operation of the chutes and jacks, solving the problems of low loading efficiency and low automation of rail scales, and achieving highly efficient and automated loading.
Patent Information
- Application Number
- CN202522562866.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-03
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2035-12-03
AI Technical Summary
The existing rail scale loading method is inefficient, has a low degree of automation, requires close cooperation from multiple operators, is prone to loading errors, and has high construction costs.
The continuous loading system based on point cloud recognition uses sensors and a central controller to monitor various parameters during the loading process, automatically adjusting the swing position, angle, and extension length of the chute. Combined with the tractor pulling the car forward or to a stop, it achieves fully automated loading.
It improved loading efficiency, reduced labor costs and labor intensity, achieved unmanned weighing and automatic measurement, reduced human intervention, and improved the degree of automation in loading.
Smart Images

Figure CN224953950U_ABST
Abstract
Description
Technical Field
[0001] This utility model relates to a continuous loading system for rail scales based on point cloud recognition, which is an automated transportation machinery and equipment, and a system for automated loading of bulk materials. Background Technology
[0002] There are two main existing methods for weighing trains for quantitative loading: one is quantitative loading that relies on weighing in a quantitative hopper above the carriage, and the other is quantitative loading that relies on weighing on a track scale below the carriage.
[0003] The former quantitative loading method offers fast loading speed, allowing the entire train to be loaded even at low speeds. It is highly efficient, highly automated, and requires minimal personnel. In some highly automated loading stations, it can even achieve unmanned operation. However, this loading method requires the construction of tall steel structures, resulting in high investment and costs, and is only suitable for loading scenarios with large throughput.
[0004] The second loading method is mainly used in situations where the loading volume is small. Its main advantages are low loading station height and low construction cost. The material is transported to the forklift compartment above the car by belt or scraper conveyor, and then unloading and loading can be done by operating the swinging scoop. The disadvantages are low loading efficiency, low degree of automation, and the need for more operators.
[0005] The main reason for the low loading efficiency of rail scales is that the timing control of its loading process is tightly linked with a lack of buffer. Multiple operators need to work closely together, and even a slight delay can cause material blockage at the unloading port of the feeding conveyor or insufficient loading of the already loaded section, requiring the car to be returned for replenishment. Even with experienced operators, as working hours increase, the strict coordination and excessive labor intensity (mental stress) can easily lead to errors in loading operations.
[0006] Improving the loading efficiency of rail scale loading, reducing human intervention during the loading process, and automating rail scale loading are problems that need to be solved. Summary of the Invention
[0007] To overcome the problems of existing technologies, this invention proposes a continuous loading system for rail scales based on point cloud recognition. The system integrates various elements of loading and unifies resource allocation, thereby achieving automation of rail scale loading.
[0008] The purpose of this utility model is achieved as follows: A continuous loading system for rail scales based on point cloud recognition includes a loading area and a weighing area arranged along the train direction. A feeding conveyor belt with a variable frequency speed controller is provided on one side of the loading area. A feeding hopper with a feeding gate is provided at the inlet of the feeding conveyor belt. A double-wing chute inlet is provided below the head of the feeding conveyor belt. The double-wing chute has a forked receiving hopper. The forked receiving hopper is hinged to a telescopic main chute equipped with a main chute swing driver and a main chute telescopic driver, and a secondary chute equipped with a secondary chute swing driver. The forked receiving hopper also has a flap with a flap driver for distributing material between the main chute and the secondary chute. Below the double-wing chute is a track through which a car pulled by an ox can pass. The weighing area is equipped with a rail scale, and below the double-wing chute is a device capable of scanning the car. The system includes a longitudinal scanning radar for the internal material pile; a main chute equipped with a main chute swing angle sensor and a telescopic position sensor; an auxiliary chute equipped with an auxiliary chute swing angle sensor; a flap rotation angle sensor on the rotating shaft of the flapper; car scanning radar groups on both sides of the track scale to determine the position of the car and the axle; a traction controller on the ox; electronic tags storing car information installed on the cars; and an electronic tag reader in the loading area. The system comprises a frequency converter, a feeding gate, a track scale, a main chute swing drive, a main chute telescopic drive, an auxiliary chute swing drive, a flapper drive, a longitudinal scanning radar, a main chute swing angle sensor, a telescopic position sensor, an auxiliary chute swing angle sensor, a flap rotation angle sensor, car scanning radar groups, a traction controller, and an electronic tag reader, all electrically connected to a central controller.
[0009] Furthermore, the cross-sectional shape of the main chute and the auxiliary chute is rectangular.
[0010] Furthermore, the main chute telescopic actuator is a hydraulic cylinder; the main chute swing actuator and the auxiliary chute swing actuator are winches and wire ropes or hydraulic cylinders; the flap actuator is an electric motor or hydraulic motor with a reducer.
[0011] Furthermore, the main chute swing angle sensor, the auxiliary chute swing angle sensor, and the flip plate rotation angle sensor are absolute encoders.
[0012] Furthermore, the central controller is an electronic device with storage and computing functions.
[0013] Furthermore, the aforementioned carriage scanning radar group includes a multi-line lidar installed at both the front and rear of the carriage.
[0014] The advantages and beneficial effects of this utility model are as follows: This utility model utilizes the connection of a track scale, a double-wing chute, a feeding belt conveyor, and an iron ox, and arranges multiple sensors to continuously monitor various parameters during the loading process. Based on these parameters, the swing position, angle, and extension length of the chute, as well as the forward or stop movement of the iron ox pulling the car, are adjusted, thereby achieving fully automated loading. This improves work efficiency while reducing labor costs and labor intensity. Attached Figure Description
[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0016] Figure 1 This is a schematic diagram of the system described in Embodiment 1 of this utility model;
[0017] Figure 2 This is a schematic diagram of the system described in Embodiment 1 of this utility model. Figure 1 PP section view;
[0018] Figure 3 This is a schematic diagram of the double-wing chute structure described in Embodiment 1 of this utility model;
[0019] Figure 4 This is a schematic block diagram of the system described in Embodiment 1 of this utility model;
[0020] Figure 5 This is a schematic diagram of the loading position of the method described in Embodiment 7 of this utility model;
[0021] Figure 6 This is a schematic diagram of the material pile in the main chute of the method described in Embodiment 9 of this utility model;
[0022] Figure 7 This is a schematic diagram of the material pile in the secondary chute of the method described in Embodiment 10 of this utility model. Detailed Implementation
[0023] Example 1:
[0024] This embodiment is a continuous loading system for rail scales based on point cloud recognition, such as... Figure 1 , 2As shown in Figures 3 and 4. The system described in this embodiment includes a loading area 1 and a weighing area 2 arranged along the train direction. A conveyor belt 3 with a variable frequency drive is provided on one side of the loading area. A feeding bin 4 with a feeding gate 401 is provided at the inlet of the conveyor belt. A double-wing chute 5 has an inlet 501 below the conveyor head. The double-wing chute has a forked receiving bin 502. The forked receiving bin is hinged to a telescopic main chute 503 with a main chute swing drive and a main chute telescopic drive, and a secondary chute 504 with a secondary chute swing drive. The forked receiving bin also has a flap 505 with a flap drive for distributing material between the main and secondary chutes. Below the double-wing chute is a track 7 through which a car 6 pulled by an ox can pass. The weighing area has a track scale 8. Below the double-wing chute is a longitudinal track capable of scanning the material pile inside the car. The system includes a scanning radar 9, a main chute swing angle sensor 10 and a telescopic position sensor 11 on the main chute, a secondary chute swing angle sensor 12 on the secondary chute, a flip-board rotation angle sensor 13 on the rotating shaft of the flip-board, a car scanning radar group 14 on both sides of the track scale to determine the position of the car and the wheel axle 601, a traction controller on the ox, electronic tags storing car information installed on the car, and an electronic tag reader in the loading area. The system comprises a frequency converter, a feeding gate, a track scale, a main chute swing drive, a main chute telescopic drive, a secondary chute swing drive, a flip-board drive, a longitudinal scanning radar, a main chute swing angle sensor, a telescopic position sensor, a secondary chute swing angle sensor, a flip-board rotation angle sensor, a car scanning radar group, a traction controller, and an electronic tag reader, all electrically connected to the central controller.
[0025] The loading system described in this embodiment is an automated loading station with a relatively small loading capacity, using double-wing chutes and a track scale to load train carriages. Because there are no buffer bins or metering bins, the overall height of the loading station is low, thus saving on the cost of steel structure frames and lifting costs. The feeding silo described in this embodiment can take various forms, such as a barrel-shaped silo or a stockpile. A gate is provided at the discharge port to allow for the discharge of materials. The belt conveyor described is usually not a simple single belt, but a combination of multiple belt conveyors; for convenience, this embodiment refers to the entire belt conveyor combination.
[0026] The material conveying section in this embodiment mainly consists of a double-wing chute, which includes a main chute and a secondary chute, as well as a branched receiving bin and flaps installed in the branched receiving bin. The flaps control whether the material flows out of the main chute or the secondary chute, or control the distribution of material flow in the main chute and the secondary chute by controlling the angle of the flaps.
[0027] The key difference between the double-wing chute in this embodiment and the conventional double-wing chute is that the open-top unloading hopper is upgraded to a four-sided enclosed, retractable, and scraping-capable square cylindrical chute. This allows for simultaneous loading and leveling of materials, achieving precise loading and leveling. Furthermore, the main chute's outlet is designed to be inclined. When the main chute is lowered, its outlet plane is parallel to the track plane. This design facilitates control of material flow, scraping, and leveling, and also helps in calculating the material filling volume of the car body.
[0028] The main chute swing drive, main chute telescopic drive, auxiliary chute swing drive, and flap drive are mechanical devices such as winches, hydraulic cylinders, and hydraulic motors, used to drive the chute's swinging, telescopic, and flap rotation.
[0029] Another key aspect of this embodiment is the inclusion of numerous sensors for controlling and adjusting the loading process, thereby achieving fully automated loading.
[0030] The carriage scanning radar group described in this embodiment typically consists of multiple lidar units installed on one or both sides of the railway. One lidar unit is usually installed at the beginning and end of the weighbridge, and one or two lidar units are installed before the carriages enter the loading area. These lidar units are used to count the number of carriages entering the loading area and to detect the carriages' speed using the Doppler effect. The above settings are for the carriage scanning radar units installed at the beginning and end of the weighbridge; each lidar unit has its own coordinate system. The carriage scanning radar group detects in real time whether the four axles of the carriages are on the weighbridge, overcoming the problem of existing detection technologies being affected by carriage reversing or minor carriage movements, thus achieving unmanned weighbridge operation and automatic metering. The carriage scanning radar group scans the carriages with the horizontal height of the radar as the dividing line; the upper half scans and identifies the carriages. Figure 2 The lower half of the α part scans and identifies the wheel axle. Figure 2 The β part.
[0031] The longitudinal scanning radar is installed between the main and auxiliary chutes of the double-wing chute, at the bottom of the bifurcated receiving bin. The longitudinal scanning radar intelligently identifies the material filling degree of the car body during the loading process, so as to realize the automatic control of the conveyor belt conveyor and the tractor traction.
[0032] The central controller is an electronic device with storage and processing capabilities, such as an industrial control computer. It is used to manage and control the entire system's material conveying, loading, and weighing operations, replacing manual labor and achieving automation.
[0033] Example 2:
[0034] This embodiment is an improvement on the above embodiment, and is a refinement of the cross-sectional shape of the main chute and the auxiliary chute in the above embodiment. The cross-sectional shape of the main chute and the auxiliary chute in this embodiment is rectangular.
[0035] The open-top sieve was replaced with a four-sided enclosed square cylindrical double-wing main and auxiliary chute. The main chute is a retractable and scraping chute. When loading, the lower lip of the chute is parallel to the track, which can achieve material leveling while loading.
[0036] Example 3:
[0037] This embodiment is an improvement on the above embodiment, and is a refinement of the above embodiment regarding the swing drive of the main chute, the auxiliary chute, and the flapper. In this embodiment, the main chute telescopic drive is a hydraulic cylinder; the main chute swing drive and the auxiliary chute swing drive are winches and wire ropes or oil cylinders; and the flapper drive is an electric motor or hydraulic motor with a reducer.
[0038] The wire rope offers fast and practical traction, while the hydraulic cylinder is stable and less prone to malfunction.
[0039] Example 4:
[0040] This embodiment is an improvement on the above embodiment and a refinement of the angle sensor in the above embodiment. The main chute swing angle sensor, the auxiliary chute swing angle sensor, and the flip plate rotation angle sensor described in this embodiment are absolute encoders.
[0041] The main and auxiliary chutes are connected to the bifurcated receiving bin via rotating bearings. An absolute encoder is installed at the rotating part of the bearing to obtain the chute swing angle. A rotating flap is installed in the bifurcated receiving bin, driven by a hydraulic cylinder. The flap is connected to the bifurcated receiving bin via bearings, and an absolute encoder is installed on the bearing to obtain the flap swing angle.
[0042] Example 5:
[0043] This embodiment is an improvement on the above embodiment and a refinement of the central controller in the above embodiment. The central controller described in this embodiment is an electronic device with storage and computing functions.
[0044] The central controller can be an industrial control computer or a similar electronic device.
[0045] Example 6:
[0046] This embodiment is an improvement on the above embodiment, and is a refinement of the above embodiment regarding the carriage scanning radar group. The carriage scanning radar group described in this embodiment includes a multi-line lidar set at the front and rear of the carriage.
[0047] The carriage scanning radar group is installed on one or both sides of the railway track. It can be several independent lidars, or it can be a grating group or a lidar with testing capabilities.
[0048] Example 7:
[0049] This embodiment is a continuous loading method for rail scales based on point cloud recognition using the above system.
[0050] The loading method described in this embodiment replaces the work of at least five people, including: a weighbridge operator, a loading operator, a tractor traction operator, and a leveling operator. The weighbridge operator is responsible for monitoring the weighbridge and confirming the measurement after ensuring all four axles of the wagon are on the scale. Generally, the weighbridge is equipped with an electromagnetic counting device to determine the number of axles on the scale. However, if the counting is inaccurate due to the wagon reversing or slight displacement causing the electromagnetic counting to malfunction, the weighbridge operator needs to adjust the axle counting operation promptly. The loading operator is responsible for operating the winch to adjust the swing angle of the two unloading scoops to regulate the unloading volume. The tractor traction operator is responsible for controlling the train speed, visually observing the material pile in real time, adjusting the wagon position as needed to match the loading speed, and working with the loading operator to ensure loading quality. The leveling operator is responsible for leveling the material pile after loading, smoothing out any protrusions to meet the requirements for leveling during loading. This embodiment completely replaces manual operation through the following steps:
[0051] Step 1, Loading Preparation: The wagon approaches the loading area (below the double-wing chute) under the traction of the tractor. The central controller receives the wagon attribute information stored in the electronic tag. The system is ready to load the next wagon, either as the first or during the loading process. At this time, the main chute is fully retracted, and the auxiliary chute is lowered at a fixed angle (this applies to wagons other than the first; if it's the first wagon, it's retracted). The main chute is closed, and the auxiliary chute is open. A coordinate system is established with the direction of travel as the y-axis. The central controller calculates the starting unloading positions A1 and B1 for the main chute and auxiliary chute based on the material and wagon information. The starting unloading position for the main chute is when the front sill of the wagon reaches point A1, and the starting unloading position for the auxiliary chute is when the front sill reaches point B1. Positions A1 and B1 can be determined based on material information and actual measurements. Figure 5 As shown.
[0052] Step 2, Start the secondary chute unloading: When the front chute reaches position B1, the flap is hit to the fork side of the main chute, and the secondary chute is lowered (if the car entering the loading area is the first car, the secondary chute needs to be lowered; if the current car is not the first car, the secondary chute will always be lowered, so no lowering action is needed), open the feeding gate of the feeding hopper and the feeding conveyor belt, and the tractor pulls the car to continue moving forward. The material is unloaded into the car through the secondary chute. The longitudinal scanning radar scans the material pile under the discharge port of the secondary chute to identify the height of the pile, which is then used to control the speed of the tractor pulls the car and the conveying rate of the feeding conveyor belt.
[0053] The position detection of the front panel of the carriage can be achieved through various recognition methods, such as array-type grating line point detection, laser point cloud recognition, or AI vision.
[0054] Step 3, Rapid Unloading from the Main Sluice: When the front chute reaches position A1, lower the main sluice (first section). The extension length of the main sluice is determined based on the car model and the material being unloaded; this extension length is the scraping height. Position the flapper to the branching side of the secondary sluice. The feeding conveyor maintains a stable feeding speed. The tractor pulls the car forward, unloading the bulk material through the main sluice into the car. The material rapidly accumulates in the car and is leveled by the lower lip of the main sluice. During unloading, the longitudinal scanning radar continuously scans the material accumulation in the car, identifying the fullness of the material pile using point cloud data. The central controller adjusts the tractor's speed and the conveyor's conveying rate based on the material accumulation status; this speed is called the adaptive speed. Simultaneously, the car scanning radar group monitors the car's movement position and the balance of the car's wheel axles. The car model can be obtained using RFID-based identification of electronic tags on the bottom of the car.
[0055] The longitudinal scanning radar installed above the loading area detects the fullness of the material pile to determine the train traction speed. At the same time, the loading speed is kept constant to ensure the continuity of loading.
[0056] As described in the method of identifying the material filling degree of the truck compartment during the loading process using laser point cloud, when the material filling degree is 60%, the rated traction speed (matching the rated loading speed) is maintained. When it is higher or lower than 60%, the speed is reduced or accelerated linearly according to the difference to achieve uniform material loading.
[0057] Step 4, detailed preparation of materials for the carriage: Detailed preparation of materials for the carriage includes the following sub-steps:
[0058] Sub-step 401, switch from fast unloading to slow unloading: After the weighbridge on the third wheel axle of the car is checked, stop or reduce the traction speed of the tractor (stop or slow down to wait for fine distribution). At this time, the traction speed is no longer controlled by the material filling degree, but by the weighing data of the track scale as the control parameter of the flap control, which is used as the basis for adjusting the unloading speed of the main chute. At this time, the front sill of the next car has passed the B1 position. At this time, the feeding speed remains unchanged. Move the flap to the middle position of the fork, reduce the unloading speed of the main chute, and unload the separated part of the material from the auxiliary chute to the next car. If it is the last car, lift the auxiliary chute and reduce the feeding rate of the feeding belt.
[0059] Sub-step 402, slow unloading: Since the fourth axle has not yet been weighed, the weighing data on the track scale is lower than the actual weighing data. When the weighing value on the track scale approaches the expected loading amount (e.g., reaching the expected loading amount W), N 90%, W N The train is slowly pulled at 1 / 4 of the loading speed (the speed of the carriage when the main chute is unloading quickly) until the fourth axle is on the balance, at which point the ox stops pulling the carriage.
[0060] The above data, including the tipping position, 8.0 tons, and 1 / 4 speed, were determined based on actual loading experience.
[0061] Sub-step 403, fine batching: Move the flap to the 1 / 4 position of the fork (at this time, the main chute discharge occupies 1 / 4 of the material and the auxiliary chute discharge occupies 3 / 4 of the material), and unload the material from the main chute to the car for fine batching. Wait until the track scale weighing data reaches W. N At that time, the flap is fully pushed to the fork side of the main chute, and the material is completely unloaded and loaded onto the next section by the auxiliary chute.
[0062] During the fine mixing process, the system needs to make safety adjustments based on the material height at the discharge port of the next auxiliary chute. For example, if the material height at the discharge port of the next auxiliary chute exceeds the safety value, the auxiliary chute needs to be raised at a certain angle to avoid overflow.
[0063] Step 5, End the loading of the current car: The main chute is fully raised, all four axles are on the scale and stabilized for 3-5 seconds, then the weight of the track scale is read to form complete loading data, and the loading cycle continues to the next car. If it is the last car, the loading of the whole car is completed.
[0064] When loading the last car, the auxiliary chute is no longer used. Instead, the feeding rate is adjusted by the feeding belt conveyor to complete the fine and precise material distribution. An intermittent feeding method is adopted to achieve the goal of emptying the conveyor belt while loading is completed. The specific process is as follows:
[0065] When loading the last car begins, loading is carried out at an appropriate speed (rapid loading). At the same time, the material accumulation and wheel axle weighing status of the last car are monitored. When the car's loading capacity (by car volume) reaches 50%, intermittent unloading is adopted. That is, before the third wheel axle is weighed, the feeding gate is opened or closed intermittently, and the tractor traction speed is adjusted. After the third wheel axle is weighed, the opening and closing frequency of the feeding gate is increased until the fourth wheel axle is weighed. After the fourth wheel axle is weighed, the tractor stops traction and intermittent material conveying is adopted to accurately reach the expected loading volume.
[0066] Specific instructions for intermittent unloading:
[0067] Taking the C80 car body as an example (car body load 80 tons).
[0068] Based on the length from the feed hopper opening above the conveyor belt to the discharge point, estimate the load W of the conveyor belt from full capacity to empty capacity. n Let's set it at 20 tons. When loaded to half the length of the car body (half of 80 tons is 40 tons, leave 8 tons for fine-tuning, which should be greater than W),... n That is, (80 / 2-8)>W n Turn off the feeder supplying material to the conveyor belt. After the load on the conveyor belt is cleared (the timing time is the length from the feed hopper opening to the conveyor head divided by the rotational speed), the timing is measured intermittently in multiple t intervals.N (Time required for the feeder to supply 2.0 tons of material to the belt) The feeder is started intermittently, and the traction speed is adjusted in real time according to the material pile-up situation until the 3 shafts are on the scale.
[0069] Once the third axle is balanced, traction is stopped, and the timing is changed to multiple intermittent t intervals. N / 2 The feeder is started intermittently, and when the weight value of the track scale approaches (W) N When the load reaches -8.0 tons, the train is pulled slowly until the fourth axle is on the scale, at which point the train traction stops.
[0070] When the fourth axis is balanced, the timing is changed to be measured in multiple intermittent intervals (t). N / 4 Start the feeder to feed material intermittently until the rated value W is reached. N Turn off the feeder, stop the belt conveyor, lift the main chute, and the last car is loaded. Ideally, at this point, there should be no material left on the belt or only a small amount of material left.
[0071] Example 8:
[0072] This embodiment is a refinement of Embodiment Seven, specifically detailing the methods for detecting the carriage position and axle position. The carriage position and axle position detection methods described in this embodiment include:
[0073] The laser point cloud is used to identify the two axes at the tail of the scale, either the first two axes (axes 1 and 2) or the last two axes (axes 3 and 4), and the distance of axes 3 and 4 from the tail of the scale is measured in real time to achieve the purpose of unmanned weighing and automatic measurement.
[0074] a. Whether the current axle is the front two wheels (1st and 2nd axles) or the rear two wheels (3rd and 4th axles) of the carriage;
[0075] b. Whether the third round was on the balance, whether the fourth round was on the balance, and how far from the end of the balance;
[0076] For wheel and axle detection, the existing method uses electromagnetic induction switches installed on the track sleeper beams at the beginning and end of the weighbridge to detect the wheels (metal). When a wheel passes the induction switch, the wheel and axle number is incremented or decremented by 1 to obtain the number of wheel and axles on the weighbridge. This method is susceptible to counting errors caused by slight movements of the carriage when the vehicle reverses or the wheels approach the induction switch, requiring manual calibration by the weighbridge operator. Therefore, this embodiment proposes using laser point cloud recognition to identify vehicle wheel and axles. The timing for weighing data is determined by real-time identification of the wheel and axle positions at the beginning and end of the weighbridge. A multi-line scanning laser (16 lines or more) is installed directly opposite the beginning and end of the weighbridge, at a distance of at least 0.5 meters from the carriage floor (the greater the distance, the better), and at a height flush with the carriage floor. The more scanning lines of the laser radar, the wider its vertical angle. A laser radar with a vertical scanning angle range that meets the requirements can be selected based on the distance between the radar and the carriage.
[0077] Taking a 64-line lidar as an example, the vertical scanning angle includes positive 30 degrees and negative 45 degrees. Positive angle laser scanning lines are used to identify the front or rear sills of the vehicle body, while negative angle laser scanning lines are used to identify the wheel axles.
[0078] Point cloud data acquired through lidar scanning includes the subdivided horizontal and vertical angles of each lidar point in space, as well as the three-dimensional coordinate data (x, y, z values) of that point. Wheel axle analysis can be performed by analyzing the lidar point clouds scanned to the sides of the vehicle and wheels. The three-dimensional lidar point cloud data follows the right-hand rule: the z-value of a positive vertical angle point cloud is greater than 0, and the z-value of a negative vertical angle point cloud is less than 0. The direction directly in front of the radar is the positive x-axis, and the direction behind the radar is the negative x-axis. The left side of the horizontal direction of the radar is the positive y-axis, and the right side is the negative y-axis.
[0079] Let the distances from the radar to the carriage and the axle be L respectively. 01 (Based on the average width of different car body models, and according to safety clearance requirements, its value is greater than 0.5 meters) L 02 (Based on the average value of the concave and convex points of the axle and its surrounding components, which is greater than 0.7 meters), the radar distance from the orbital height is H0, the axle diameter is R, and the distance between the two axles is L. 03 All data were obtained from on-site measurements.
[0080] The point cloud features can be summarized from the explicit physical characteristics of the carriages and axles as follows:
[0081] (1) Wheel and axle sets are arranged from the bottom plate of the carriage down to the track space. Each carriage has 4 pairs of wheel and axles. Every two pairs of wheel and axles and auxiliary devices form a wheel and axle set. The two wheel and axles near the front rail of the carriage are the front wheel and axles near the rear rail of the carriage are the rear wheel and axles (rear wheel and axle sets). The laser point cloud falling on the wheel and axle sets is presented in the form of a patch area, and its x value is approximately L. 02 It is clearly distinguishable from the tubes / rods under the carriage (whose x value is approximately L).02 +0.5).
[0082] (2) Above the floor of the carriage are the side panels. The laser point cloud falling on the side panels is presented in the form of a patch area, with an x-value of approximately L. 01 The difference in feature values between point clouds falling within empty or non-existent carriages lies in the fact that the x-value of the point cloud falling within an empty or non-existent carriage is greater than (L). 01 +0.2).
[0083] (3) The front axle assembly is close to the front sill of the carriage, and the rear axle assembly is close to the rear sill of the carriage. As described in (1), when the axle assembly is detected, the radar is checked for gaps or no carriages on the left and right sides respectively in the manner described in (2). If there is a gap or no carriage on the left side of the axle assembly, it indicates that the axle assembly is the front axle assembly of the carriage. Similarly, if there is a gap or no carriage on the right side of the axle assembly, it indicates that the axle assembly is the rear axle assembly of the carriage.
[0084] Detection Step 1: First, select the point cloud in the negative z-axis direction in real time for analysis. Use the point cloud to identify the wheel and axle group (the feature value x of the laser point cloud falling on the wheel and axle group is approximately L). 02 The process involves collecting point clouds that meet certain criteria, and then sequentially searching the point clouds on the collected axle wheel assembly at horizontal subdivision angles (every 0.2 degrees). The search criteria are: at least 20 laser points (limited to a range of +50 degrees to -50 degrees) are found along the positive or negative y-axis at a given laser point (let's call it point M). If more than 85% of the x-values of these laser points are greater than L... 02 +0.2, for example: 17 out of 20 x values are greater than L. 02 +0.2 indicates that the laser point M is located at the front or rear end of the wheel assembly. Based on the wheel radius R, the y-coordinates of the contact points between the two wheels and the track can be determined: If point M is at the front end of the wheel assembly, the y-coordinate of the front wheel axle (y... M -R), y M Here is the y-axis coordinate of laser point M, and the y-coordinate of the rear wheel axle of the wheel assembly (y... M -RL 03 If point M is the rear end of the wheel assembly, the y-coordinate of the front wheel axle of the wheel assembly (y... M +R+L 03 ), the y-coordinate of the rear wheel axle of the wheel set ((y M +R).
[0085] Detection Step 2: After locating the wheel axle assembly, analyze the point cloud in real time along the positive z-axis. First, use the point cloud to identify the side panel of the carriage (the feature value x of the laser point cloud falling on the side panel of the carriage is approximately L). 01And the quantity reaches a certain amount) and collect point clouds that meet the conditions. On the collected axle wheel group, the point cloud is polled sequentially (every 0.2 degrees) according to the horizontal subdivision angle. The search condition is a certain laser point (let's say point N), and 20 laser points are searched along the positive or negative y-axis at horizontal angles (limited to a range of positive 50 degrees to negative 50 degrees). If the x-value of these 20 laser points is greater than (L 01 If the number of +0.2) is greater than 17, it can be determined whether the laser point N is located at the front or rear of the carriage, and at the same time, it can be determined whether the wheel and axle assembly is the front wheel and axle assembly or the rear wheel and axle assembly.
[0086] Detection step 3: After determining the front / rear axle of the wheel and axle assembly, and identifying whether it is the front or rear wheel and axle assembly, it can be determined whether the front / rear axle is the 1st / 2nd axle or the 3rd / 4th axle. The radar is installed directly opposite the head or tail of the scale. The distance between the axle and the head / tail of the scale can be obtained from detection step 1.
[0087] Example 9:
[0088] This embodiment is an improvement upon the above embodiment, refining the method of using laser point clouds to identify the fullness of materials in the loading compartment during the loading process. The method of using laser point clouds to identify the fullness of materials in the loading compartment during the loading process described in this embodiment includes:
[0089] The filling degree of the cargo compartment is identified by laser point cloud during the loading process, and the train traction speed is adjusted according to the filling degree to achieve automatic control of the tractor traction speed.
[0090] The physical characteristics of the main chute during unloading are determined based on the actual loading operation conditions, such as... Figure 6 As shown:
[0091] a. The downward angle of the main chute is determined by the scraping characteristics of the main chute discharge port, that is, the material outlet of the main chute must be parallel to the track. Conventional double-wing chute main or auxiliary chute is scoop-shaped, that is, the top is open. However, in order to effectively control the flow of materials, this case uses a chute with a closed top and a rectangular cross-section. This allows the chute to not only limit the material flow but also to scrape and level the material. This requires that the material outlet of the main chute be parallel to the track, that is, the line connecting the front lip and the rear lip of the chute is parallel to the track plane.
[0092] b. Depending on the type of wagon and the type of material to be loaded, the expansion and contraction of the main chute during unloading and scraping can be determined based on actual experience. Since the main chute opening is responsible for scraping and leveling material, it effectively controls the height of the material pile within the wagon. Therefore, the expansion and contraction of the chute needs to be adjusted during loading to regulate the height of the material pile. Factors affecting the expansion and contraction include the height of the wagon's side rails and the unit weight of the material.
[0093] c. Determine the spatial coordinates of the line segment L(AB) between the front lip A and the rear lip B of the main chute; here, the front and rear of the line segment L(AB) are determined based on the magnitude of the y-coordinate value (absolute value), such as... Figure 4 As shown, the origin of the longitudinal scanning radar is on the right side of the main chute. Therefore, the value of point A on the front lip of the main chute is greater than the value of point B on the rear lip, so it is set as "front" and the corresponding point B is set as "rear".
[0094] d. Determine the lateral tilt angle of the material piled in the car based on the current angle of repose, and then determine the generatrix L(CD) of the tilted surface of the material pile; since the angle of repose of the same material is constant when piled up, i.e., Figure 4 The angle between line segment L (CD) and the car floor is fixed.
[0095] e. The longitudinal scanning radar determines the intersection point E of the extension line of line segment L(CD) and line segment L(AB) to determine whether the traction speed of the tractor is reasonable. The length of line segment L(AE) relative to the length of line segment L(AB) is a representation of the material filling degree. The closer point E is to point B, the more likely the material in the chute is to be nearly full, and the train needs to be tractioned to move forward. Otherwise, the material in the bifurcated receiving bin above the chute may overflow. Conversely, the closer point E is to point A, the more likely the material in the chute is to be empty, and the train traction speed needs to be reduced or the train needs to be stopped. Otherwise, the material may be piled up in the loading process, resulting in insufficient loading.
[0096] Example 10
[0097] This embodiment is an improvement upon Embodiment Nine, detailing the point cloud filtering process for longitudinal scanning radar in Embodiment Nine. The point cloud filtering process for longitudinal scanning radar described in this embodiment includes:
[0098] The spatial coordinates of line segment L(CD) are obtained by a lidar scanning the material pile slope below the main chute along the longitudinal direction of the car body. A multi-line lidar (longitudinal scanning lidar) is suspended below the double-wing chute. As the car body passes through the main chute, the lidar scan obtains the laser point cloud on the material pile slope below the main chute inlet. The point cloud selection process includes: first, acquiring the point cloud along the laser scan line that runs longitudinally through the car body and is close to the center of the car body; then, selecting points based on their distance from the car body floor from 1.0 meter to 2.0 meters. Figure 4 The line segment L(CD) can be obtained by performing discrete point fitting on the selected point cloud.
[0099] As the car passes through the secondary chute, the longitudinal scanning radar acquires the material stacking status of the secondary chute within the car, such as... Figure 7As shown, since the secondary chute is only used as a transition point between the two carriages during unloading, the unloading is done in a scattered, piled-up manner. The secondary chute unloads at a fixed angle, with point G below the unloading port as the landing point. The height of point F, located 0.5 meters along the y-axis, is used as a representation of the pile height. When the height at point A is detected to be too high (e.g., reaching 50% of the expected pile height in the carriage), the tractor needs to be immediately ordered to pull the train forward to prevent the pile from becoming too high and interfering with the unloading of the main chute.
[0100] Finally, it should be noted that the above is only used to illustrate the technical solution of this utility model and not to limit it. Although this utility model has been described in detail with reference to the preferred arrangement, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solution of this utility model (such as the form of loading station and chute, the characteristics of various materials, the order of steps, etc.) without departing from the spirit and scope of the technical solution of this utility model.
Claims
1. A continuous loading system for rail scales based on point cloud recognition, comprising a loading area and a weighing area arranged along the train direction, wherein a feeding conveyor belt with a variable frequency speed controller is provided on one side of the loading area, a feeding bin with a feeding gate is provided at the inlet of the feeding conveyor belt, a double-wing chute inlet is provided below the head of the feeding conveyor belt, the double-wing chute is provided with a forked receiving bin, the forked receiving bin is hinged to a telescopic main chute and a main chute with a swing drive and a telescopic drive, and a secondary chute with a swing drive, the forked receiving bin is also provided with a flap with a flap drive for distributing material between the main chute and the secondary chute, a rail is provided below the double-wing chute for a car pulled by an ox to pass through, and a rail scale is provided in the weighing area, characterized in that... The double-wing chute is equipped with a longitudinal scanning radar capable of scanning the material pile inside the car. The main chute is equipped with a main chute swing angle sensor and a telescopic position sensor. The auxiliary chute is equipped with an auxiliary chute swing angle sensor. The rotating shaft of the flap is equipped with a flap rotation angle sensor. The two sides of the track scale are equipped with car scanning radar groups to determine the position of the car and the position of the wheel axle. The ox is equipped with a traction controller. The car is equipped with an electronic tag that stores the car information. The loading area is equipped with an electronic tag reader. The frequency converter, feeding gate, track scale, main chute swing drive, main chute telescopic drive, auxiliary chute swing drive, flap drive, longitudinal scanning radar, main chute swing angle sensor, telescopic position sensor, auxiliary chute swing angle sensor, flap rotation angle sensor, car scanning radar group, traction controller, and electronic tag reader are electrically connected to the central controller.
2. The system according to claim 1, characterized in that, The main chute and the auxiliary chute have rectangular cross-sectional shapes.
3. The system according to claim 2, characterized in that, The main chute telescopic actuator is a hydraulic cylinder; the main chute swing actuator and the auxiliary chute swing actuator are winches and wire ropes or oil cylinders; the flap actuator is an electric motor or hydraulic motor with a reducer.
4. The system according to claim 3, characterized in that, The main chute swing angle sensor, the auxiliary chute swing angle sensor, and the flip plate rotation angle sensor are absolute encoders.
5. The system according to claim 4, characterized in that, The central controller is an electronic device with storage and computing functions.
6. The system according to claim 5, characterized in that, The aforementioned carriage scanning radar group includes a multi-line lidar installed at the front and rear of the carriage.
