A method for automatic control of a truck loading process based on a TOF camera

Abstract

The application discloses a kind of automatic control method of truck loading process based on TOF camera, comprising: S100, through TOF_1 camera acquisition cargo box position data and loading point position data;S101, cargo box position data and loading point position data are analyzed, and the azimuth difference is obtained and generates loading point parking offset distance data and loading point parking offset angle data;S102, loading point parking offset distance data and loading point parking offset angle data are used as input, and the control result is output through position control model;S200, through TOF_1 camera acquisition parking stationary duration data;S201, stationary duration data are judged to vehicle, and generate determination result;S300, through TOF_1 camera acquisition cargo box space size data, and preprocessing obtains cargo box cross section data;S301, cargo box cross section data are analyzed, and generate loading quantity estimate, start to carry out primary loading.

Description

Technical Field

[0001] This invention relates to the field of transportation management technology, and specifically to an automatic control method for truck loading process based on a TOF camera. Background Technology

[0002] Traditional loading processes typically rely on manual operation, which has certain limitations, including operator fatigue, uncertainty in subjective judgment, and working conditions in harsh environments. Meanwhile, during the loading process, accuracy is crucial to ensure accurate material loading. Traditional methods may not be able to meet the requirements of high precision and efficiency. However, an automatic control system based on a TOF camera can provide more accurate and real-time data, thereby improving the accuracy and efficiency of the loading process.

[0003] Although automatic control systems using TOF cameras are already in use during truck loading, TOF cameras still cannot accurately control the amount of material being loaded. In addition, most TOF cameras only monitor whether there is overloading of goods, but different goods have different loading characteristics. For example, sand is loaded in a cone shape, which results in a lot of wasted filling space in the cargo box or on the cargo plate when the height is detected to be too high.

[0004] To address the aforementioned issues, an automatic control method for truck loading process based on a TOF camera is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic control method for truck loading process based on a TOF camera, so as to overcome the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The automatic control method for truck loading process based on TOF camera includes a TOF_1 camera and a TOF_2 camera installed on the side wall of the material cylinder, and specifically includes the following steps:

[0008] S100: Collect cargo box location data and loading point location data using a TOF_1 camera;

[0009] S101. Perform data analysis on the cargo box location data and the loading point location data, obtain the orientation difference, and generate loading point docking offset distance data and loading point docking offset angle data.

[0010] S102. Take the data of the stopping offset distance and the stopping offset angle of the loading point as input, and output the control result through the position control model.

[0011] S200: Collects parking and stationary duration data using a TOF_1 camera;

[0012] S201. Determine whether a vehicle is stationary based on the stationary duration data and generate a determination result.

[0013] S300: Collects cargo box space dimension data through TOF_1 camera and preprocesses to obtain cargo box cross-sectional data;

[0014] S301. Analyze the cross-sectional data of the cargo box, generate an estimated loading quantity, and start the first-level loading;

[0015] S400: Collects the maximum height data of the goods after the first stage of loading is completed using a TOF_1 camera, and obtains the remaining cargo box size data using a TOF_2 camera;

[0016] S401. Analyze and process the maximum height data of the goods, generate loading compensation data, and perform loading compensation.

[0017] S402. Analyze and process the remaining cargo box size data, generate data for adjusting the feeding amount, and adjust the feeding amount accordingly.

[0018] S403. Perform secondary analysis and processing on the remaining cargo box size data, generate step distance control data and perform step distance control until the remaining cargo box size data in the cargo box is 0, and output the "Completed loading" label.

[0019] Furthermore, the cargo box position data is the cargo box position collected in real time by a TOF_1 camera, with the loading point docking line as the reference target. The loading point position is the loading point docking line data collected in real time by a TOF_1 camera. The loading point docking offset angle data is the angle difference between the cargo box position and the loading point docking line, with the loading point docking line as the reference target. The loading point docking offset distance data is the maximum distance between the cargo box position and the loading point docking line, with the loading point docking line as the reference target.

[0020] Furthermore, the construction steps of the position control model are as follows:

[0021] Step Q1: Obtain n sets of historical material point stopping offset distance data and historical material point stopping offset angle data as test classes, and collect the corresponding historical control distance and historical control angle of n test classes as control results;

[0022] Step Q2: Use n test classes as input attribute values ​​for the position control model, use the control results as the output of the position control model, use each control result as a prediction target, and use the prediction accuracy of the position control model as the training target.

[0023] Step Q3: Train the machine learning models for n test classes. Based on the prediction accuracy threshold L set in advance before the training step, stop training when the prediction accuracy reaches the prediction accuracy threshold ki o; and mark the machine learning model that has been trained for the nth test class as ki o. The position control model is a deep neural network model, decision tree or support vector machine.

[0024] Step Q4: Send the trained position control model kio to the processor and store it.

[0025] Furthermore, the logic for determining vehicle stillness based on stationary duration data is as follows:

[0026] The stationary duration Gx is collected using a TOF_1 camera. Stationary duration comparison thresholds Tz1 and Tz2 are set, where Tz1 is less than Tz2. If the stationary duration Gx is greater than 0 and less than Tz1, a "Stationary adjustment in progress" label is generated; if the stationary duration Gx is greater than Tz1 and less than Tz2, a "Adjustment about to be completed" label is generated; and if the stationary duration Gx is greater than Tz2, a "Adjustment completed" label is generated.

[0027] Furthermore, the spatial dimension data includes cargo box depth data and cargo box width data. The preprocessing performs a product operation on the cargo box depth and cargo box width to generate cargo box cross-sectional data.

[0028] Furthermore, the steps for generating the estimated material loading volume are as follows:

[0029] By connecting to the network, the stacking coefficients of different cargo materials are obtained, the stacking coefficient of the currently loaded cargo is retrieved, and the cross-sectional area data of the cargo box is multiplied with the stacking coefficient of the currently loaded cargo.

[0030] Furthermore, the maximum height data is the maximum stacking height of the goods after they are placed in the material cylinder (100), and the remaining cargo box size data includes the visible remaining cargo box size and the total cargo box size.

[0031] Furthermore, the logic for generating the material feeding compensation data is as follows:

[0032] The compensation coefficients for different cargo materials are obtained through network connection. The compensation coefficients obtained during the current loading are retrieved. The maximum height data of the cargo is subtracted from the cargo box depth data obtained by the TOF_1 camera. The calculation result is multiplied with the compensation coefficient of the current cargo to generate the loading amount compensation data.

[0033] Furthermore, the logic for generating the controlled feeding amount data is as follows:

[0034] The remaining cargo box size data is integrated and analyzed, and the estimated amount of material above the commercial vehicle is obtained by dividing the visible remaining cargo box size by the total cargo box size, and the material compensation data is added together.

[0035] Furthermore, the logic for generating the step distance control data is as follows:

[0036] A secondary analysis is performed on the remaining cargo box size data. The stepping coefficients of different cargo materials are obtained through the network. The stepping coefficient of the current cargo material is multiplied by the quotient of the visible remaining cargo box size divided by the total cargo box size. The result is then multiplied by the estimated loading quantity to obtain the product.

[0037] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0038] In this invention, by setting up two sets of TOF cameras and constructing a position control model, accurate stopping and stationary determination of the loading truck are achieved during the time period from before the loading truck stops to after the loading is completed. In addition, by collecting various types of data from the cargo box end through the two sets of TOF cameras, precise real-time control of the loading amount, loading amount compensation, step distance control, and loading stop control of the cargo in the cargo box is achieved. At the same time, the automatic control method based on TOF cameras can realize the automation of the truck loading process, which not only improves work efficiency but also reduces the interference of human factors and the possibility of errors. Automated control also helps to improve the production capacity of the entire loading system. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0040] Figure 1 This is a flowchart illustrating the steps of an automatic control method for truck loading process based on a TOF camera according to the present invention.

[0041] Figure 2 This is a demonstration diagram of an embodiment of the automatic control method for truck loading process based on a TOF camera according to the present invention;

[0042] Figure 3 This is a bidirectional structural diagram of the material cylinder in an automatic control method for truck loading process based on a TOF camera according to the present invention.

[0043] Figure 4 This is a diagram showing the installation structure of the material cylinder and the TOF camera in an automatic control method for truck loading process based on a TOF camera according to the present invention.

[0044] Figure 5This invention relates to the first-level loading operation of an automatic control method for truck loading process based on a TOF camera.

[0045] Figure 6 This is a diagram illustrating the secondary loading operation of an automatic control method for truck loading process based on a TOF camera, according to the present invention.

[0046] Figure 7 This is a diagram showing the actual output of a TOF camera in an automatic control method for truck loading process based on a TOF camera, according to the present invention.

[0047] In the diagram: 100, material cylinder; 101, TOF_1 camera; 102, TOF_2 camera. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 As shown in this embodiment, an automatic control method for truck loading process based on a TOF camera includes a TOF_1 camera 101 and a TOF_2 camera (102) disposed on the side wall of the material cylinder 100. The specific method includes the following steps:

[0050] S100: Collect cargo box location data and loading point location data using TOF_1 camera 101;

[0051] The cargo box position data is the cargo box position collected in real time by a TOF_1 camera 101 with the loading point docking line as the reference target. The loading point position is the loading point docking line data collected in real time by a TOF_1 camera 101.

[0052] S101. Perform data analysis on the cargo box location data and the loading point location data, obtain the orientation difference, and generate loading point docking offset distance data and loading point docking offset angle data.

[0053] The loading point stopping offset angle data is the angle difference between the cargo box position and the loading point stopping point, with the loading point stopping line as the reference target. The loading point stopping offset distance data is the maximum distance between the cargo box position and the loading point stopping point, with the loading point stopping line as the reference target.

[0054] S102. Take the data of the stopping offset distance and the stopping offset angle of the loading point as input, and output the control result through the position control model.

[0055] The construction steps of the position control model are as follows:

[0056] Step Q1: Obtain n sets of historical material point stopping offset distance data and historical material point stopping offset angle data as test classes, and collect the corresponding historical control distance and historical control angle of n test classes as control results;

[0057] Step Q2: Use n test classes as input attribute values ​​for the position control model, use the control results as the output of the position control model, use each control result as a prediction target, and use the prediction accuracy of the position control model as the training target.

[0058] Step Q3: Train the machine learning models for n test classes. Based on the prediction accuracy threshold L set in advance before the training step, stop training when the prediction accuracy reaches the prediction accuracy threshold ki o; and mark the machine learning model that has been trained for the nth test class as ki o. The position control model is a deep neural network model, decision tree or support vector machine.

[0059] Step Q4: Send the trained position control model kio to the processor and store it.

[0060] S200: Collects parking time data using TOF_1 camera 101;

[0061] S201. Determine whether a vehicle is stationary based on the stationary duration data and generate a determination result.

[0062] The logic for determining vehicle stillness based on stationary duration data is as follows:

[0063] The stationary duration Gx is collected using a TOF-1 camera 101. Stationary duration comparison thresholds Tz1 and Tz2 are set, where Tz1 is less than Tz2. If the stationary duration Gx is greater than 0 and less than Tz1, a "Stationary adjustment in progress" label is generated; if the stationary duration Gx is greater than Tz1 and less than Tz2, a "Adjustment about to be completed" label is generated; and if the stationary duration Gx is greater than Tz2, a "Adjustment completed" label is generated.

[0064] S300: The TOF_1 camera 101 collects cargo box space dimension data and preprocesses it to obtain cargo box cross-sectional data;

[0065] The spatial dimension data includes cargo box depth data and cargo box width data. The preprocessing performs a product operation on the cargo box depth and cargo box width to generate cargo box cross-sectional data.

[0066] S301. Analyze the cross-sectional data of the cargo box, generate an estimated loading quantity, and start the first-level loading;

[0067] The steps to generate the estimated material feeding quantity are as follows:

[0068] By acquiring the stacking coefficients of different cargo materials through network connection, retrieving the stacking coefficient of the currently loaded cargo, and multiplying the cross-sectional area data of the cargo box with the stacking coefficient of the currently loaded cargo, the specific formula for generating the estimated loading quantity is as follows:

[0069] γ=SL*α

[0070] Where γ is the estimated loading quantity, SL is the cross-sectional area data of the cargo box, and α is the stacking coefficient of the loaded goods.

[0071] It should be noted that, due to differences in the internal viscosity and friction of different materials, the same amount of material placed in the feed cylinder will result in different stacked cone heights. To ensure consistent stacking heights after feeding, a coefficient must be multiplied from the standard feeding amount for different materials to guarantee consistent stacking heights. This coefficient is called the stacking coefficient in the model. The stacking coefficient can be confirmed experimentally. For example, taking the stacking coefficient of sand as a baseline of 1, the feeding amount required for other materials to stack to the same height divided by the feeding amount of sand gives the corresponding stacking coefficient.

[0072] S400: Collect the maximum height data of the goods after the first-stage loading is completed by the TOF_1 camera 101, and obtain the remaining cargo box size data by the TOF_2 camera (102);

[0073] The maximum height data is the maximum stacking height of goods after they are currently placed in the material cylinder 100. The remaining cargo box size data includes the visible remaining cargo box size and the total cargo box size.

[0074] S401. Analyze and process the maximum height data of the goods, generate loading compensation data, and perform loading compensation.

[0075] The logic for generating the material feeding compensation data is as follows:

[0076] The compensation coefficients for different cargo materials are obtained through network connection. The compensation coefficients obtained during current loading are retrieved. The maximum height data of the cargo is subtracted from the cargo box depth data obtained by the TOF_1 camera 101. The calculation result is multiplied by the current cargo compensation coefficient to generate loading quantity compensation data. The analysis formula for generating the loading quantity compensation data is as follows:

[0077] Bs l=(Hc-hw)*β;

[0078] Where Bs l is the material loading compensation data, Hc is the cargo box depth data, hw is the maximum cargo height data, and β is the compensation coefficient obtained from the current material loading.

[0079] It should be noted that the compensation coefficient is similar to the stacking coefficient. Due to the different materials, the amount of material required to compensate for the same height will also be different. Taking the amount of material required to compensate for 1 meter with sand as the benchmark, the amount of material required to compensate for 1 meter with other materials divided by the amount of sand is the corresponding compensation coefficient.

[0080] S402. Analyze and process the remaining cargo box size data, generate data for adjusting the feeding amount, and adjust the feeding amount accordingly.

[0081] The logic for generating the controlled material feeding data is as follows:

[0082] The remaining cargo box size data is integrated and analyzed. The estimated material quantity (based on the sum of the quotient and above) is calculated by dividing the visible remaining cargo box size by the total cargo box size and adding the material quantity compensation data. The specific analysis formula is as follows:

[0083]

[0084] Where δ is the controlled feeding amount data, Bs l is the feeding amount compensation data, γ is the estimated feeding amount, Chs is the visible remaining cargo box size, and Chz is the total cargo box size.

[0085] S403. Perform secondary analysis and processing on the remaining cargo box size data, generate step distance control data and perform step distance control until the remaining cargo box size data in the cargo box is 0, and output the "Completed loading" label.

[0086] The logic for generating the step distance control data is as follows:

[0087] A secondary analysis is performed on the remaining cargo box size data. The stepping coefficients for different cargo materials are obtained via network connection. The quotient of the current cargo material's stepping coefficient multiplied by the visible remaining cargo box size divided by the total cargo box size is calculated. This result is then multiplied by the estimated loading quantity to obtain the product. The specific analysis formula is as follows:

[0088]

[0089] Where Bt l is the step distance control data, Jbx is the step coefficient of the current cargo material, Chs is the visible remaining cargo box size, Chz is the total cargo box size, and γ is the estimated loading amount.

[0090] It should be noted that due to different materials, the space occupied by the bottom of the stack of goods of the same height will be different, and the stepping distance required for the next loading vehicle will also be different. Taking a sand pile as an example, under the standard loading condition, the stepping coefficient is 1. The ratio of the stepping distance of other materials to that of sand material is the stepping coefficient.

[0091] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0092] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0093] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0094] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0095] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0096] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0097] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0098] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An automatic control method for truck loading process based on a TOF camera, characterized in that, The method includes a TOF_1 camera (101) and a TOF_2 camera (102) disposed on the side wall of the material cylinder (100), and includes the following steps: S100, Collect cargo box position data and loading point position data through TOF_1 camera (101); S101. Perform data analysis on the cargo box location data and the loading point location data, obtain the orientation difference, and generate loading point docking offset distance data and loading point docking offset angle data. S102. Take the data of the stopping offset distance and the stopping offset angle of the loading point as input, and output the control result through the position control model. S200: Collects parking stationary time data through TOF_1 camera (101); S201. Determine whether a vehicle is stationary based on the stationary duration data and generate a determination result. S300: The cargo box space dimension data is collected by a TOF_1 camera (101), and the cargo box cross-sectional data is obtained by preprocessing. S301. Analyze the cross-sectional data of the cargo box to generate a feeding quantity estimate, and begin primary feeding; the steps for generating the feeding quantity estimate are as follows: By connecting to the network, the stacking coefficients of different cargo materials are obtained, the stacking coefficient of the currently loaded cargo is retrieved, and the cross-sectional area data of the cargo box is multiplied with the stacking coefficient of the currently loaded cargo. S400: The maximum height data of the goods after the first-stage loading is completed is collected by the TOF_1 camera (101), and the remaining cargo box size data is obtained by the TOF_2 camera (102). S401. Analyze and process the maximum height data of the goods to generate loading amount compensation data and perform loading compensation; wherein, the generation logic of the loading amount compensation data is as follows: The compensation coefficients for different cargo materials are obtained through the network, the compensation coefficients obtained for the current loading are retrieved, the cargo box depth data obtained by the TOF_1 camera (101) are subtracted from the maximum height data of the cargo, and the calculation result is multiplied with the compensation coefficient of the current cargo to generate the loading amount compensation data. S402. Analyze and process the remaining cargo box size data, generate data for adjusting the feeding amount, and adjust the feeding amount accordingly. S403. Perform secondary analysis and processing on the remaining cargo box size data to generate step distance control data and perform step distance control until the remaining cargo box size data in the cargo box is 0, and output a "loading complete" flag; wherein, the logic for generating the step distance control data is as follows: A secondary analysis is performed on the remaining cargo box size data. The stepping coefficients of different cargo materials are obtained through the network. The stepping coefficient of the current cargo material is multiplied by the quotient of the visible remaining cargo box size divided by the total cargo box size. The result is then multiplied by the estimated loading quantity to obtain the product.

2. The automatic control method for truck loading process based on a TOF camera according to claim 1, characterized in that, The cargo box position data is the cargo box position collected in real time by a TOF_1 camera (101) with the loading point docking line as the reference target. The loading point position is the loading point docking line data collected in real time by a TOF_1 camera (101). The loading point docking offset angle data is the angle difference between the cargo box position and the loading point docking line as the reference target. The loading point docking offset distance data is the maximum distance between the cargo box position and the loading point docking line as the reference target.

3. The automatic control method for truck loading process based on a TOF camera according to claim 2, characterized in that, The construction steps of the position control model are as follows: Step Q1: Obtain n sets of historical material point stopping offset distance data and historical material point stopping offset angle data as test classes, and collect the corresponding historical control distance and historical control angle of n test classes as control results; Step Q2: Use n test classes as input attribute values ​​for the position control model, use the control results as the output of the position control model, use each control result as a prediction target, and use the prediction accuracy of the position control model as the training target. Step Q3: Train the machine learning models for n test classes. Based on the prediction accuracy threshold L set in advance before the training step, stop training when the prediction accuracy reaches the prediction accuracy threshold kio; and mark the machine learning model that has been trained for the nth test class as kio. The position control model is a deep neural network model, decision tree or support vector machine. Step Q4: Send the trained position control model kio to the processor and store it.

4. The automatic control method for truck loading process based on a TOF camera according to claim 3, characterized in that, The logic for determining vehicle stillness based on stationary duration data is as follows: The parking still time Gx is collected by a TOF_1 camera (101). The still time comparison thresholds Tz1 and Tz2 are set. The time comparison threshold Tz1 is less than the time comparison threshold Tz2. If the parking still time Gx is greater than 0 and the parking still time Gx is less than Tz1, a "Stationary adjustment in progress" judgment label is generated; if the parking still time Gx is greater than Tz1 and the parking still time Gx is less than Tz2, a "Adjustment is about to be completed" judgment label is generated; if the parking still time Gx is greater than Tz2, a "Adjustment completed" judgment label is generated.

5. The automatic control method for truck loading process based on a TOF camera according to claim 4, characterized in that, The spatial dimension data includes cargo box depth data and cargo box width data. The preprocessing performs a product operation on the cargo box depth and cargo box width to generate cargo box cross-sectional data.

6. The automatic control method for truck loading process based on a TOF camera according to claim 5, characterized in that, The maximum height data is the maximum stacking height of the goods after they are placed in the material cylinder (100). The remaining cargo box size data includes the visible remaining cargo box size and the total cargo box size.

7. The automatic control method for truck loading process based on a TOF camera according to claim 6, characterized in that... The logic for generating the controlled feeding data is as follows: The remaining cargo box size data is integrated and analyzed, and the estimated amount of material above the commercial vehicle is obtained by dividing the visible remaining cargo box size by the total cargo box size, and the material compensation data is added together.

Images