Container loading and unloading detection control method and system based on artificial intelligence
Through the container loading and unloading detection and control method based on artificial intelligence, using data fusion and neural network model, the problems of low efficiency and poor accuracy of traditional detection methods are solved, and the accurate measurement and automated decision-making of container weight are realized, which improves transportation safety and efficiency.
Patent Information
- Application Number
- CN202411009562.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional container inspection methods rely on manual inspection, are inefficient and susceptible to human factors, resulting in mis-checking or missed inspection. The accumulated error of existing weight sensors during the measurement process and interference from environmental factors lead to a decrease in measurement accuracy.
Adopt container loading and unloading detection and control methods based on artificial intelligence, and use weight sensors and temperature and humidity sensors to collect data for data fusion and correction processing, train neural network models to predict container weight, and combine environmental compensation algorithms to reduce errors and environmental interference to achieve accurate Measure and automate decision making.
It improves the accuracy and work efficiency of container weight measurement, reduces the need for manual intervention, enhances the stability and reliability of the system, ensures the safety and smoothness of container transportation, and improves the degree of automation of operations.
Smart Images

Figure CN119004012A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of container detection, and in particular to an artificial intelligence-based container loading and unloading detection control method and system. Background Art
[0002] With the continuous growth of trade, container transportation, as the main mode of logistics, has become increasingly important. The efficiency, convenience and standardization of container transportation make it an important link connecting trade. However, with the increase in transportation volume, the inspection and management of containers are also facing huge challenges. Traditional container inspection methods mainly rely on manual inspection, which is not only inefficient but also easily affected by human factors, resulting in false detection or missed detection. In recent years, artificial intelligence technology has made significant progress, and the maturity of artificial intelligence related technology has provided new solutions for container inspection.
[0003] In the prior art, for railway containers, during the weighing and lifting process, multiple weight sensors are used to detect the weight of the container to determine the weight threshold of the railway container, and then realize the lifting operation of the railway container. However, multiple weight sensors will have problems of error accumulation and interference from environmental factors during the measurement process, which will affect the measurement accuracy and thus lead to a decrease in measurement accuracy. For this reason, a container loading and unloading detection control method and system based on artificial intelligence are proposed. Summary of the invention
[0004] The purpose of the present invention is to provide a container loading and unloading detection control method and system based on artificial intelligence to solve the problems raised in the above background technology.
[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:
[0006] In the first aspect, a container loading and unloading detection control method based on artificial intelligence comprises the following steps:
[0007] Step 1: Deploy weight sensors and temperature and humidity sensors to collect raw data on railway container weight and environmental parameters, and set container weight thresholds based on transportation requirements and safety standards to determine whether the container needs to be lifted;
[0008] Step 2, based on the acquired original data of railway container weight and environmental parameters, data fusion and correction processing are performed on the original data, wherein the original data includes historical data and real-time data;
[0009] Step 3: Train a neural network model based on historical data to predict the weight of the container, match it with the set container weight threshold, and determine the lifting conditions;
[0010] Step 4: Combine the real-time data of environmental parameters and apply environmental compensation algorithms to adjust sensor readings to reduce interference from environmental factors, input the fused and compensated data into the trained neural network model, and output the predicted container weight;
[0011] Step 5: Evaluate the weight of the container and the preset container weight threshold, make a decision to lift or not lift, and continuously monitor the operating status of the sensor and the changing trend of the measurement data to issue an early warning of abnormal situations.
[0012] A further improvement of the technical solution of the present invention is that in step 1, the collection process of the original data of the railway container weight and environmental parameters is as follows:
[0013] Step 101, deploying four weight sensors in the railway container weighing area to ensure uniform distribution and accurate measurement of the weight of the container, and deploying temperature and humidity sensors in the weighing area to monitor the environmental conditions of weight measurement and collect environmental parameter data;
[0014] Step 102, setting up a data acquisition system to collect readings of the weight sensor and the temperature and humidity sensor, obtaining raw data, ensuring the timeliness and accuracy of the data, and cleaning, denoising and outlier processing of the raw data;
[0015] Step 103, according to the transportation requirements and safety standards, a weight threshold of the container is set. If the weight of the container exceeds the threshold, it is determined that it needs to be lifted; otherwise, it is determined that it does not need to be lifted.
[0016] A further improvement of the technical solution of the present invention is that in step 2, the data fusion and correction processing process of the original data is:
[0017] Step 201, dividing the acquired original data of railway container weight and environmental parameters into historical data and real-time data, and extracting the historical data of the weight sensor and the historical data of the temperature and humidity sensor to obtain a historical data set and a real-time data set;
[0018] Step 202: weighted average is adopted to assign different weights according to the accuracy and stability of each sensor, and the raw data of the four weight sensors are fused to reduce the error accumulation and reflect the actual weight of the container. The purpose of fusion is to reduce the overall error accumulation caused by the error of a single sensor. The fused data is considered to more accurately reflect the actual weight of the container.
[0019] Step 203, analyzing the historical data of the temperature and humidity sensor, determining the relationship between the environmental parameters and the weight measurement, and performing environmental parameter correction on the fused weight data according to the change of the environmental parameters. The purpose of the correction is to adjust the reading of the weight sensor to make it closer to the actual weight of the container under standard environmental conditions;
[0020] Step 204, based on the degree of influence of environmental parameters on weight measurement, an environmental compensation model is constructed in combination with historical data of the temperature and humidity sensors, and a temperature compensation index and a humidity compensation index are obtained respectively. The temperature compensation index and the humidity compensation index are combined to obtain a comprehensive environmental compensation index, and the influence of comprehensive temperature and humidity on the measurement results is adjusted to adjust the reading of the weight sensor to eliminate or reduce the influence of environmental factors on the measurement results.
[0021] A further improvement of the technical solution of the present invention is that the expression of the data fusion is:
[0022]
[0023] Among them, W i is the weighted average weight, which means the average value of the weight sensor readings considering different weights, w i is the weight of the i-th weight sensor, which is used to reflect its contribution or importance to the total weight. i is the reading of the i-th weight sensor, i ranges from 1 to 4;
[0024] The expression of the temperature compensation index is:
[0025] W T =W raw ·(1+α·ΔT);
[0026] Among them, W T is the temperature compensation index, W raw is the original weight reading, α is the temperature coefficient, which indicates the material's sensitivity to temperature, and ΔT is the difference between the current temperature and the reference temperature;
[0027] The expression of the humidity compensation index is:
[0028] W H =W T ·(1+β·ΔRH);
[0029] Among them, W H is the humidity compensation index, W T is the temperature compensation index, which indicates the weight after temperature compensation, β is the humidity system, which indicates the sensitivity of the material to humidity, and ΔRH is the difference between the current relative humidity and the reference humidity;
[0030] The expression of the comprehensive environmental compensation index is:
[0031] W EZ =W raw ·(1+α·ΔT+β·ΔRH+γ·ΔT·ΔRH);
[0032] Among them, WEZ is the comprehensive environmental compensation index, and γ is the cross-term coefficient, which represents the combined effect of temperature and humidity.
[0033] A further improvement of the technical solution of the present invention is that in step 3, the prediction process of the container weight is:
[0034] Step 301, extracting the weight sensor reading and related temperature and humidity data from the historical data set, obtaining a feature data set, and dividing the feature data set into a training set and a test set;
[0035] Step 302: training a container weight prediction model based on the neural network model and the data of the training set, adjusting the parameters of the container weight prediction model through an iterative optimization algorithm so that the container weight prediction model can accurately predict the weight of the container, and using a test data set to evaluate the container weight prediction model, combining the Huber loss function to evaluate the prediction performance of the container weight prediction model, calculating the gap between the model prediction value and the actual observation value, and measuring the accuracy of the model prediction;
[0036] Step 303, deploy the container weight prediction model, input real-time data, obtain the predicted weight of the container, and compare it with the set container weight threshold. If the predicted weight exceeds the threshold, the lifting condition is triggered, and the model prediction results and actual lifting conditions are collected.
[0037] A further improvement of the technical solution of the present invention is that the output calculation formula of each neuron in the container weight prediction model is:
[0038]
[0039] Among them, a j is the output of the jth neuron, σ is the Sigmoid activation function, w jk is the weight from the kth input to the jth neuron, x k is the kth input feature, b j is the bias term of the jth neuron, and m is the number of input features;
[0040] The expression of the Huber loss function is:
[0041]
[0042] Among them, HL is the Huber loss value of the entire data set, which is the sum of all sample losses, and ρ is the function used to calculate the loss of a single sample in the Huber loss function. is the residual of a single sample, i.e. the model prediction value The actual observed value y q, δ is the threshold parameter in the Huber loss function, which is used to distinguish between square loss and linear loss to balance the sensitivity of square loss and the robustness of linear loss. q is the sample index, indicating the qth sample in the data set, and n is the total number of samples in the data set.
[0043] A further improvement of the technical solution of the present invention is that in step 4, the predicted container weight output process is:
[0044] Step 401, extracting the real-time readings of the weight sensor and the temperature and humidity sensor, fusing the readings of the four weight sensors into a single weight value, and using an environmental compensation algorithm to adjust the reading of the weight sensor according to the real-time temperature and humidity data to reduce the interference of environmental factors on weight measurement;
[0045] Step 402, inputting the data of the real-time weight sensor and the temperature and humidity sensor into the trained neural network model, and predicting the weight of the container based on the neural network model and the compensated data;
[0046] Step 403: output the predicted container weight, and determine whether the lifting conditions are met according to the set threshold.
[0047] A further improvement of the technical solution of the present invention is that in step 5, the monitoring process of the sensor operating state is:
[0048] Step 501, compare the compensated weight data with the preset weight threshold. If the weight of the container exceeds the threshold, the lifting mechanism is triggered. If the weight of the container is lower than or equal to the threshold, the lifting is not performed. A heavy button and an empty button are set to switch between different states to decide whether to perform the lifting or release operation automatically or manually.
[0049] Step 502, continuously monitor the operating status of the weight sensor and the temperature and humidity sensor to ensure that they work normally and transmit data in real time, and regularly collect the measurement data of the sensors, analyze the change trend of the sensor measurement data, identify abnormal conditions, and set alarm conditions to detect abnormal conditions of the sensor data, such as sudden changes, continuous deviations, etc.;
[0050] Step 503, when an abnormal situation is detected, a warning signal is immediately issued, and corresponding measures are taken, such as checking the sensor, adjusting the compensation algorithm, recalibrating, etc.
[0051] In the second aspect, an artificial intelligence-based container loading and unloading detection control system is used to implement an artificial intelligence container detection control method, including a container detection center, wherein the container detection center is communicatively connected with a sensor module, a data acquisition and processing module, a data fusion module, an environmental compensation module, a prediction module, a decision execution module, a state switching control module, and an abnormality detection module, wherein electrical signals are connected between the modules;
[0052] The sensor module includes four weight sensors and temperature and humidity sensors. The weight sensor is used to detect the weight of the railway container. The four sensors are distributed in different positions of the container weighing area to ensure the accuracy and stability of weight measurement. The total weight of the container is calculated through the readings of the four sensors. The temperature and humidity sensor is used to collect temperature and humidity data. Combined with the environmental compensation algorithm, it reduces the interference of environmental factors on weight measurement and improves the accuracy of measurement.
[0053] The data acquisition and processing module is used to collect data from the sensor module and perform data processing, including weight and temperature and humidity information, and to adjust the reading of the weight sensor using an environmental compensation algorithm to reduce the impact of environmental factors on the measurement, and to fuse the readings of the four weight sensors into a single weight value to improve the reliability of the measurement;
[0054] The data fusion module is used to integrate the data of the four weight sensors, reduce the impact of single sensor errors and failures, and improve the accuracy and robustness of weight measurement;
[0055] The environmental compensation module is used to adjust the weight reading according to the temperature and humidity data in combination with the application of the environmental compensation algorithm, eliminate or reduce the interference of environmental factors on the weight measurement, and improve the measurement accuracy;
[0056] The prediction module is used to construct a neural network model to predict the container weight and evaluate the fused weight value according to a preset container weight threshold;
[0057] The decision execution module is used to evaluate the fused weight value according to a preset container weight threshold, and if it is detected that the railway container weight is higher than the threshold, a lifting instruction is issued to control the lifting operation; if it is detected that the railway container weight is not higher than the threshold, a non-lifting instruction is issued;
[0058] The state switching control module allows the operator to switch the system state and perform calibration through the "one heavy and one empty" buttons, providing flexible control of the system state and facilitating maintenance and calibration;
[0059] The anomaly detection module is used to continuously monitor the operating status of the sensor and the changing trend of the measurement data, monitor the sensor data and prediction results, detect abnormal situations and trigger early warnings, promptly discover and respond to potential problems, and ensure stable operation of the system.
[0060] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:
[0061] 1. The present invention provides an artificial intelligence-based container loading and unloading detection control method and system, which realizes accurate measurement of container weight by integrating sensors and data processing technology. The use of four weight sensors ensures the stability and accuracy of the measurement, and the environmental compensation algorithm further reduces the interference of environmental factors on the measurement, which not only improves the measurement accuracy of container weight, but also greatly shortens the detection time, improves the overall work efficiency, autonomously collects data, processes data, makes decisions and executes corresponding control operations, and reduces the need for manual intervention.
[0062] 2. The present invention provides an artificial intelligence-based container loading and unloading detection control method and system. By continuously monitoring the operating status of the sensor and the changing trend of the measurement data, abnormal situations can be discovered and handled in a timely manner. The real-time monitoring and early warning mechanism greatly enhances the stability and reliability of the system, ensuring the safety and smoothness of container transportation. Through training with historical data, the weight of the container is predicted, reducing dependence on manual intervention, and a decision on whether to lift is made based on a preset threshold, further improving the degree of automation of the operation.
[0063] 3. The present invention provides an artificial intelligence-based container loading and unloading detection control method and system, which optimizes the lifting operation of the container through precise measurement and intelligent decision-making, quickly responds to the weight information of the container, reduces delays caused by waiting for weight measurement results, and the automated decision-making process reduces decision-making time and improves operational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0064] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0065] Figure 1 is a flow chart of the method of the present invention;
[0066] Figure 2 The data fusion and correction process flow chart of the raw data of the present invention;
[0067] Figure 3 is a flow chart for predicting container weight of the present invention;
[0068] Figure 4 It is a module structure diagram of the present invention. DETAILED DESCRIPTION
[0069] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0070] Embodiment 1, as Figure 1-3 As shown, the present invention provides a container loading and unloading detection control method based on artificial intelligence, comprising the following steps:
[0071] Step 1: Deploy weight sensors and temperature and humidity sensors to collect the original data of railway container weight and environmental parameters, and set the container weight threshold according to transportation requirements and safety standards to determine whether the container needs to be lifted. Deploy four weight sensors in the railway container weighing area to ensure uniform distribution and accurate measurement of the container weight, and deploy temperature and humidity sensors in the weighing area to monitor the environmental conditions for weight measurement and collect environmental parameter data. Set up a data acquisition system to collect the readings of weight sensors and temperature and humidity sensors, obtain original data, ensure the timeliness and accuracy of the data, and clean, denoise and process the original data to improve the efficiency and accuracy of subsequent data processing. Set the container weight threshold according to transportation requirements and safety standards. If the weight of the container exceeds the threshold, it is judged to be lifted; otherwise, it is judged not to be lifted;
[0072] Step 2, based on the obtained original data of railway container weight and environmental parameters, the original data are fused and corrected, wherein the original data includes historical data and real-time data. The obtained original data of railway container weight and environmental parameters are divided into historical data and real-time data, and the historical data of weight sensor and the historical data of temperature and humidity sensor are extracted to obtain historical data set and real-time data set. The weighted average is adopted to assign different weights according to the accuracy and stability of each sensor, and the original data of four weight sensors are fused to reduce error accumulation and reflect the actual weight of the container. The purpose of fusion is to reduce the overall error accumulation caused by the error of a single sensor. The fused data is regarded as more accurate. It accurately reflects the actual weight of the container, analyzes the historical data of the temperature and humidity sensors, determines the relationship between environmental parameters and weight measurement, and performs environmental parameter correction on the fused weight data according to the changes in environmental parameters. The purpose of the correction is to adjust the reading of the weight sensor to make it closer to the actual weight of the container under standard environmental conditions. According to the degree of influence of environmental parameters on weight measurement, an environmental compensation model is constructed in combination with the historical data of the temperature and humidity sensors to obtain the temperature compensation index and humidity compensation index respectively, and the comprehensive environmental compensation index is obtained by combining the temperature compensation index and the humidity compensation index. The influence of temperature and humidity on the measurement results is comprehensively considered, and the reading of the weight sensor is adjusted to eliminate or reduce the influence of environmental factors on the measurement results.
[0073] Furthermore, the expression of data fusion is:
[0074]
[0075] Among them, W i is the weighted average weight, which means the average value of the weight sensor readings considering different weights, w i is the weight of the i-th weight sensor, which is used to reflect its contribution or importance to the total weight. i is the reading of the i-th weight sensor, i ranges from 1 to 4;
[0076] The expression of temperature compensation index is:
[0077] W T =W raw ·(1+α·ΔT);
[0078] Among them, W T is the temperature compensation index, W raw is the original weight reading, α is the temperature coefficient, which indicates the material's sensitivity to temperature, and ΔT is the difference between the current temperature and the reference temperature;
[0079] The expression of humidity compensation index is:
[0080] W H=W T ·(1+β·ΔRH);
[0081] Among them, W H is the humidity compensation index, W T is the temperature compensation index, which indicates the weight after temperature compensation, β is the humidity system, which indicates the sensitivity of the material to humidity, and ΔRH is the difference between the current relative humidity and the reference humidity;
[0082] The expression of comprehensive environmental compensation index is:
[0083] W EZ =W raw ·(1+α·ΔT+β·ΔRH+γ·ΔT·ΔRH);
[0084] Among them, W EZ is the comprehensive environmental compensation index, γ is the cross-term coefficient, which represents the combined effect of temperature and humidity;
[0085] Step 3: Train a neural network model based on historical data to predict the weight of the container and match the set container weight threshold to determine the lifting conditions. Extract the weight sensor readings and related temperature and humidity data from the historical data set to obtain a feature data set. Divide the feature data set into a training set and a test set. Train the container weight prediction model based on the neural network model combined with the data of the training set. Adjust the parameters of the container weight prediction model through an iterative optimization algorithm so that the container weight prediction model can accurately predict the weight of the container. Use the test data set to evaluate the container weight prediction model. Combine the Huber loss function to evaluate the prediction performance of the container weight prediction model. Calculate the gap between the model prediction value and the actual observation value to measure the accuracy of the model prediction. Deploy the container weight prediction model and input real-time data to obtain the predicted weight of the container. Compare it with the set container weight threshold. If the predicted weight exceeds the threshold, the lifting condition is triggered, and the model prediction results and actual lifting conditions are collected.
[0086] Furthermore, the output calculation formula of each neuron in the container weight prediction model is:
[0087]
[0088] Among them, a j is the output of the jth neuron, σ is the Sigmoid activation function, w jk is the weight from the kth input to the jth neuron, x k is the kth input feature, b j is the bias term of the jth neuron, and m is the number of input features;
[0089] Furthermore, the expression of Huber loss function is:
[0090]
[0091] Among them, HL is the Huber loss value of the entire data set, which is the sum of all sample losses, and ρ is the function used to calculate the loss of a single sample in the Huber loss function. is the residual of a single sample, i.e. the model prediction value The actual observed value y q The difference between , δ is the threshold parameter in the Huber loss function, which is used to distinguish between the square loss and the linear loss to balance the sensitivity of the square loss and the robustness of the linear loss. q is the sample index, which means the qth sample in the data set. n is the total number of samples in the data set. It should be noted that when hour, for That is, the mean square error, which is sensitive to small errors and helps the model learn small fluctuations in the data. for That is, linear loss, which reduces the sensitivity to large errors and helps reduce the impact of outliers on the model. For small errors, Huber loss is the same as mean square error, and the loss decreases rapidly as the error decreases. For large errors, Huber loss is transformed into a linear form, which reduces the sensitivity to outliers and reduces the impact of individual large errors on the overall loss. When the error value is close to δ, Huber loss smoothly transitions from square loss to linear loss, avoiding the discontinuity of the loss function.
[0092] Step 4: Combine the real-time data of environmental parameters and apply the environmental compensation algorithm to adjust the sensor readings, reduce the interference of environmental factors, input the fused and compensated data into the trained neural network model, output the predicted container weight, extract the real-time readings of the weight sensor and the temperature and humidity sensor, fuse the readings of the four weight sensors into a single weight value, and use the environmental compensation algorithm to adjust the readings of the weight sensor according to the real-time temperature and humidity data to reduce the interference of environmental factors on weight measurement. Input the data of the real-time weight sensor and the temperature and humidity sensor into the trained neural network model, predict the container weight according to the neural network model and the compensated data, output the predicted container weight, and determine whether the lifting conditions are met according to the set threshold;
[0093] Step 5: Evaluate the weight of the container and the preset container weight threshold, make a decision on whether to lift or not, and continuously monitor the operating status of the sensor and the changing trend of the measurement data, issue an early warning of abnormal conditions, compare the compensated weight data with the preset weight threshold, and trigger the lifting mechanism if the weight of the container exceeds the threshold. If the weight of the container is lower than or equal to the threshold, the lifting will not be performed. Set a heavy button and an empty button to switch between different states, and decide whether to automatically or manually perform the lifting or release operation. Continuously monitor the operating status of the weight sensor and the temperature and humidity sensor to ensure that they work normally and transmit data in real time, and regularly collect the measurement data of the sensors, analyze the changing trend of the sensor measurement data, identify abnormal conditions, set alarm conditions to detect abnormal conditions in the sensor data, such as sudden changes, continuous deviations, etc. When an abnormal condition is detected, an early warning signal is immediately issued, and corresponding measures are taken, such as checking the sensor, adjusting the compensation algorithm, and recalibrating.
[0094] Embodiment 2, as Figure 4 As shown, on the basis of Example 1, the present invention further provides an artificial intelligence-based container loading and unloading detection control system for realizing an artificial intelligence container detection control method, including a container detection center, the container detection center is communicatively connected with a sensor module, a data acquisition and processing module, a data fusion module, an environmental compensation module, a prediction module, a decision execution module, a state switching control module and an abnormality detection module, wherein the modules are connected by electrical signals;
[0095] The sensor module includes four weight sensors and temperature and humidity sensors. The weight sensor is used to detect the weight of the railway container. The four sensors are distributed in different positions of the container weighing area to ensure the accuracy and stability of weight measurement. The total weight of the container is calculated through the readings of the four sensors. The temperature and humidity sensor is used to collect temperature and humidity data. Combined with the environmental compensation algorithm, it reduces the interference of environmental factors on weight measurement and improves the accuracy of measurement;
[0096] The data acquisition and processing module is used to collect data from the sensor module and process the data, including weight, temperature and humidity information, and apply the environmental compensation algorithm to adjust the reading of the weight sensor to reduce the impact of environmental factors on the measurement, and fuse the readings of the four weight sensors into a single weight value to improve the reliability of the measurement;
[0097] Data fusion module, used to integrate the data of four weight sensors, reduce the impact of single sensor errors and failures, and improve the accuracy and robustness of weight measurement;
[0098] Environmental compensation module, used to combine and apply environmental compensation algorithm, adjust weight reading according to temperature and humidity data, eliminate or reduce the interference of environmental factors on weight measurement, and improve measurement accuracy;
[0099] The prediction module is used to build a neural network model to predict the container weight and evaluate the fused weight value according to the preset container weight threshold;
[0100] The decision-making execution module is used to evaluate the fused weight value according to the preset container weight threshold. If it is detected that the railway container weight is higher than the threshold, a lifting instruction is issued to control the lifting operation. If it is detected that the railway container weight is not higher than the threshold, a non-lifting instruction is issued;
[0101] The state switching control module allows the operator to switch the system state and perform calibration through the "one heavy and one empty" button, providing flexible control of the system state and facilitating maintenance and calibration;
[0102] The anomaly detection module is used to continuously monitor the operating status of the sensor and the changing trend of the measurement data, monitor the sensor data and prediction results, detect abnormal situations and trigger early warnings, promptly discover and respond to potential problems, and ensure stable operation of the system.
[0103] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. Container loading and unloading detection and control method based on artificial intelligence, Features: The following steps are involved: Step 1: Deploy weight sensors and temperature and humidity sensors to collect raw data on railway container weight and environmental parameters, and set container weight thresholds based on transportation requirements and safety standards to determine whether the container needs to be lifted; Step 2, based on the acquired original data of railway container weight and environmental parameters, data fusion and correction processing are performed on the original data, wherein the original data includes historical data and real-time data; Step 3: Train a neural network model based on historical data to predict the weight of the container, match it with the set container weight threshold, and determine the lifting conditions; Step 4: Combine the real-time data of environmental parameters and apply environmental compensation algorithms to adjust sensor readings to reduce interference from environmental factors, input the fused and compensated data into the trained neural network model, and output the predicted container weight; Step 5: Evaluate the weight of the container and the preset container weight threshold, make a decision to lift or not lift, and continuously monitor the operating status of the sensor and the changing trend of the measurement data to issue an early warning of abnormal situations.
2. According to the artificial intelligence-based container loading and unloading detection control method of claim 1, Features: In step 1, the process of collecting the original data of railway container weight and environmental parameters is as follows: Step 101, deploying four weight sensors in the railway container weighing area, and deploying temperature and humidity sensors in the weighing area to monitor the environmental conditions of weight measurement and collect environmental parameter data; Step 102, setting a data acquisition system to collect readings of the weight sensor and the temperature and humidity sensor, obtaining raw data, and cleaning, denoising and outlier processing the raw data; Step 103, according to the transportation requirements and safety standards, a weight threshold of the container is set. If the weight of the container exceeds the threshold, it is determined that it needs to be lifted; otherwise, it is determined that it does not need to be lifted.
3. According to the artificial intelligence-based container loading and unloading detection control method of claim 2, Features: In step 2, the data fusion and correction process of the original data is as follows: Step 201, dividing the acquired original data of railway container weight and environmental parameters into historical data and real-time data, and extracting the historical data of the weight sensor and the historical data of the temperature and humidity sensor to obtain a historical data set and a real-time data set; Step 202: weighted average is used to assign different weights to each sensor according to its accuracy and stability, and the raw data of the four weight sensors are fused to reduce error accumulation and reflect the actual weight of the container. Step 203, analyzing the historical data of the temperature and humidity sensor, determining the relationship between the environmental parameters and the weight measurement, and performing environmental parameter correction on the fused weight data according to the change of the environmental parameters; Step 204, based on the influence of environmental parameters on weight measurement, an environmental compensation model is constructed in combination with historical data of the temperature and humidity sensors to obtain a temperature compensation index and a humidity compensation index respectively, and a comprehensive environmental compensation index is obtained by combining the temperature compensation index and the humidity compensation index to comprehensively consider the influence of temperature and humidity on the measurement results.
4. According to the artificial intelligence-based container loading and unloading detection control method of claim 3, Features: The expression of the data fusion is: Among them, W i is the weighted average weight, w i is the weight of the i-th weight sensor, We i is the reading of the i-th weight sensor, i ranges from 1 to 4; The expression of the temperature compensation index is: W T =W raw ·(1+α·ΔT); Among them, W T is the temperature compensation index, W raw is the original weight reading, α is the temperature coefficient, which indicates the material's sensitivity to temperature, and ΔT is the difference between the current temperature and the reference temperature; The expression of the humidity compensation index is: W H =W T ·(1+β·ΔRH); Among them, W H is the humidity compensation index, W T is the temperature compensation index, β is the humidity system, which indicates the sensitivity of the material to humidity, and ΔRH is the difference between the current relative humidity and the reference humidity; The expression of the comprehensive environmental compensation index is: W EZ =W raw ·(1+α·ΔT+β·ΔRH+γ·ΔT·ΔRH); Among them, W EZ is the comprehensive environmental compensation index, and γ is the cross-term coefficient, which represents the combined effect of temperature and humidity.
5. According to the artificial intelligence-based container loading and unloading detection control method of claim 4, Features: In step 3, the container weight prediction process is: Step 301, extracting the weight sensor reading and related temperature and humidity data from the historical data set, obtaining a feature data set, and dividing the feature data set into a training set and a test set; Step 302: training a container weight prediction model based on the neural network model and the data of the training set, adjusting the parameters of the container weight prediction model through an iterative optimization algorithm, and evaluating the container weight prediction model using a test data set, and evaluating the prediction performance of the container weight prediction model in combination with the Huber loss function; Step 303, deploy the container weight prediction model, input real-time data, obtain the predicted weight of the container, and compare it with the set container weight threshold. If the predicted weight exceeds the threshold, the lifting condition is triggered, and the model prediction results and actual lifting conditions are collected.
6. The container loading and unloading detection and control method based on artificial intelligence according to claim 5, Features: The expression of the Huber loss function is: Among them, HL is the Huber loss value of the entire data set, which is the sum of all sample losses, and ρ is the function used to calculate the loss of a single sample in the Huber loss function. is the residual of a single sample, i.e. the model prediction value The actual observed value y q , δ is the threshold parameter in the Huber loss function, which is used to distinguish between square loss and linear loss, q is the sample index, indicating the qth sample in the data set, and n is the total number of samples in the data set.
7. The container loading and unloading detection and control method based on artificial intelligence according to claim 6, Features: In step 4, the predicted container weight output process is: Step 401, extracting the real-time readings of the weight sensor and the temperature and humidity sensor, fusing the readings of the four weight sensors into a single weight value, and using an environmental compensation algorithm to adjust the reading of the weight sensor according to the real-time temperature and humidity data; Step 402, inputting the data of the real-time weight sensor and the temperature and humidity sensor into the trained neural network model, and predicting the weight of the container based on the neural network model and the compensated data; Step 403: output the predicted container weight, and determine whether the lifting conditions are met according to the set threshold.
8. The container loading and unloading detection and control method based on artificial intelligence according to claim 7, Features: In step 5, the monitoring process of the sensor operating status is: Step 501, compare the compensated weight data with the preset weight threshold. If the weight of the container exceeds the threshold, the lifting mechanism is triggered. If the weight of the container is lower than or equal to the threshold, the lifting is not performed. A heavy button and an empty button are set to switch between different states to decide whether to perform the lifting or release operation automatically or manually. Step 502, continuously monitor the operating status of the weight sensor and the temperature and humidity sensor, and regularly collect the measurement data of the sensors, analyze the change trend of the sensor measurement data, identify abnormal conditions, and set alarm conditions to detect abnormal conditions of the sensor data; Step 503: When an abnormal situation is detected, an early warning signal is immediately issued and corresponding measures are taken.
9. A container loading and unloading detection control system based on artificial intelligence, used to implement the container loading and unloading detection control method based on artificial intelligence according to any one of claims 1 to 8, comprising a container detection center, Features: The container detection center is communicatively connected with a sensor module, a data acquisition and processing module, a data fusion module, an environmental compensation module, a prediction module, a decision execution module, a state switching control module and an abnormality detection module, wherein electrical signals are connected between the modules; The sensor module includes four weight sensors and a temperature and humidity sensor. The weight sensor is used to detect the weight of the railway container. The total weight of the container is calculated through the readings of the four sensors. The temperature and humidity sensor is used to collect temperature and humidity data. The data acquisition and processing module is used to collect data from the sensor module, perform data processing, including weight and temperature and humidity information, and fuse the readings of the four weight sensors into a single weight value; The data fusion module is used to integrate the data of four weight sensors; The environmental compensation module is used to adjust the weight reading according to the temperature and humidity data in combination with the application of the environmental compensation algorithm; The prediction module is used to construct a neural network model to predict the container weight and evaluate the fused weight value according to a preset container weight threshold; The decision execution module is used to evaluate the fused weight value according to a preset container weight threshold, and if it is detected that the railway container weight is higher than the threshold, a lifting instruction is issued to control the lifting operation; if it is detected that the railway container weight is not higher than the threshold, a non-lifting instruction is issued; The state switching control module allows the operator to switch the system state and perform calibration through the "one heavy and one empty" button; The anomaly detection module is used to continuously monitor the operating status of the sensor and the change trend of the measurement data, monitor the sensor data and prediction results, detect abnormal situations and trigger early warnings.
Citation Information
Cited By
Automatic weighing system based on enameled wire
CN119642952A
Automobile signal detection system
CN119806103A
Control method and system for loading and unloading detection of container reach stacker
CN119873626A
Intelligent package fresh-keeping control method and system
CN121008636A
Container reach stacker loading and unloading detection method and system based on artificial intelligence
CN121516743A
