Cargo status prediction method, device and multi-classification modeling method
By using a multi-classification model of vehicle-mounted weighing equipment to automatically predict the status of goods, the problem of poor timeliness and accuracy of goods status identification under manual identification methods is solved, and efficient and low-cost goods status identification is achieved.
Patent Information
- Application Number
- CN202210054273.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-11
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2040-09-11
AI Technical Summary
Existing methods for identifying cargo status based on manual identification are not timely or accurate, and are costly.
By acquiring continuous raw data points from vehicle-mounted weighing equipment, a multi-classification model is used to predict the status of goods, including the processing and feature generation of both whole and segmented data, and to establish a multi-classification model for automated prediction.
It significantly improves the accuracy and timeliness of cargo status identification, reduces costs, and requires no human intervention.
Smart Images

Figure CN114548508B_ABST
Abstract
Description
[0001] This application is a divisional application of the Chinese patent application with application number 202010956370.8, application date September 11, 2020, and invention name “Method, device and multi-classification modeling method for cargo status prediction”. Technical Field
[0002] The present application relates to the field of logistics technology, and in particular to a method, device and multi-classification modeling method for cargo status prediction. Background Art
[0003] In the field of logistics, especially in the field of truck freight or trunk road freight, it is very important to judge the status of goods. Accurate and timely identification of the status of goods can help managers improve vehicle management efficiency and reduce transportation costs.
[0004] At present, cargo status identification is mainly carried out manually, that is, the driver or terminal manager manually counts the actual status of the cargo and then reports it, but this method has information lag; if the terminal manager is required to manually report the cargo status in real time, it will greatly increase the labor cost; if the driver is required to report the cargo status in real time, the effect is not satisfactory, because the driver's job is to drive the vehicle, and he needs to rest after parking the vehicle, and will not pay attention to the cargo status. If the driver is required to report the cargo status in real time, it will inevitably affect the driver's rest and cause safety hazards.
[0005] In summary, the existing method of identifying the status of goods based on manual identification has poor timeliness and accuracy, and is costly. Summary of the invention
[0006] The main purpose of the present application is to provide a method, device and multi-classification modeling method for cargo status prediction to solve the problems of poor timeliness and accuracy and high cost of existing cargo status identification methods based on manual identification.
[0007] In order to achieve the above objective, according to a first aspect of the present application, a method for predicting cargo status is provided.
[0008] The method for cargo status prediction according to the present application includes:
[0009] Obtaining N consecutive raw data points most recently generated by the vehicle-mounted weighing device, wherein the raw data points record the raw measurement data, where N is an integer greater than or equal to 2;
[0010] According to N consecutive original data points, 1 whole segment data and M segment data are obtained, where M is an integer greater than or equal to 2;
[0011] One whole segment of data and M segmented data are input into a multi-classification model to obtain a current prediction state. The multi-classification model is a model that can predict the current state of the cargo based on the original measurement data of the vehicle-mounted weighing device.
[0012] Optionally, the step of inputting one whole segment of data and M segmented data into a multi-classification model further includes:
[0013] One whole segment of data and M segmented data are processed separately to obtain processed data, and the processed data are input into the multi-classification model.
[0014] Optionally, the processing of the 1 whole segment of data and the M segmented data to obtain processed data, and inputting the processed data into the multi-classification model further includes:
[0015] The 1 whole segment data and the M segment data are characterized to obtain feature data, and the feature data are input into the multi-classification model.
[0016] Optionally, the original measurement data includes at least one of deformation data, resistance data, voltage data, current data or time data.
[0017] In order to achieve the above objective, according to a second aspect of the present application, a method for multi-classification modeling is provided.
[0018] The method of multi-classification modeling according to the present application includes:
[0019] Acquire N consecutive original data points and their corresponding cargo states from the data set for multiple times, wherein the data set includes multiple original data points and their corresponding cargo states, wherein the original data points record original measurement data, and N is an integer greater than or equal to 2;
[0020] According to each of the N consecutive original data points and their corresponding cargo status, 1 whole segment data, M segment data and status data are obtained respectively, where M is an integer greater than or equal to 2;
[0021] All 1 whole segment data, M segment data and state data are input as training data into the multi-classification model to be trained to obtain the trained multi-classification model.
[0022] Optionally, obtaining M segmented data according to each of the consecutive N original data points includes:
[0023] The N original data points are segmented in such a way that each segmented data has data point intersections with any other segmented data to obtain M segmented data.
[0024] Optionally, the status data obtained according to each of the N consecutive original data points and their corresponding cargo statuses include:
[0025] The cargo status corresponding to the last raw data point among the N raw data points is used as the status data; or,
[0026] The status data is determined according to the number or proportion of the corresponding cargo statuses in the N original data points.
[0027] Optionally, all of the 1 whole segment data, the M segment data and the state data are input as training data into the multi-classification model to be trained, and further include:
[0028] Each of the 1 whole segment data and the M segment data is subjected to characterization processing to obtain feature data, and the feature data and state data corresponding to each of the 1 whole segment data and the M segment data are input into the multi-classification model to be trained.
[0029] In order to achieve the above-mentioned purpose, according to the third aspect of the present application, a device for predicting the status of goods is provided.
[0030] The device for predicting the cargo status according to the present application includes:
[0031] An acquisition unit, used to acquire N consecutive raw data points newly generated by the vehicle-mounted weighing device, wherein the raw data points record raw measurement data, and N is an integer greater than or equal to 2;
[0032] A data processing unit, used to process the data acquired by the acquisition unit to obtain input data required by the multi-classification model;
[0033] The prediction unit is used to input the input data obtained by the data processing unit into the multi-classification model to obtain the current prediction state, wherein the multi-classification model is a model that can predict the current state of the goods based on the original measurement data of the vehicle-mounted weighing equipment or the feature data derived from the original measurement data; the prediction unit stores the trained multi-classification model.
[0034] Optionally, the data processing unit includes:
[0035] A segmentation module, used to obtain 1 whole segment of data and M segmented data according to N consecutive original data points obtained by the acquisition unit, where M is an integer greater than or equal to 2;
[0036] The feature module is used to perform feature processing on one whole segment of data and / or M segmented data obtained by the segmentation module to obtain feature data.
[0037] In order to achieve the above object, according to a fourth aspect of the present application, a multi-classification modeling device is provided, the device comprising:
[0038] An acquisition unit, used for acquiring N consecutive original data points and their corresponding cargo states from the data set for multiple times, wherein the data set includes multiple original data points and their corresponding cargo states, wherein the original data points record original measurement data, and N is an integer greater than or equal to 2;
[0039] A data processing unit, used to obtain 1 whole segment data, M segment data and status data according to each of the N consecutive original data points and their corresponding cargo status, where M is an integer greater than or equal to 2;
[0040] The training unit is used to input all 1 whole segment data, M segment data and state data as training data into the multi-classification model to be trained to obtain the trained multi-classification model.
[0041] In order to achieve the above object, according to a fifth aspect of the present application, a method for predicting cargo status is provided, the method comprising:
[0042] Obtaining N consecutive raw data points most recently generated by the vehicle-mounted weighing device, wherein the raw data points record the raw measurement data, where N is an integer greater than or equal to 2;
[0043] Performing characterization processing on N consecutive original data points to obtain characteristic data;
[0044] The feature data is input into the multi-classification model to predict the status of the goods.
[0045] In order to achieve the above-mentioned purpose, according to the sixth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method for predicting the status of goods described in any one of the first aspect or the fifth aspect and / or the method for multi-classification modeling described in any one of the second aspect.
[0046] In order to achieve the above-mentioned purpose, according to the seventh aspect of the present application, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor executes the method for predicting the status of goods described in any one of the first aspect or the fifth aspect and / or the method for multi-classification modeling described in any one of the second aspect.
[0047] In the method and device for predicting the cargo state described in the present application, the latest N continuous raw data points generated by the vehicle-mounted weighing device are obtained, and the raw data points record the original measurement data; 1 whole segment data and M segment data are obtained according to the continuous N raw data points; 1 whole segment data and M segment data are input into a multi-classification model to obtain the current predicted state. Among them, the multi-classification model is a model that can predict the current state of the cargo according to the raw measurement data of the vehicle-mounted weighing device.
[0048] In the multi-classification modeling method of the present application, N consecutive original data points and their corresponding cargo states are obtained multiple times from a data set, the data set includes multiple original data points and their corresponding cargo states, and the original measurement data are recorded in the original data points; according to each of the N consecutive original data points and their corresponding cargo states, 1 whole segment data, M segment data and status data are obtained respectively; all of the 1 whole segment data, M segment data and status data are input as training data into the multi-classification model to be trained, and a trained multi-classification model is obtained.
[0049] When predicting the status of goods, this application makes predictions based on a multi-classification model. No manual identification is required during the prediction process, which can significantly improve the accuracy and timeliness of cargo status identification while reducing costs. In addition, when predicting, only the original data of the on-board weighing equipment is needed to predict the status of the goods. There is no need to obtain a series of other data such as vehicle speed and GPS, nor is there a need to establish a separate prediction model for a specific vehicle. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The schematic embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application. In the drawings:
[0051] Figure 1 It is a flow chart of a method for multi-classification modeling provided according to an embodiment of the present application;
[0052] Figure 2 It is a flow chart of a method for predicting cargo status provided according to an embodiment of the present application;
[0053] Figure 3 It is a block diagram of a device for predicting the status of goods provided according to an embodiment of the present application;
[0054] Figure 4 It is a block diagram of another device for predicting the state of goods provided in an embodiment of the present application;
[0055] Figure 5 It is a block diagram of the composition of a multi-classification modeling device provided according to an embodiment of the present application. DETAILED DESCRIPTION
[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0057] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged in appropriate circumstances to make the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.
[0058] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0059] According to an embodiment of the present application, a method for multi-classification modeling is provided, such as Figure 1 As shown, the method comprises the following steps:
[0060] S101. Obtain N consecutive original data points and their corresponding cargo states from the data set multiple times.
[0061] The cargo status corresponding to each original data point in the data set is obtained through data annotation. For example, the historical data contains multiple original data points generated by the on-board weighing equipment, and there are many cargo statuses (which can be set manually). When annotating the data, it can be marked according to the cargo status that needs to be predicted. Assuming that the cargo status that needs to be predicted is loading start, loading completion, unloading start, and unloading completion, then the original data points in the historical data can be annotated with the four cargo statuses of loading start, loading completion, unloading start, and unloading completion according to the actual situation. For the original data points that do not match these four cargo statuses in the actual situation, they can be annotated with others, thus forming a data set.
[0062] For example, a certain raw data point can correspond to a cargo state, such as the cargo state corresponding to raw data point 1 is other states, the cargo state corresponding to raw data point 2 is other states, the cargo state corresponding to raw data point 3 is the start of loading state, and so on. This data set contains multiple raw data points and their corresponding cargo states. In this example, there are five raw data points and their corresponding cargo states, namely, start of loading, loading completed, start of unloading, unloading completed, and other cargo states. It should be noted that the above five cargo states are an example, and in actual applications, the cargo states can also be adaptively adjusted according to actual needs.
[0063] Furthermore, in order to ensure the accuracy of the model, the number of raw data points corresponding to each cargo state is set in the data set. Taking the above five cargo states as an example, the number of raw data points corresponding to each cargo state is not less than 10% of the total number of raw data points in the data set (specifically, the number of raw data points in the data set can be adjusted by random downsampling); or, the number of raw data points corresponding to the four states of starting loading, loading completion, starting unloading, and unloading completion is not less than 10% of the total number of raw data points in the data set; or, the ratio between the raw data points corresponding to the five states of starting loading, loading completion, starting unloading, unloading completion, and other states is 1:1:1:1:2. It should be noted that the specific data values of 10%, 1:1:1:1:2 in the above examples can be adaptively adjusted according to the type of cargo state and actual needs in actual applications.
[0064] The original data points record the original measurement data generated by the vehicle-mounted weighing device. For example, a vehicle-mounted weighing device that uses the deformation data of an object to obtain weighing data (or weight data) through a specific calculation process will first obtain the deformation data of the object through its own measuring device, and then use this data to calculate the weighing data. The original measurement data here is the deformation data of the object; for another example, a vehicle-mounted weighing device that uses the resistance data of an object to obtain weighing data through a specific calculation process will first obtain the resistance data of the object through its own measuring device, and then use this data to calculate the weighing data. The original measurement data here is the resistance data of the object; of course, there are also vehicle-mounted weighing devices manufactured using other measurement principles, such as vehicle-mounted weighing devices manufactured using measured current or voltage data, and vehicle-mounted weighing devices manufactured using time data generated when ultrasonic waves measure distance (such as the time it takes for ultrasonic waves to go back and forth). Therefore, as long as it is the original measurement data in the vehicle-mounted weighing device, it will be sufficient.
[0065] It should be noted that the original measurement data is not the weighing data or weight data finally obtained by the on-board weighing equipment, because the existing on-board weighing equipment will adjust the calculation process based on influencing factors such as ambient temperature and vehicle speed, so that the weighing data is close to the actual value of the cargo weight in real time. Therefore, different original measurement data may produce the same weighing data after calculation. If these weighing data are used, it is impossible to establish a good prediction model.
[0066] Since the original data points in the data set and their corresponding cargo status are obtained from historical data, the order relationship between each original data point is also known (such as the order relationship obtained by the time when the original data point was generated), so it is also possible to obtain N consecutive original data points and their corresponding cargo status from the data set. Where N is an integer greater than or equal to 2, for example, N = 30, 40, etc.
[0067] A continuous set of N original data points and their corresponding cargo status cannot meet the data volume requirement for model training. The more training data there is, the higher the accuracy of the final model will be. Therefore, multiple continuous N original data points and their corresponding cargo status are needed to obtain training data.
[0068] S102. Obtain 1 whole segment of data, M segmented data and status data according to each of the N consecutive original data points and their corresponding cargo status.
[0069] Among them, 1 whole segment of data contains N continuous original data points. M (M is an integer greater than or equal to 2) segmented data are obtained by segmenting the N original data points in a way that each segmented data has a data point intersection with any other segmented data. That is, each segment of data contains multiple continuous original data points among the continuous N original data points, and each segment of data has an intersection with at least one other segmented data. It should be noted that segmenting the data is also to improve the accuracy of the model.
[0070] A specific example is given for illustration, for example, N=30, M=3, and the three segmented data are: segmented data 1: the 1st to 16th original data points; segmented data 2: the 7th to 23rd original data points; segmented data 3: the 15th to 30th original data points.
[0071] In this embodiment, two methods for determining status data are provided:
[0072] The first method is to use the cargo status corresponding to the last raw data point among the N consecutive raw data points as the status data (e.g., if the cargo status corresponding to the last raw data point is the start of loading, then the status data is the start of loading);
[0073] The second method is to determine the status data based on the number or proportion of the corresponding cargo states among the N original data points (e.g., among the 30 original data points, 10 of the original data points correspond to other cargo states, and 20 of the original data points correspond to the start loading cargo state. Since the number of original data points corresponding to the start loading cargo state is relatively large, the status data is the start loading cargo state, that is, the cargo state corresponding to the largest number of the N original data points is used as the status data).
[0074] In addition, it should be noted that in the process of obtaining the whole data, segmented data and status data, the original data points and their corresponding cargo status can be preprocessed according to actual needs, such as adjusting the format and digitizing, etc., to form the whole data, segmented data and status data required for model training. The preprocessing method can use existing technology.
[0075] S103. All 1 whole segment data, M segment data and state data are input as training data into the multi-classification model to be trained to obtain the trained multi-classification model.
[0076] All the 1 whole segment data, M segment data and status data obtained in the above steps are used as training data, and input into the multi-classification model to be trained for training to obtain a multi-classification model. The input of the multi-classification model is 1 whole segment data and M segment data, and the output of the model is the current cargo status. For the above example, the output result of the trained model is one of the five types: start loading, complete loading, start unloading, complete unloading and others (others can also be states that do not need to be predicted, etc.). There are many specific model training methods, which are not limited here, as long as a trained multi-classification model can be obtained. In addition, the multi-classification model to be trained can be any model that can achieve multi-classification.
[0077] In addition, the whole segment of data and the M segmented data may be processed separately to obtain processed data, and the processed data and the state data may be input into the multi-classification model to be trained to obtain the trained multi-classification model. More specifically, the processing includes preprocessing and / or characterization processing, or other processing means.
[0078] Furthermore, before all the 1 whole segment data, M segment data and state data are input as training data into the multi-classification model to be trained, each of the 1 whole segment data and M segment data can be characterized to obtain feature data, and then the feature data and state data corresponding to each of the 1 whole segment data and M segment data are input into the multi-classification model to be trained to obtain the multi-classification model. Specifically, the characterization process is described as follows:
[0079] (1) The median of each segment of data is used as feature data.
[0080] That is, the median of 1 whole segment data and M segment data is used as feature data. In combination with the above specific example, 1 whole segment data is 30 original measurement data, and the 3 segment data are segment data 1: the 1st to 16th original measurement data; segment data 2: the 7th to 23rd original measurement data; segment data 3: the 15th to 30th original measurement data. According to the calculation method of the median, the median of each segment data is obtained.
[0081] (2) The ratio of the medians of each segment of data is used as feature data.
[0082] According to the method in (1) above, the median of each segment of data can be obtained, and then the medians of each segment are calculated in pairs to obtain the ratio of each median to obtain the feature data.
[0083] (3) The standard deviation of the difference value of each segment of data is used as the feature data.
[0084] The difference value of each segment of data is calculated to obtain multiple difference values, and the standard deviation of the multiple difference values is calculated to obtain the characteristic data.
[0085] It should be noted that the above (1)-(3) are three characterization processing methods. In practical applications, any one or a combination of them can be selected for characterization processing.
[0086] From the above description, it can be seen that in the multi-classification modeling method of the embodiment of the present application, when performing data processing, the original data points selected are the original measurement values of the vehicle-mounted weighing equipment, and segmented processing is also performed, thereby improving the accuracy of the model.
[0087] In addition, it should be noted that in actual applications, multi-classification modeling can also add vehicle speed data, vehicle positioning data and other data related to the status of the cargo as input data of the model during training, which still falls within the scope of protection of this technical solution.
[0088] In addition, the present invention also provides another multi-classification modeling method, which includes the following steps:
[0089] First, N consecutive original data points and their corresponding cargo states are obtained from the data set multiple times. The data set contains multiple original data points and their corresponding cargo states. The original data points record the original measurement data, and N is an integer greater than or equal to 2.
[0090] The implementation of this step can be found in Figure 1 The implementation method of step S101 will not be described in detail here.
[0091] Secondly, according to each of the N consecutive original data points and their corresponding cargo status, a whole segment of data and status data are obtained respectively;
[0092] This step is implemented in the same way as Figure 1 The difference in the implementation of step S102 is that it is not necessary to segment a whole segment of data to obtain M segmented data. The implementation methods for obtaining 1 whole segment of data and state data are the same and will not be repeated here. It should be noted that the fact that M segmented data are not obtained here and that M segmented data are not used in subsequent training is the difference between this multi-classification modeling method and the aforementioned multi-classification modeling method. Apart from this, other processing methods are similar.
[0093] Finally, each whole segment of data is characterized to obtain feature data, and the feature data and state data corresponding to each whole segment of data are input into the multi-classification model to be trained to obtain the multi-classification model.
[0094] The implementation method of characterizing each whole segment of data to obtain feature data is the same as the implementation method of characterizing 1 whole segment of data and M segmented data in the aforementioned embodiment, which will not be repeated here. After obtaining the feature data, the feature data and status data corresponding to all the whole segments of data are input into the multi-classification model to be trained for training to obtain a multi-classification model. The input of the multi-classification model is 1 whole segment of data, and the output result of the model is the current cargo status. For the aforementioned example, the output result of the trained model is one of the five types: starting loading, completing loading, starting unloading, completing unloading, and others (others may also be states that do not need to be predicted, etc.). There are many specific model training methods, which are not limited here, as long as a trained multi-classification model can be obtained. In addition, the multi-classification model to be trained can be any model that can achieve multi-classification.
[0095] According to an embodiment of the present application, a method for predicting cargo status is provided, such as Figure 2 As shown, the method comprises the following steps:
[0096] S201. Obtain the latest N consecutive raw data points generated by the vehicle-mounted weighing device, in which the raw data points record the raw measurement data.
[0097] It should be noted that the value of N must be consistent with the value of N selected during the multi-classification model training. The properties of the original data points are also the same as those of the original data points involved in the multi-classification model training process. That is, if the original data points used in the multi-classification model training are deformation data, then the data obtained in this step is also deformation data; if it is resistance data, then the data obtained is also resistance data; and so on.
[0098] S202. Obtain 1 whole segment of data and M segmented data according to N consecutive original data points.
[0099] This step is implemented in the same way as Figure 1 The method of obtaining one whole segment of data and M segmented data according to N consecutive original data points in step S102 is the same and will not be repeated here. In addition, the values of N and M are the same as those selected during actual training.
[0100] S203. Input 1 whole segment of data and M segmented data into the multi-classification model to obtain the current prediction state.
[0101] The 1 whole segment data and M segment data obtained in step S202 are input into the above Figure 1 In the embodiment, the multi-classification model trained is used for prediction to obtain the current prediction state, that is, the current cargo state. The current prediction state is any one of the output results when the multi-classification model is trained. For example, the current prediction state can be one of the five types: starting loading, completing loading, starting unloading, completing unloading, and others.
[0102] In addition, the whole segment of data and the M segmented data may be processed separately to obtain processed data, and the processed data may be input into the multi-classification model. More specifically, the processing includes preprocessing and / or characterization processing, or other processing means.
[0103] In addition, it should be noted that, corresponding to the training process of the multi-classification model, if a whole segment of data and M segmented data are preprocessed and / or characterized, then when making predictions, the whole segment of data and M segmented data obtained from the latest N consecutive original data points also need to be preprocessed and / or characterized in the same way. That is, the whole segment of data and M segmented data are characterized to obtain feature data, and the feature data are input into the trained multi-classification model to obtain the current prediction state.
[0104] From the above description, it can be seen that in the method for predicting the cargo state of this embodiment, the latest continuous N original data points generated by the vehicle-mounted weighing device are obtained, and the original measurement data is recorded in the original data points; 1 whole segment data and M segment data are obtained according to the continuous N original data points; 1 whole segment data and M segment data are input into the multi-classification model to obtain the current predicted state. Among them, the multi-classification model is a model that can predict the current state of the cargo according to the original measurement data of the vehicle-mounted weighing device. Specifically, in the multi-classification modeling method, N continuous original data points and their corresponding cargo states are obtained from the data set for multiple times, and the data set contains multiple original data points and their corresponding cargo states, and the original measurement data is recorded in the original data points; 1 whole segment data, M segment data and state data are obtained according to each continuous N original data point and its corresponding cargo state; all 1 whole segment data, M segment data and state data are input into the multi-classification model to be trained as training data to obtain the trained multi-classification model.
[0105] Therefore, in this application, when predicting the cargo status, the prediction is made based on a multi-classification model, and no manual identification is required during the prediction process, which can significantly improve the accuracy and timeliness of cargo status identification, while also reducing costs. In addition, when predicting, only the original data of the on-board weighing equipment is needed to predict the cargo status, without the need to obtain a series of other data such as vehicle speed and GPS, and without the need to establish a separate prediction model for a specific vehicle.
[0106] In addition, the embodiment of the present application also provides a method for predicting the status of goods, which specifically includes the following steps.
[0107] First, obtain the latest N consecutive raw data points generated by the vehicle-mounted weighing equipment, in which the raw data points record the raw measurement data, where N is an integer greater than or equal to 2;
[0108] This step is implemented in the same way Figure 2 The implementation method of step S201 is the same and will not be repeated here.
[0109] Secondly, N consecutive original data points are characterized to obtain feature data, and the feature data is input into the multi-classification model to obtain the current prediction state.
[0110] It should be noted that characterizing N consecutive raw data points to obtain feature data is equivalent to obtaining one whole segment of data based on N consecutive raw data points, and characterizing one whole segment of data to obtain feature data. Specifically, the processing method for characterizing one whole segment of data to obtain feature data is the same as the method for characterizing one whole segment of data during multi-classification model training, which will not be repeated here. The corresponding feature data is obtained after characterization processing. Then the obtained feature data is input into the multi-classification model obtained by previous training to obtain the current prediction state, that is, the current cargo state. The current prediction state is any one of the output results when the multi-classification model is trained. For example, the current prediction state can be one of the five types of loading, loading completion, unloading, unloading completion, and others. It should be noted that the multi-classification model for feature data input here is a multi-classification model obtained by inputting the feature data and state data corresponding to each whole segment of data into the multi-classification model to be trained.
[0111] From the above description, it can be seen that in the method for predicting the cargo state of the embodiment of the present application, the latest continuous N raw data points generated by the vehicle-mounted weighing equipment are obtained, and the raw data points record the original measurement data; the continuous N raw data points are characterized to obtain feature data, and the feature data is input into the multi-classification model to obtain the current predicted state. Among them, the multi-classification model is a model that can predict the current state of the cargo based on the raw measurement data of the vehicle-mounted weighing equipment. Specifically, in the multi-classification modeling method, N continuous raw data points and their corresponding cargo states are obtained from the data set for multiple times, and the data set contains multiple raw data points and their corresponding cargo states, and the raw data points record the original measurement data; according to each continuous N raw data point and its corresponding cargo state, 1 whole segment of data and state data are obtained respectively; each 1 whole segment of data is characterized to obtain feature data, and the feature data and state data corresponding to each 1 whole segment of data are input into the multi-classification model to be trained to obtain the trained multi-classification model.
[0112] Therefore, in this application, when predicting the cargo status, the prediction is made based on a multi-classification model, and no manual identification is required during the prediction process, thereby improving the accuracy and timeliness of cargo status identification, while also reducing costs. In addition, when predicting, only the original data of the on-board weighing equipment is needed to predict the cargo status, without the need to obtain a series of other data such as vehicle speed and GPS, nor is it necessary to establish a separate prediction model for a specific vehicle.
[0113] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0114] According to an embodiment of the present application, there is also provided a method for implementing the above Figure 2 Method for predicting the state of goods, such as Figure 3 As shown, the device comprises:
[0115] An acquisition unit 31 is used to acquire N consecutive raw data points newly generated by the vehicle-mounted weighing device, wherein the raw data points record raw measurement data, and N is an integer greater than or equal to 2;
[0116] The data processing unit 32 is used to process the data acquired by the acquisition unit 31 to obtain the input data required by the multi-classification model;
[0117] The prediction unit 33 is used to input the input data obtained by the data processing unit 32 into the multi-classification model to obtain the current prediction state;
[0118] Wherein, the multi-classification model is a model capable of predicting the current state of the cargo based on the original measurement data of the vehicle-mounted weighing equipment or the feature data derived from the original measurement data;
[0119] Among them, the prediction unit stores the trained multi-classification model.
[0120] Further, such as Figure 4 As shown, the data processing unit 32 also includes,
[0121] The segmentation module 321 is used to obtain 1 whole segment of data and M segmented data according to the N consecutive original data points obtained by the acquisition unit 31, where M is an integer greater than or equal to 2;
[0122] Further, such as Figure 4 As shown, the data processing unit 32 further includes a feature module 322 for performing feature processing on one whole segment of data and / or M segmented data obtained by the segmentation module 321 to obtain feature data.
[0123] Specifically, the specific process of each unit and module in the device of the embodiment of the present application to realize its function can be found in the relevant description in the method embodiment, which will not be repeated here.
[0124] According to an embodiment of the present application, there is also provided a method for implementing the above Figure 1 The method of multi-classification modeling is as follows Figure 5 As shown, the device comprises:
[0125] An acquisition unit 41 is used to acquire N consecutive original data points and their corresponding cargo states from the data set for multiple times, wherein the data set includes multiple original data points and their corresponding cargo states, wherein the original data points record original measurement data, and N is an integer greater than or equal to 2;
[0126] The data processing unit 42 is used to obtain 1 whole segment data, M segment data and status data according to each of the N consecutive original data points and their corresponding cargo status, where M is an integer greater than or equal to 2;
[0127] The training unit 43 is used to input all 1 whole segment data, M segment data and state data as training data into the multi-classification model to be trained to obtain the trained multi-classification model.
[0128] Specifically, the specific process of each unit and module in the device of the embodiment of the present application to realize its function can be found in the relevant description in the method embodiment, which will not be repeated here.
[0129] From the above description, it can be seen that in the multi-classification modeling device of the embodiment of the present application, when performing data processing, the original data points selected are the original measurement values of the vehicle-mounted weighing equipment, and segmented processing is also performed, thereby improving the accuracy of the model.
[0130] According to an embodiment of the present application, a computer-readable storage medium is also provided, which stores computer instructions, and the computer instructions are used to enable the computer to execute the method for predicting cargo status and / or the method for multi-classification modeling in the above method embodiments.
[0131] According to an embodiment of the present application, an electronic device is also provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor executes the method for predicting cargo status and / or the method for multi-classification modeling in the above-mentioned method embodiment.
[0132] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the present application is not limited to any specific combination of hardware and software.
[0133] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A method for predicting cargo status, It is characterized in that The method comprises: Obtaining N continuous raw data points currently generated by the vehicle-mounted weighing device, wherein the raw data points record raw measurement data, N is an integer greater than or equal to 2, and the raw measurement data includes at least one of deformation data, resistance data, voltage data, current data or time data; According to the continuous N original data points, a whole segment of data and M segmented data are obtained, where M is an integer greater than or equal to 2, and the M segmented data are obtained by segmenting the N original data points, and each segmented data in the M segmented data has a data point intersection with any other segmented data; The one whole segment of data and the M segmented data are input into a multi-classification model to obtain a current prediction state, wherein the multi-classification model is a model that can predict the current state of the cargo based on the original measurement data of the on-board weighing equipment, and the multi-classification model is trained based on continuous original data points in the data set and the cargo states corresponding to the continuous original data points in the data set.
2. The method for cargo status prediction according to claim 1, It is characterized in that The step of inputting the one whole segment of data and the M segmented data into a multi-classification model comprises: The one whole segment of data and the M segmented data are processed respectively to obtain processed data, and the processed data are input into the multi-classification model, wherein the processing includes at least one of preprocessing and characterization processing, and the preprocessing includes adjusting the format and digitizing.
3. The method for cargo status prediction according to claim 2, It is characterized in that The processing is characterization processing, and the processed data is feature data, which includes at least one of the following: the median of the whole segment data and the M segmented data, the pairwise ratio of the medians, and the standard deviation of the difference value between the whole segment data and the M segmented data.
4. A method for multi-classification modeling, It is characterized in that The method is used to obtain a multi-classification model in the method for cargo status prediction according to any one of claims 1 to 3, and the method comprises: Acquire N consecutive raw data points and their corresponding cargo states from the data set for multiple times, wherein the data set includes multiple raw data points and their corresponding cargo states, wherein the raw data points record raw measurement data, N is an integer greater than or equal to 2, and the raw measurement data includes at least one of deformation data, resistance data, voltage data, current data, or time data; According to each of the continuous N original data points and the corresponding cargo status, a whole segment of data, M segmented data and status data are obtained respectively, where M is an integer greater than or equal to 2, and the M segmented data are obtained by segmenting the N original data points, and each segmented data in the M segmented data has a data point intersection with any other segmented data; All of the entire segment data, M segmented data and status data are input as training data into the multi-classification model to be trained to obtain a trained multi-classification model, which is a model that can predict the current status of the cargo based on the original measurement data of the on-board weighing equipment.
5. A device for predicting the status of goods, It is characterized in that The device is used to implement the method for predicting the status of goods according to any one of claims 1 to 3, and the device comprises: An acquisition unit, used to acquire N continuous raw data points currently generated by the vehicle-mounted weighing device, wherein the raw data points record raw measurement data, N is an integer greater than or equal to 2, and the raw measurement data includes at least one of deformation data, resistance data, voltage data, current data or time data; A data processing unit, used to process the N original data points acquired by the acquisition unit to obtain input data required by the multi-classification model, wherein the input data includes M segmented data, and the M segmented data are obtained by segmenting the N original data, and each segmented data in the M segmented data has a data point intersection with any other segmented data; A prediction unit is used to input the input data obtained by the data processing unit into a multi-classification model to obtain a current prediction state, wherein the multi-classification model is a model that can predict the current state of the cargo based on the original measurement data of the vehicle-mounted weighing equipment or the feature data derived from the original measurement data, and the multi-classification model is trained based on continuous original data points in the data set and the cargo states corresponding to the continuous original data points in the data set, and the prediction unit stores the trained multi-classification model.
6. The device for predicting cargo status according to claim 5, It is characterized in that The data processing unit comprises: A segmentation module, used for obtaining a whole segment of data and M segmented data according to the N consecutive original data points acquired by the acquisition unit, where M is an integer greater than or equal to 2; The feature module is used to perform feature processing on a whole segment of data and / or M segmented data obtained by the segmentation module to obtain feature data, and the input data is the feature data.
7. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for predicting the status of goods described in any one of claims 1 to 3 and / or the method for multi-classification modeling described in claim 4.
8. An electronic device, It is characterized in that include: at least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores a computer program, and the computer program is executed by the at least one processor so that the at least one processor executes the method for predicting the status of goods described in any one of claims 1 to 3 and / or the method for multi-classification modeling described in claim 4.
