A fresh agricultural product transport truck intelligent rapid detection method based on multi-source data fusion

Abstract

The present application relates to the field of expressway green channel violation monitoring, and particularly relates to a kind of fresh agricultural products transport truck intelligent rapid detection method based on multi-source data fusion. Including the following steps: step one, build goods packaging loading proportion benchmark database and empty car mass benchmark database;Step two, collect total weight of loaded vehicle, actual vehicle information and transport goods information, obtain goods packaging loading proportion;Step three, obtain difference absolute value according to actual goods loading quality and theoretical maximum goods loading quality;Step four, determine trust level according to the ratio of difference absolute value and theoretical load;The trust level includes mild suspicion and severe suspicion;For the vehicle determined as severe suspicion, it is suggested to carry out artificial strict inspection, and the vehicle determined as mild suspicion is normally inspected. The present application can automatically identify insufficient loading, mixed loading goods and other various violations, significantly improve the accuracy and efficiency of green channel vehicle detection.

Description

Technical Field

[0001] This invention relates to the field of monitoring violations in highway green channels, specifically to an intelligent and rapid detection method for trucks transporting fresh agricultural products based on multi-source data fusion. Background Technology

[0002] The transportation of fresh agricultural products is eligible for toll reductions or exemptions. Vehicles enjoying green channel benefits must carry fresh agricultural products accounting for more than 80% of the vehicle's approved load capacity or cargo volume.

[0003] However, in actual operation, violations such as insufficient loading capacity, mixing of non-green channel goods, and false reporting of goods types exist, leading to the abuse of the policy and loss of toll fees. Currently, the inspection of green channel vehicles mainly relies on manual inspection, which has the following limitations: low inspection efficiency, affecting traffic efficiency; and strong subjectivity in manual judgment, lacking unified standards. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and propose an intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion.

[0005] A method for intelligent and rapid detection of fresh agricultural products transport trucks based on multi-source data fusion, such as... Figure 1 As shown, it includes the following steps: Step 1: Construct a benchmark database for cargo packaging loading ratio and an empty vehicle weight benchmark database; Step 2: Collect the total weight of the cargo vehicle, actual vehicle information, and transported cargo information to obtain the cargo packaging loading ratio; Step 3: Calculate the predicted empty vehicle weight and the theoretical maximum cargo load weight. Then, calculate the actual cargo load weight using the total weight of the loaded vehicle and the predicted empty vehicle weight. Obtain the absolute value of the difference between the actual cargo load weight and the theoretical maximum cargo load weight. Step 4: Determine the trust level based on the ratio of the absolute value of the difference to the theoretical load capacity; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection should be conducted.

[0006] Furthermore, a benchmark database of cargo packaging loading proportions was constructed using field measurement and theoretical estimation methods to obtain the proportion data of the whole container.

[0007] Furthermore, for goods that are approximately spherical, the mass of a single item and the specific gravity of the entire container are derived through physical modeling. Let the diameter of a single item be d (m) and its density be ρ (kg / m³), then the formula for calculating the mass of a single item is: (1) Where: ρ is the density of a single cargo, in kg / m³; d is the diameter of a single cargo; Using the cubic close packing theory, and assuming the standard container dimensions are l×w×h (m), the theoretical maximum number of components that can be loaded into the standard container is: (2) Where: l is the length of the standard container, w is the width of the standard container, h is the height of the standard container, and d is the diameter of a single item; The specific gravity of a container is defined as the ratio of the total mass of the goods to the volume of the container. (3) Where: n max m represents the maximum number of items that can be loaded into a theoretical standard container; m is the mass of a single item. The theoretical estimation method specifically filters samples by setting a space utilization threshold (η≥0.52), and the formula for calculating the space utilization rate is as follows: (4) Where: n max V represents the maximum number of components that can be loaded into a theoretical standard housing. single Let l be the volume of a single box, w be the length of a standard box, and h be the height of a standard box.

[0008] Furthermore, the on-site measurement method specifically involves: first, using a measuring tape to accurately measure the length, width, and height of the box, and then using an electronic scale to accurately weigh the goods and the entire box to obtain the actual specific gravity data of the entire box.

[0009] Furthermore, the method for calculating the predicted empty vehicle weight is as follows: (9) in: This is the first stage prediction value. This is the predicted value for the second stage.

[0010] Furthermore, in step three, the theoretical maximum cargo loading mass is specifically: (14) (15) Where: M theory n is the total weight of the cargo vehicle. max The maximum number of loads in a theoretical standard housing; m single n1 represents the theoretical mass of a single container; n2 represents the spatial arrangement constraint; n3 represents the mass limit constraint; and n4 represents the volume utilization constraint.

[0011] Furthermore, in step three, the actual cargo loading mass is calculated using the predicted values ​​of the total weight of the cargo vehicle and the empty vehicle weight, specifically as follows: (16) Where: M actual M represents the total weight of the cargo vehicle. empty This refers to the empty vehicle weight.

[0012] Furthermore, in step four, the determination of the trust level is specifically as follows: (17) Where: M theory M represents the theoretical maximum cargo loading mass. actual This refers to the total weight of the cargo vehicle.

[0013] On the other hand, the present invention provides a system for intelligent and rapid detection of fresh agricultural product transport trucks based on multi-source data fusion as described in any one of the above claims, comprising: Data acquisition module: Collects the total weight of the cargo vehicle, actual vehicle information and transported cargo information, and obtains the cargo packaging loading ratio; Data calculation module: Calculates the predicted empty vehicle weight and the theoretical maximum cargo loading weight, then calculates the actual cargo loading weight using the total weight of the cargo vehicle and the predicted empty vehicle weight, and obtains the absolute value of the difference between the actual cargo loading weight and the theoretical maximum cargo loading weight; Judgment Module: Determines the trust level based on the ratio of the absolute value of the difference to the theoretical load; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection is performed.

[0014] Compared with the prior art, the present invention has the following beneficial effects: To achieve an automatic detection method for abnormal loading of green channel vehicles on highways, this invention first constructs a cargo packaging loading proportion database through a combination of field surveys and theoretical estimations. This database includes key parameters such as the proportion of whole containers and packaging dimensions of green channel goods. Second, addressing the difficulty in accurately obtaining empty vehicle weight, an empty vehicle weight prediction method based on two-stage residual regression is proposed, predicting the empty vehicle weight based on features such as curb weight, vehicle type, cargo box type, and vehicle brand. Third, a loading rationality judgment model based on three-dimensional packing theory is established. The theoretical maximum loading capacity is calculated through spatial arrangement constraints, mass limitation constraints, and volume utilization constraints, and the loading rationality is determined by comparing it with actual weighing data. Through the comprehensive application of the above methods, this invention can automatically identify various violations such as insufficient loading and mixed cargo, significantly improving the accuracy and efficiency of green channel vehicle detection. Attached Figure Description

[0015] The accompanying drawings are incorporated in and form part of this specification, and together with the description serve to explain the principles of the invention.

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the intelligent rapid detection method for fresh agricultural product transport trucks according to the present invention; Figure 2 This is a flowchart illustrating the intelligent and rapid detection method for transporting fresh agricultural products by truck, as described in this invention. Detailed Implementation

[0018] Exemplary embodiments will now be described in detail. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples consistent with some aspects of the invention as detailed in the appended claims.

[0019] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Example

[0020] A method for intelligent and rapid detection of fresh agricultural products transport trucks based on multi-source data fusion, such as... Figure 1-2 As shown, it includes the following steps: Step 1: Construct a benchmark database for cargo packaging loading ratio and an empty vehicle weight benchmark database; Step 2: Collect the total weight of the cargo vehicle, actual vehicle information, and transported cargo information to obtain the cargo packaging loading ratio; 2.1 Data Acquisition and Preprocessing At the toll station entrance, vehicle information is obtained through a license plate recognition system, drivers declare the type of goods being transported, origin and other information, the total weight of the vehicle at the entrance is obtained through a weighbridge system, the dimensions of the cargo compartment are obtained through a 3D laser scanning system, and parameters such as curb weight, vehicle model, cargo compartment type and vehicle brand are extracted from the vehicle registration information to complete the data collection and preprocessing work.

[0021] 2.2 Cargo Packaging Loading Proportion Database Query Based on the type of goods declared by the driver, the system retrieves the corresponding container weight and box dimensions, including key information such as length, width, and height, from the cargo packaging loading weight database. This database contains parameters such as the container weight and packaging dimensions of green channel goods, serving as the foundation for subsequent calculations of theoretical loading capacity.

[0022] Step 3: Calculate the predicted empty vehicle weight and the theoretical maximum cargo load weight. Then, calculate the actual cargo load weight using the total weight of the loaded vehicle and the predicted empty vehicle weight. Obtain the absolute value of the difference between the actual cargo load weight and the theoretical maximum cargo load weight. 3.1 Empty vehicle weight prediction After cleaning the data and removing outliers whose weight gain coefficients do not meet the reasonable range, a linear regression model is first established using the curb weight to obtain the first-stage predicted value. The residual between the actual value and the predicted value is calculated. Then, features such as the type of carriage and the vehicle brand are input into the machine learning model to predict the residual. Finally, the first-stage predicted value and the prediction residual are added together to obtain the empty vehicle weight predicted value.

[0023] 3.2 Inspection of the rationality of cargo loading The bulk density and packaging dimensions of the declared cargo are retrieved from the cargo packaging loading density database. The weight of a single container is calculated as the length, width, height, and bulk density of the container. Based on the vehicle dimensions and approved load capacity, the theoretical maximum number of loads is calculated using a triple constraint model. This model includes spatial arrangement constraints, mass limitation constraints, and volume utilization constraints, and the minimum value of these three constraints is taken. The theoretical maximum cargo loading weight is then calculated as the theoretical maximum number of loads multiplied by the weight of a single container. The actual loading weight is then calculated as the total weighed weight minus the empty vehicle weight. The difference between the theoretical and actual values ​​is calculated, and finally, the trust level is determined based on the ratio of the absolute value of the difference to the theoretical load capacity, completing the rationality check for a single type of cargo loading.

[0024] Step 4: Determine the trust level based on the ratio of the absolute value of the difference to the theoretical load capacity; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection should be conducted.

[0025] 4.1 Comprehensive Evaluation and Report Generation The results of the loading rationality inspection are summarized, and manual review is recommended for vehicles deemed seriously suspicious or exhibiting multiple anomalies. An inspection report is generated, including vehicle information, cargo information, various inspection indicators, judgment results, and recommended measures. The inspection results are then stored in a database for subsequent big data analysis and model optimization.

[0026] In addition, continuous model optimization is necessary. New measured data should be collected regularly to update the cargo packaging loading ratio database, and more empty vehicle weight data should be accumulated to retrain the empty vehicle weight prediction model. Based on feedback from actual applications, various judgment thresholds and weights should be adjusted, and the algorithm should be continuously optimized to improve detection accuracy and efficiency.

[0027] Furthermore, a benchmark database of cargo packaging loading proportions was constructed using field measurement and theoretical estimation methods to obtain the proportion data of the whole container.

[0028] Furthermore, the theoretical estimation method specifically filters samples by setting a space utilization threshold, and the formula for calculating the space utilization rate is as follows: (4) Where: n max V represents the maximum number of components that can be loaded into a theoretical standard housing. single Let l be the volume of a single box, w be the length of a standard box, and h be the height of a standard box.

[0029] A. Database construction strategy The database of cargo packaging loading proportions was constructed using a dual-track parallel strategy, fully combining field measurement and theoretical estimation methods. 1. On-site measurement method: For goods for which samples are easy to obtain, conduct on-site surveys by going to fruit wholesale markets, vegetable wholesale markets and toll stations, use a tape measure to accurately measure the length, width and height of the box, and use an electronic scale to accurately weigh the goods and the whole box to obtain real and reliable whole box specific gravity data.

[0030] 2. Theoretical Estimation Method: For goods that are difficult to cover through on-site surveys due to seasonal or geographical limitations, a theoretical modeling method is used. For goods that are approximately spherical (such as apples, oranges, pears, etc.), the mass of a single item and the specific gravity of the whole box are derived through physical modeling. For goods that are approximately spherical, let the diameter of a single item be d (m) and the density be ρ (kg / m³), then the formula for calculating the mass of a single item is: (1) Using the cubic close packing theory, and assuming the standard box dimensions are l×w×h (m), the theoretical maximum number of loads is: (2) The specific gravity of a container is defined as the ratio of the total mass of the goods to the volume of the container. (3) Formula (4) is obtained from the above formulas (1)-(3). The samples are screened by setting a space utilization threshold (η≥0.52) to ensure the rationality of the theoretical data. Finally, a database of packaging loading proportions covering major green channel goods categories is formed by actual measurement data and theoretical estimation.

[0031] Furthermore, the on-site measurement method specifically involves: first, using a measuring tape to accurately measure the length, width, and height of the box, and then using an electronic scale to accurately weigh the goods and the entire box to obtain the actual specific gravity data of the entire box.

[0032] Furthermore, the method for calculating the predicted empty vehicle weight is as follows: (9) in: This is the first stage prediction value. This is the predicted value for the second stage.

[0033] B. Empty vehicle weight estimation method Empty vehicle weight is a key parameter for calculating actual load. To address the problems of traditional statistical methods, such as the complexity of vehicle brands, unbalanced data distribution, and interference from modification factors, this study proposes a two-stage residual regression model, organically combining the interpretability of linear regression with the nonlinear fitting capability of machine learning.

[0034] The weight gain factor d is defined to reflect the relative difference between the empty vehicle weight and the curb weight: (5) Where y is the actual empty vehicle weight, and x is the curb weight. Samples meeting the criteria (0 ≤ d ≤ 0.3) are retained after data cleaning.

[0035] The first stage establishes a linear regression relationship between curb weight and empty vehicle weight: (6) Where α is the intercept and β is the regression coefficient. These are the predicted values ​​for the first stage.

[0036] Calculate the residual between the actual value and the first-stage predicted value: (7) The second stage uses machine learning methods (CatBoost, XGBoost, RandomForest, LightGBM) to predict residuals, with features such as carriage type and vehicle brand as input: (Carriage type, vehicle brand) (8) The final empty vehicle weight prediction value is obtained by superimposing the results of the two stages, as shown in formula (9). Furthermore, in step three, the theoretical maximum cargo loading mass is specifically: (14) (15) Where: Mtheory n is the total weight of the cargo vehicle. max The maximum number of loads in a theoretical standard housing; m single n1 represents the theoretical mass of a single container; n2 represents the spatial arrangement constraint; n3 represents the mass limit constraint; and n4 represents the volume utilization constraint.

[0037] C. Detection Methods for Green Channel Vehicles with Abnormal Loading 1. Theoretical mass estimation of a single box Based on the aforementioned database and prediction model, a triple-constraint packing model is proposed for a single cargo transportation scenario. The full container weight is extracted from the cargo packaging loading weight database. Given the dimensions of the packaging box (l, w, h), calculate the theoretical mass of a single box: (10) 2. Triple constraint model calculation The geometric dimensions (L, W, H) and rated load capacity of the carriage were obtained using 3D laser scanning. By introducing the loading clearance factor g, a triple constraint model is established: Spatial layout constraints: (11) Quality constraints: (12) Volume utilization constraints: (13) From formulas (10)-(13), we can obtain formulas (14) and (15).

[0038] Furthermore, in step three, the actual cargo loading mass is calculated using the predicted values ​​of the total weight of the cargo vehicle and the empty vehicle weight, specifically as follows: (16) Where: M actual M represents the total weight of the cargo vehicle. empty To predict the weight of an empty vehicle.

[0039] Furthermore, in step four, the determination of the trust level is specifically as follows: (17) Where: M theory M represents the theoretical maximum cargo loading mass. actual This refers to the total weight of the cargo vehicle.

[0040] Trust level criteria are established based on the difference indicators: reasonable, doubtful, and highly doubtful.

[0041] For mixed cargo scenarios, the theoretical maximum loading mass M is calculated independently for each type of cargo i.theory,i Define statistical indicators: (18) Calculate the degree of deviation between the actual total weight and the statistical indicators, and establish rules for judging mixed loading scenarios.

[0042] On the other hand, this invention discloses a system for intelligent and rapid detection of fresh agricultural product transport trucks based on multi-source data fusion as described in any one of the above claims, comprising: Data acquisition module: Collects the total weight of the cargo vehicle, actual vehicle information and transported cargo information, and obtains the cargo packaging loading ratio; Data calculation module: Calculates the predicted empty vehicle weight and the theoretical maximum cargo loading weight, then calculates the actual cargo loading weight using the total weight of the cargo vehicle and the predicted empty vehicle weight, and obtains the absolute value of the difference between the actual cargo loading weight and the theoretical maximum cargo loading weight; Judgment Module: Determines the trust level based on the ratio of the absolute value of the difference to the theoretical load; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection is performed.

[0043] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention.

[0044] It should be understood that the present invention is not limited to the content already described above, and various modifications and changes can be made without departing from its scope. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for intelligent and rapid detection of fresh agricultural products transport trucks based on multi-source data fusion, characterized in that, Includes the following steps: Step 1: Construct a benchmark database for cargo packaging loading ratio and an empty vehicle weight benchmark database; Step 2: Collect the total weight of the cargo vehicle, actual vehicle information, and transported cargo information to obtain the cargo packaging loading ratio; Step 3: Calculate the predicted empty vehicle weight and the theoretical maximum cargo load weight. Then, calculate the actual cargo load weight using the total weight of the loaded vehicle and the predicted empty vehicle weight. Obtain the absolute value of the difference between the actual cargo load weight and the theoretical maximum cargo load weight. Step 4: Determine the trust level based on the ratio of the absolute value of the difference to the theoretical load capacity; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection should be conducted.

2. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 1, characterized in that, A benchmark database of cargo packaging loading proportions was constructed by using field measurement and theoretical estimation methods to obtain the proportion data of whole containers.

3. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 2, characterized in that, The theoretical estimation method specifically filters samples by setting a space utilization threshold, and the formula for calculating the space utilization rate is as follows: （4） Where: n max V represents the maximum number of components that can be loaded into a theoretical standard housing. single Let l be the volume of a single box, w be the length of a standard box, and h be the height of a standard box.

4. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 2, characterized in that, The on-site measurement method specifically involves: first, using a tape measure to accurately measure the length, width, and height of the box; then, using an electronic scale to accurately weigh the goods and the entire box to obtain the specific gravity data of the entire box.

5. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 1, characterized in that, The method for calculating the predicted empty vehicle weight is as follows: in: This is the first stage prediction value. This is the predicted value for the second stage.

6. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 1, characterized in that, In step three, the theoretical maximum cargo loading mass is specifically as follows: （14） （15） Where: M theory n represents the theoretical maximum cargo loading mass. max The maximum number of loads in a theoretical standard housing; m single n1 represents the theoretical mass of a single container; n2 represents the spatial arrangement constraint; n3 represents the mass limit constraint; and n4 represents the volume utilization constraint.

7. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 1, characterized in that, In step three, the actual cargo loading mass is calculated using the predicted values ​​of the total weight of the cargo vehicle and the empty vehicle weight, specifically as follows: （16） Where: M weigh M represents the total weight of the cargo vehicle. empty This refers to the empty vehicle weight.

8. The intelligent and rapid detection method for fresh agricultural product transport trucks based on multi-source data fusion according to claim 1, characterized in that, In step four, the trust level is determined as follows: （17） Where: M theory M represents the theoretical maximum cargo loading mass. actual This refers to the total weight of the cargo vehicle.

9. A system for intelligent and rapid detection of fresh agricultural product transport trucks based on multi-source data fusion as described in any one of claims 1 to 8, characterized in that, include: Data acquisition module: Collects the total weight of the cargo vehicle, actual vehicle information and transported cargo information, and obtains the cargo packaging loading ratio; Data calculation module: Calculates the predicted empty vehicle weight and the theoretical maximum cargo loading weight, then calculates the actual cargo loading weight using the total weight of the cargo vehicle and the predicted empty vehicle weight, and obtains the absolute value of the difference between the actual cargo loading weight and the theoretical maximum cargo loading weight; Judgment Module: Determines the trust level based on the ratio of the absolute value of the difference to the theoretical load; the trust level includes mild suspicion and severe suspicion; for vehicles determined to be severely suspicious, it is recommended to conduct a strict manual inspection, and for vehicles determined to be mildly suspicious, a normal inspection is performed.

Images