CN114742503A

CN114742503A - Intelligent logistics car sharing method and device based on deep learning

Info

Publication number: CN114742503A
Application number: CN202210390740.5A
Authority: CN
Inventors: 许旭然
Original assignee: Guangzhou Huisuan New Generation Information Technology Co ltd
Current assignee: Shandong Yalian Supply Chain Management Group Co ltd
Priority date: 2022-04-14
Filing date: 2022-04-14
Publication date: 2022-07-12
Anticipated expiration: 2042-04-14
Also published as: CN114742503B

Abstract

The invention discloses a smart logistics car sharing method and device based on deep learning, which comprises the steps of obtaining a waiting car sharing goods list, training a goods combination accident prediction model, combining a first goods with a second goods in a car sharing mode, shooting an image of the first goods after the first goods are loaded as a first image, shooting an image of the second goods after the second goods are unloaded as a second image, calculating all difference matrixes and coordinate positions of the first image and the second image, determining an image difference point set, listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule, and sending confusion verification reminding information to a carrier according to the quasi point transportation weight of the goods in the confusion reminding list. The method and the system effectively carry out car sharing planning through the block chain and the deep learning technology, improve the accuracy of the confusion reminding model, and improve the transportation timeliness and accuracy.

Description

Intelligent logistics car sharing method and device based on deep learning

Technical Field

The invention relates to the technical field of intelligent logistics, in particular to an intelligent logistics carpooling method and device based on deep learning.

Background

Although the existing logistics platform can find a vehicle suitable for matching, the transportation of goods is not only finished vehicle transportation, but also a lot of logistics orders with small goods quantity, so that when the goods quantity is small, the loading rate of the vehicle is low, the idle load rate is high, and further the logistics cost of a logistics company is high and the logistics order processing efficiency is low.

The carpooling behavior refers to the behavior that goods with multiple destinations or multiple departure places are loaded on the same vehicle for transportation in the transportation process, the existing carpooling solution usually costs logistics service quality, transportation cost is saved, meanwhile, goods are easily confused due to the fact that multiple goods share the same truck, and therefore wrong goods are unloaded to wrong destinations, unnecessary influences are caused, the goods are combined in a carpooling mode, the accident situation of goods transportation is reduced, and the problem is urgently needed to be solved by transportation manufacturers.

Disclosure of Invention

In view of at least one of the above technical problems, the present invention provides an intelligent logistics car-sharing method and apparatus based on deep learning.

On one hand, the embodiment of the invention comprises an intelligent logistics car-sharing method based on deep learning, which comprises the following steps:

the method comprises the steps of obtaining a waiting car sharing cargo list, determining cargo information in the waiting car sharing cargo list, obtaining truck information, selecting first cargo to be transported by a truck owner at a client according to the waiting car sharing cargo list, and determining the transportable residual volume and the transportable residual weight of the truck according to the cargo information of the first cargo and the truck information;

acquiring historical car pooling information-accident occurrence condition table, counting the goods combination condition of the historical car pooling, and the road transportation length, the road jolting condition, the transportation acceleration condition and the type name of an accident under the goods combination condition, taking the historical car pooling information-accident occurrence condition table, the road transportation length and the road jolting condition as characteristic values, taking the type name of the accident as a predicted value, and training a goods combination accident prediction model by combining a deep learning or machine learning classification algorithm;

eliminating the goods which are combined with the first goods in a wrong way from the waiting car sharing goods list, screening a candidate car sharing goods list according to the total volume of the goods, the weight of the goods, the transportable residual volume of the truck, the transportable residual weight of the truck, the destination of the goods and the initial position of the goods, predicting and traversing the candidate car sharing goods list based on the goods combination accident prediction model, and inputting the goods information of which the prediction result is no accident risk into a no accident risk goods list; selecting the second cargo based on the list of accident-free risk cargos;

carrying out car sharing transportation according to the first goods and the second goods, and arranging the placing positions of the first goods and the second goods;

counting the goods combination condition of the historical carpools, the road transportation length, the road jolt condition, the transportation acceleration condition and the goods confusion distance under the goods combination condition, and establishing a historical carpool information-confusion distance table;

determining the roi distance grade of goods confusion detection according to the historical car pooling information-confusion distance table, shooting an image after the first goods are loaded as a first image, shooting an image after the second goods are unloaded as a second image, calculating all difference matrixes and coordinate positions of the first image and the second image, determining an image difference point set, obtaining the total height of the first goods and the total height of the second goods according to the image difference point set, calculating the height difference of the goods, and listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule;

and sending confusion verification reminding information to a carrier according to the quasi-point transportation weight of the goods in the confusion reminding list and by combining the highest number of the goods which can be verified by the carrier.

Preferably, the obtaining a list of cargoes waiting for carpooling, determining cargo information in the list of cargoes waiting for carpooling, obtaining truck information, selecting a first cargo to be transported by a truck owner at a client according to the list of cargoes waiting for carpooling, and determining a remaining transportable volume and a remaining transportable weight of the truck according to the cargo information of the first cargo and the truck information includes:

the list of cargoes waiting for sharing is acquired by collecting and collecting in a freight transportation intermediary webpage, the freight transportation intermediary webpage contains freight information required by delivery, and the freight is composed of a pile of single cargoes with the same type, name and specification;

the cargo information includes: the method comprises the following steps of (1) obtaining a cargo name, a cargo weight, a cargo total volume, a cargo type, a cargo starting position, a cargo destination, a cargo total height, a cargo total length, a cargo total number, a cargo transportation reward, a quasi-point transportation weight of a cargo, a single cargo image, a single cargo width, a single cargo height and a single cargo weight;

the quasi-point transportation weight of the goods is the weight of the goods which must be transported to a destination at a quasi-point, and the weight is determined through system presetting;

the freight transportation reward is obtained after a truck owner transports the freight, and is recorded and set through an online trading platform;

acquiring truck information, wherein the truck information comprises: the total transportable volume of the truck, the total transportable weight of the truck, and the width of the boxcar;

the determining the remaining transportable volume and the remaining transportable weight of the truck according to the cargo information of the first cargo and the truck information includes: subtracting the cargo weight of the cargo information of the first cargo from the total transportable weight of the truck to obtain the remaining transportable weight of the truck; and subtracting the total cargo volume of the cargo information of the first cargo from the total cargo volume transportable by the truck to obtain the remaining transportable cargo volume of the truck.

Preferably, the obtaining, at the server, the historical car pooling information-accident occurrence table, counting the cargo combination situation of the historical car pooling, and under the cargo combination situation, the road transportation length, the road jolt situation, the transportation acceleration situation, and the type name of the occurred accident, taking the historical car pooling information-accident occurrence table, the road transportation length, and the road jolt situation as characteristic values, and taking the type name of the occurred accident as a predicted value, in combination with a deep learning or machine learning classification algorithm, training a cargo combination accident prediction model, includes:

the machine learning classification algorithm comprises: a Support Vector Machine (SVM), a naive Bayes classifier, a decision tree; the deep learning classification algorithm comprises a convolutional neural network classification algorithm; when the machine learning training data volume is less than 10000, selecting a machine learning classification algorithm, and when the machine learning training data volume is more than 10000, selecting a deep learning classification algorithm;

the goods combination condition of the historical carpools refers to the goods information of two types of carpool goods of the carpool;

the type name of the accident comprises: confusion, damage and no accident risk occur; the road transport length, comprising: the total road mileage for the common transportation of the two cargos; the road bump condition comprising: detecting the times of road jolting exceeding a threshold value by combining a vibration sensor to obtain the road jolting condition;

the traffic acceleration condition, comprising: and detecting the times that the acceleration exceeds the threshold value in the transportation process through an acceleration sensor to obtain the transportation acceleration condition.

Preferably, the step of excluding the cargo combined with the first cargo in the waiting car-sharing cargo list, screening a candidate car-sharing cargo list according to the total cargo volume, the cargo weight, the transportable residual volume of the truck, the transportable residual weight of the truck, the cargo destination and the cargo start position, predicting and traversing the candidate car-sharing cargo list based on the cargo combination accident prediction model, inputting the cargo information of which the prediction result is accident-risk-free into an accident-risk-free cargo list, and selecting the second cargo based on the accident-risk-free cargo list includes:

judging whether the cargo combination of the historical car pooling information-accident occurrence condition table is in fire or corrosion, and if so, listing the cargo combination as an error combination; excluding from the list of carpooling goods items a good that is in a wrong combination with the first good;

selecting a cargo list in which the cargo total volume, the cargo weight and the cargo remaining weight are smaller than the transportable remaining volume and the transportable remaining weight of the truck, the cargo destination is before the destination of the first cargo, and the cargo starting position is after the cargo starting position of the first cargo, from the waiting car-sharing cargo list to obtain a candidate car-sharing cargo list;

traversing the candidate car pooling cargo list, temporarily combining each cargo information with the first cargo, acquiring the predicted road transportation length under the condition of the cargo combination based on map software, and acquiring the predicted road jolting condition under the condition of the predicted road transportation length based on the historical car pooling information-accident occurrence condition table, inputting the cargo information of the temporary cargo combination, the predicted road transportation length and the predicted road jolting condition into the cargo combination accident prediction model to obtain a prediction result of the type name of the accident, if the prediction result is no accident risk, inputting the traversed cargo information into an accident-free risk cargo list, and acquiring the accident-free risk cargo list after traversing;

the selecting the second good based on the accident-free risk goods list comprises: selecting the goods with the highest freight transportation reward as the second goods based on the accident-risk-free goods list;

or

Traversing the cargo information in the accident-risk-free cargo list based on the accident-risk-free cargo list, and determining a freight price required for reissuing the confused cargo as a cargo reissue price according to the distance between the destination of the cargo and the destination of the first cargo in combination with the internet freight price;

displaying the reissue price of each cargo in the accident-free risk cargo list and the confidence coefficient of the prediction result in the cargo combined accident prediction model at a client, recommending that the confidence coefficient of the prediction result in the cargo combined accident prediction model is higher than a preset first threshold value, and recommending n cargos with the reissue price lower than a preset second threshold value to the truck owner for selection, thereby confirming the second cargo.

Preferably, the car sharing transportation according to the first goods and the second goods, and the arranging of the placing positions of the first goods and the second goods comprises: the first goods are placed on the inner side of the carriage, the second goods are placed on the outer side of the carriage, a goods gap is formed between the first goods and the second goods to distinguish the first goods from the second goods, and the goods gap must be larger than the width of the single goods of the first goods and the second goods.

Preferably, the method includes the steps of counting the cargo combination condition of the historical car pooling, and the road transportation length, the road jolt condition, the transportation acceleration condition and the cargo confusion distance under the cargo combination condition, and establishing a historical car pooling information-confusion distance table, including:

and in each transportation, a carrier checks whether the second goods are mixed by the first goods after the second goods are unloaded, judges whether the second goods belong to the first goods by scanning the two-dimensional codes on the goods, inputs the distance between the second goods and the goods gap by the carrier through a mobile client if the second goods belong to the second goods, and uploads the distance between the second goods and the goods gap and the road transportation length, road bump condition and transportation acceleration condition of the transportation of the time to the historical car pooling information-mixing distance table.

Preferably, the determining, according to the historical car pooling information-confusion distance table, a roi distance grade of goods confusion detection, taking an image after the first goods are loaded as a first image, taking an image after the second goods are unloaded as a second image, calculating all difference matrices and coordinate positions of the first image and the second image, and determining an image difference point set specifically includes:

according to the historical car pooling information-confusion distance table, in combination with the situation of the combination of the cargos to be transported at this time, the road transportation length, the road bumping situation and the transportation acceleration situation, similarity calculation is carried out on the historical data in the historical car pooling information-confusion distance table, so that the combination of the cargos to be transported at one time with the highest similarity to be transported at this time and the distance of the cargos to be confused are obtained, and the roi distance grade of cargo confusion detection is determined;

the distance for mixing the goods which are transported once and have the highest transportation similarity comprises the following steps: the average distance of goods for confusion, and the farthest distance of goods for confusion;

the roi distance rating, comprising: setting 2 roi distance grades, a first roi distance grade and a second roi distance grade; taking the average distance of the mixed cargos as the length of a first roi distance grade, and taking the farthest distance of the mixed cargos as the length of a second roi distance grade; the widths of the first roi distance grade and the second roi distance grade are the widths of the boxcar;

the initial coordinate of the roi distance grade is the coordinate of the cargo gap;

taking the image of the first goods after being loaded as a first image and the image of the second goods after being unloaded as a second image at the same angle, and calculating all difference matrixes and coordinate positions of the first image and the second image by using OpenCV, Python and scimit-image technologies in combination with a Structural Similarity Index (SSIM) so as to determine an image difference point set;

the capturing of the image after the second cargo is unloaded includes: when the second goods destination is reached, judging whether goods exist in the second goods placement area through a target detection algorithm, and if not, judging that the second goods are unloaded;

the second cargo placement area coordinate obtaining method is that in a shot picture, the coordinates of the cargo gap are used as initial coordinates, the total cargo length of the second cargo is used as the length, and the width of a boxcar is used as the width.

Preferably, the obtaining of the total height of the goods of the first goods and the total height of the goods of the second goods according to the image difference point set, calculating a height difference of the goods, and listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule specifically includes:

when the height difference of the goods is less than 20cm or less than the height of a single piece of goods of the first goods or the second goods, traversing each difference matrix and the coordinate position of the difference matrix of the image difference point set, if the coordinate position of a certain difference matrix is in the range of the goods clearance area, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm, and if the difference matrix belongs to the second goods, listing the difference matrix into the confusion reminding list;

the method for acquiring the cargo clearance area comprises the steps of taking the coordinates of the cargo clearance as initial coordinates, taking the length of the cargo clearance as the length and taking the width of a boxcar as the width in a shot picture;

the coordinate acquisition method of the first goods placement area comprises the steps of taking the coordinate of the innermost part of the truck as a starting coordinate, taking the total length of goods of the first goods as the length and taking the width of a truck carriage as the width in a shot picture;

when the height difference of the goods is higher than 20cm or higher than the height of the single goods of the first goods or the second goods, if the first goods is higher, traversing each difference matrix and the coordinate position of the difference matrix of the image set, if the coordinate position of one difference matrix is in the range of the first goods placing area, counting the number of the single goods of the second picture in the range of the first goods placing area through a target detection algorithm, judging whether the number of the single goods of the second picture in the range of the first goods placing area is the same as that of the first picture, if the number of the single goods of the first picture is the same as that of the single goods of the first picture, not reminding, if the number of the single goods of the second picture is different, sending confusion reminding information to a background to remind the second goods of possibly confusing the first goods, and listing all the second goods in a second confusion reminding list; if the difference matrix is in the range of the goods clearance area, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm, and if the difference matrix belongs to the second goods, listing the difference matrix in the confusion reminding list; when the height difference of the goods is higher than 20cm or higher than the height of the single goods of the first goods or the second goods, traversing each difference matrix and the coordinate position of the difference point set of the images if the height difference of the second goods is higher, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm if the coordinate position of a certain difference matrix is in the range of the first goods placing area, judging the roi distance grade of the difference matrix if the difference matrix belongs to the second goods, listing the difference matrix into the confusion reminding list if the difference is in the first roi distance grade and not in the second roi distance grade, and listing the difference matrix into the confusion reminding list if the difference is in the second roi distance grade, wherein the confusion possibility is lower according to the historical car pooling information-confusion distance list, and calculating whether data of cargo combination, road transportation length, road bump condition, transportation acceleration condition and cargo confusion distance similarity exceeding a third threshold exist in the table, and if so, listing the difference matrix in the confusion reminding list.

Preferably, the sending confusion verification reminding information to the carrier according to the quasi-point transportation weight of the goods in the confusion reminding list and by combining the highest number of the goods which can be verified by the carrier includes: and calculating whether the quasi-point transportation weight of the goods in the confusion reminding list is higher than a preset fourth threshold, sending confusion verification reminding information to a carrier for each goods in the confusion reminding list when the quasi-point transportation weight is higher than the preset fourth threshold, presetting the highest number of the goods which can be verified by the carrier in the system when the quasi-point transportation weight is lower than the preset fourth threshold, carrying out similarity calculation on each difference matrix in the confusion reminding list and the single-piece goods image of the second goods, sorting in a descending order according to the similarity, taking the difference matrix with the highest number of the goods which can be verified by the carrier and the highest similarity, and sending the confusion verification reminding information to the carrier.

In another aspect, an embodiment of the present invention further includes a computer apparatus, including a memory and a processor, where the memory is used to store at least one program, and the processor is used to load the at least one program to perform the method of the embodiment.

The beneficial effects of the invention are:

1. the method comprises the steps of obtaining historical car pooling information-accident occurrence condition table, counting the goods combination condition of the historical car pooling, taking the road transportation length, the road jolt condition, the transportation acceleration condition and the type name of an accident under the goods combination condition, taking the historical car pooling information-accident occurrence condition table, the road transportation length and the road jolt condition as characteristic values, taking the type name of the accident as a predicted value, training a goods combination accident prediction model by combining a deep learning or machine learning classification algorithm, and training a goods combination accident prediction model based on the historical car pooling information-accident occurrence condition table, so that goods combinations are screened, the accuracy of goods combination recommendation is improved, and accidents such as follow-up confusion are reduced in a combination stage.

2. Predicting and traversing the candidate car sharing cargo list based on the cargo combined accident prediction model, and inputting cargo information with a prediction result of no accident risk into a no accident risk cargo list; and selecting the second goods based on the accident-risk-free goods list, and recommending the second goods based on the goods transportation reward and the accident risk, so that the loss of error transportation is reduced.

3. According to the historical car pooling information-confusion distance table, the roi distance grade of goods confusion detection is determined, an image obtained after the first goods are loaded is shot to serve as a first image, an image obtained after the second goods are unloaded is shot to serve as a second image, all difference matrixes and coordinate positions of the first image and the second image are calculated, an image difference point set is determined, according to the image difference point set, the total height of the first goods and the total height of the second goods are obtained, the height difference of the goods is calculated, partial image difference points in the image difference point set are listed in a confusion reminding list based on a preset rule, the height difference of the goods is considered, confusion reminding confirmation is carried out, and the accuracy of confusion reminding is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments of the present application will be briefly described below. It is obvious that the drawings described below are only some embodiments of the application, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.

Fig. 1 is a schematic flowchart of an intelligent logistics car-sharing method based on deep learning according to an embodiment of the present application;

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

In the description of the present application, it is to be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like indicate orientations or positional relationships based on those shown in the drawings, and are used merely for convenience of description and for simplicity of description, and do not indicate or imply that the referenced device or element must have a particular orientation, be constructed in a particular orientation, and be operated, and thus should not be considered as limiting the present application. Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.

In this application, the word "exemplary" is used to mean "serving as an example, instance, or illustration. Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. The following description is presented to enable any person skilled in the art to make and use the application. In the following description, details are set forth for the purpose of explanation. It will be apparent to one of ordinary skill in the art that the present application may be practiced without these specific details. In other instances, well-known structures and processes are not set forth in detail in order to avoid obscuring the description of the present application with unnecessary detail. Thus, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

It should be noted that, since the method in the embodiment of the present application is executed in the electronic device, the processing objects of each electronic device all exist in the form of data or information, for example, time, which is substantially time information, and it is understood that, if the size, the number, the position, and the like are mentioned in the following embodiments, all corresponding data exist so as to be processed by the electronic device, and details are not described herein.

s 1: the method comprises the steps that a waiting car sharing cargo list is obtained, cargo information in the waiting car sharing cargo list is determined, truck information is obtained, a truck owner selects first cargo needing to be transported at a client according to the waiting car sharing cargo list, and the transportable residual volume and the transportable residual weight of a truck are determined according to the cargo information of the first cargo and the truck information;

the list of cargoes waiting for sharing is acquired by collecting and collecting in a webpage of a cargo transportation intermediary, the webpage of the cargo transportation intermediary contains information of cargoes required for delivery, and the cargoes generally consist of a pile of single cargoes with the same type, name and specification, such as a single piece, a bag of 50 jin of rice, and 70 bags in total, so that the cargoes are delivered into the list of cargoes waiting for sharing;

the cargo information includes: the method comprises the following steps of (1) cargo name, cargo weight, cargo total volume, cargo type, cargo initial position, cargo destination, cargo total height, cargo total length, cargo total number, cargo transportation reward, cargo quasi-point transportation weight, single-piece cargo image, single-piece cargo width, single-piece cargo height and single-piece cargo weight;

the quasi-point transportation weight of the goods is the weight that the goods must be transported to the destination at the quasi-point, the higher the weight is, the more the quasi-point transportation is needed to reach the destination, and the weight is determined by system presetting, wherein the higher the weight is, the lower the freshness of the goods purchased for transportation insurance is, the higher the weight is, and the weight value is 1-100;

the freight transportation reward is the reward which can be obtained by a truck owner after transporting the freight, and the recording setting is carried out through an online trading platform;

the determining the remaining transportable volume and the remaining transportable weight of the truck according to the cargo information of the first cargo and the truck information includes: subtracting the cargo weight of the cargo information of the first cargo from the total transportable weight of the truck to obtain the remaining transportable weight of the truck; subtracting the total cargo volume of the cargo information of the first cargo from the total cargo volume transportable by the truck to obtain remaining transportable volume transportable by the truck;

s 2: acquiring a historical car pooling information-accident occurrence condition table at a server, counting the goods combination condition of the historical car pooling, and the road transportation length, the road jolting condition, the transportation acceleration condition and the type name of an accident under the goods combination condition, taking the historical car pooling information-accident occurrence condition table, the road transportation length and the road jolting condition as characteristic values, taking the type name of the accident as a predicted value, and training a goods combination accident prediction model by combining a deep learning or machine learning classification algorithm;

the machine learning classification algorithm comprises: a Support Vector Machine (SVM), a naive Bayes classifier, and a decision tree; the deep learning classification algorithm comprises a convolutional neural network classification algorithm; when the machine learning training data volume is less than 10000, selecting a machine learning classification algorithm, and when the machine learning training data volume is more than 10000, selecting a deep learning classification algorithm;

the type name of the accident comprises: confusion, damage and no accident risk occur;

the road transport length comprising: the total road mileage for the common transportation of the two cargoes;

the road bump condition comprises: detecting the times of road jolting exceeding a threshold value by combining a vibration sensor to obtain the road jolting condition;

the traffic acceleration condition, comprising: detecting the times that the acceleration exceeds a threshold value in the transportation process through an acceleration sensor to obtain the transportation acceleration condition;

the historical car pooling information-accident occurrence condition table is stored in a server background, and the goods combination condition of the historical car pooling, the road transportation length, the road bump condition, the transportation acceleration condition and the type name of an accident are recorded under the goods combination condition;

s 3: eliminating the goods which are combined with the first goods in a wrong way from the waiting car sharing goods list, screening a candidate car sharing goods list according to the total volume of the goods, the weight of the goods, the transportable residual volume of the truck, the transportable residual weight of the truck, the destination of the goods and the initial position of the goods, predicting and traversing the candidate car sharing goods list based on the goods combination accident prediction model, and inputting the goods information of which the prediction result is no accident risk into a no accident risk goods list; selecting the second goods based on the accident-free risk goods list, specifically comprising:

judging whether the cargo combination of the historical car pooling information-accident occurrence condition table is in fire or corrosion, and if so, listing the cargo combination as an error combination; excluding from the list of waiting for carpool shipments the shipments made up in the wrong combination with the first shipment;

selecting a cargo list with a cargo starting position behind a cargo starting position of the first cargo in the waiting carpooling cargo list, wherein the cargo total volume, the cargo weight and the cargo remaining weight are less than the transportable remaining volume and the transportable remaining weight of the truck, the cargo ending destination is before the ending destination of the first cargo, and the candidate carpooling cargo list is obtained;

the selecting the second good based on the accident-free risk goods list comprises:

selecting the goods with the highest freight transportation reward as the second goods based on the accident-free risk goods list;

or

displaying the reissue price of each cargo in the accident-free risk cargo list and the confidence coefficient of the prediction result in the cargo combined accident prediction model at a client, recommending that the confidence coefficient of the prediction result in the cargo combined accident prediction model is higher than a preset first threshold value, and selecting n cargos of which the reissue price is lower than a preset second threshold value to the truck owner so as to confirm the second cargo;

the first threshold may be 0.6;

the second threshold may be 100 bins;

s 4: carrying out car sharing transportation according to the first goods and the second goods, and arranging the placing positions of the first goods and the second goods;

arranging the placement positions of the first goods and the second goods, specifically comprising: placing the first goods on the inner side of the carriage, placing the second goods on the outer side of the carriage, and arranging a goods gap between the first goods and the second goods to distinguish the first goods from the second goods, wherein the goods gap must be larger than the width of a single piece of goods of the first goods and the second goods;

s5: counting the goods combination condition of the historical car pooling, and under the goods combination condition, the road transportation length, the road jolt condition, the transportation acceleration condition and the goods confusion distance, and establishing a historical car pooling information-confusion distance table, which specifically comprises the following steps: for each transportation, after the second goods are unloaded, a carrier checks whether the first goods are mixed with the second goods, whether the second goods belong to is judged by scanning the two-dimensional codes on the goods, if the second goods belong to the second goods, the carrier inputs the distance between the second goods and the gap between the second goods through the mobile client, and the distance between the second goods and the gap between the second goods and the historical car pooling information-mixed distance table under the condition of combination of the second goods, road bumping condition and transportation acceleration condition are uploaded;

s6: according to the historical car pooling information-confusion distance table, determining a roi distance grade of goods confusion detection, shooting an image of the first goods after loading the goods as a first image, shooting an image of the second goods after unloading the goods as a second image, calculating all difference matrixes and coordinate positions of the first image and the second image, determining a set of image difference points, acquiring a total height of the goods of the first goods and a total height of the goods of the second goods according to the set of image difference points, calculating a height difference of the goods, and listing partial image difference points in the set of image difference points into a confusion reminding list based on a preset rule, wherein the confusion reminding method specifically comprises the following steps:

according to the historical car pooling information-confusion distance table, in combination with the condition of the combination of the goods transported at this time, the road transportation length, the road bumping condition and the transportation acceleration condition, similarity calculation is carried out on the historical data in the historical car pooling information-confusion distance table to obtain the combination of the goods transported at one time with the highest similarity of the transportation at this time and the distance of the goods in confusion, so that the roi distance grade of goods confusion detection is determined;

the distance for mixing the goods which are transported once and have the highest transportation similarity comprises the following steps: the average distance of the goods to be mixed, and the farthest distance of the goods to be mixed;

the roi distance rating, comprising: setting 2 roi distance grades, a first roi distance grade and a second roi distance grade; taking the average distance of the goods subjected to confusion as the length of a first roi distance grade, and taking the farthest distance of the goods subjected to confusion as the length of a second roi distance grade; the widths of the first roi distance grade and the second roi distance grade are the widths of the boxcar;

the roi distance grade is used for assisting in judging the roi of the target detection algorithm and the image difference point detection;

the second goods placement area coordinate obtaining method is that in a shot picture, the coordinates of the goods clearance are used as initial coordinates, the total length of the second goods is used as the length, and the width of a boxcar is used as the width;

according to the image difference point set, acquiring the total height of the first goods and the total height of the second goods, calculating the height difference of the goods, and listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule, wherein the confusion reminding list specifically comprises the following steps:

if the difference matrix is in the range of the first cargo pile area, no processing is carried out, and because the cargo height difference is small, the cargo confusion under the height difference is not processed, so that the misjudgment is reduced;

when the cargo height difference is higher than 20cm or higher than the single cargo height of the first cargo or the second cargo, if the first good is taller, then goods confusion occurs more in the second good and the goods gaps, each difference matrix and its coordinate position are traversed through the set of image differences points, if the coordinate position of one of the difference matrixes is in the range of the first goods placing area, counting the number of single first cargos in the range of the first cargo placement area by the second picture through a target detection algorithm, judging whether the number of the single first cargos in the range of the first cargo placement area is the same as that of the single first cargos in the range of the first cargo placement area by the first picture, if the first goods are the same, reminding is not carried out, if the first goods are different, confusion reminding information is sent to a background to remind that the second goods possibly confuse the first goods, and all the second goods are listed in a second confusion reminding list; if the difference matrix is in the range of the goods clearance area, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm, and if the difference matrix belongs to the second goods, listing the difference matrix into the confusion reminding list;

when the cargo height difference is higher than 20cm or higher than the height of the single cargo of the first cargo or the second cargo, if the second cargo is higher, then more cargo confusion occurs in the first cargo and the cargo gap, traversing each difference matrix and the coordinate position of the difference point set of the images, if the coordinate position of a certain difference matrix is in the range of the first cargo placing area, calculating whether the difference matrix belongs to the first cargo or the second cargo through an image classification algorithm, if the difference matrix belongs to the second cargo, judging that the difference matrix is in the roi distance grade, if the difference matrix is in the first roi distance grade and not in the second roi distance grade, then the confusion probability is higher, then the difference matrix is listed in the confusion reminding list, if the difference matrix is in the second roi distance grade, then the confusion probability is lower, according to the historical car pooling information-confusion distance table, whether data of the combination of the cargos transported with the current cargos, the road transportation length, the road bump condition, the transportation acceleration condition and the distance similarity of the confusion of the cargos exceeding a third threshold exist in the table is calculated, and if the data exist, the difference matrix is listed in the confusion reminding list;

the third threshold may be 0.6;

s7: according to the quasi-point transportation weight of the goods in the confusion reminding list, the highest number of the goods can be verified by a carrier, confusion verification reminding information is sent to the carrier, and the method specifically comprises the following steps:

calculating whether the quasi-point transportation weight of the goods in the confusion reminding list is higher than a preset fourth threshold, sending confusion verification reminding information to a carrier for each goods in the confusion reminding list when the quasi-point transportation weight is higher than the preset fourth threshold, presetting the highest number of the goods which can be verified by the carrier in a system when the quasi-point transportation weight is lower than the preset fourth threshold, carrying out similarity calculation on each difference matrix in the confusion reminding list and the single-piece goods image of the second goods, sorting in a descending order according to the similarity, taking the difference matrix with the highest number of the goods which can be verified by the carrier and the highest similarity, and sending the confusion verification reminding information to the carrier;

presetting the highest number of cargos, such as 100 cargos, which can be verified by a carrier in the system;

the preset fourth threshold may be 60;

the confusion verification means that a carrier acquires a coordinate position of a difference matrix to be verified, finds a corresponding cargo, and judges whether cargo confusion occurs or not by scanning a two-dimensional code on the cargo;

by calculating whether the quasi-point transportation weight of the goods in the confusion reminding list is higher than a preset fourth threshold value or not, and when the quasi-point transportation weight of the goods in the confusion reminding list is higher than the preset fourth threshold value, confusion verification reminding information is sent to a carrier for each goods in the confusion reminding list, so that the situation that goods which cannot be confused are confused to miss a carrying opportunity can be avoided, for example: certain goods are fruit food, and even if the goods are transported again after being transported by mistake, the goods can be rotten, so that the condition of wrong transportation cannot occur;

when the number of the cargos is lower than a preset fourth threshold value, presetting the highest number of cargos which can be verified by a carrier in the system, calculating the similarity of each difference matrix in the confusion reminding list and the single cargo image of the second cargo, sorting the difference matrixes in a descending order according to the similarity, taking the difference matrix with the highest number of cargos which can be verified by the carrier and the highest similarity, and sending confusion verification reminding information to the carrier, wherein the number of the cargos which can be verified by the carrier is preset in the system is 100, so that the situation that the carrier needs confusion verification in sequence is avoided, and the labor is saved;

and based on the blockchain technology, after the goods transportation is finished, performing source tracing and chain linking, wherein on the blockchain, when all the goods in the confusion prompt list are not confused and verified by a carrier, all the single goods in the first goods are marked as suspected confused goods. Marking all single goods in the second goods in all the second confusion prompt lists as suspected confusion goods so as to facilitate the tracing of the following goods.

In this embodiment, a computer device includes a memory and a processor, the memory is used for storing at least one program, and the processor is used for loading the at least one program to execute the intelligent logistics car-sharing method based on deep learning in embodiments S1-S7, so as to achieve the same technical effects as those described in the embodiments.

It should be noted that, unless otherwise specified, when a feature is referred to as being "fixed" or "connected" to another feature, it may be directly fixed or connected to the other feature or indirectly fixed or connected to the other feature. Furthermore, the descriptions of upper, lower, left, right, etc. used in the present disclosure are only relative to the mutual positional relationship of the constituent parts of the present disclosure in the drawings. As used in this disclosure, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. In addition, unless defined otherwise, all technical and scientific terms used in this example have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in the description of the embodiments herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this embodiment, the term "and/or" includes any combination of one or more of the associated listed items.

It will be understood that, although the terms first, second, third, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element of the same type from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of the present disclosure. The use of any and all examples, or exemplary language ("e.g.," such as "or the like") provided with this embodiment is intended merely to better illuminate embodiments of the invention and does not pose a limitation on the scope of the invention unless otherwise claimed.

It should be recognized that embodiments of the present invention can be realized and implemented in computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer readable memory. The methods may be implemented in a computer program using standard programming techniques, including a non-transitory computer-readable storage medium configured with the computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner, according to the methods and figures described in the detailed description. Each program may be implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language. Furthermore, the program can be run on a programmed application specific integrated circuit for this purpose.

Further, the operations of the processes described in this embodiment can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes described by the present embodiments (or variations and/or combinations thereof) may be performed under the control of one or more computer systems configured with executable instructions and may be implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications) collectively executed on one or more processors, by hardware, or combinations thereof. The computer program includes a plurality of instructions executable by one or more processors.

Further, the method may be implemented in any type of computing platform operatively connected to a suitable interface, including but not limited to a personal computer, mini computer, mainframe, workstation, networked or distributed computing environment, separate or integrated computer platform, or in communication with a charged particle tool or other imaging device, and the like. Aspects of the invention may be embodied in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optically read and/or write storage medium, RAM, ROM, or the like, such that it may be read by a programmable computer, which when read by the storage medium or device, is operative to configure and operate the computer to perform the procedures described herein. Further, the machine-readable code, or portions thereof, may be transmitted over a wired or wireless network. The invention described in this embodiment includes these and other different types of non-transitory computer-readable storage media when such media include instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. The invention also includes the computer itself when programmed according to the methods and techniques described herein.

A computer program can be applied to input data to perform the functions described in the present embodiment to convert the input data to generate output data that is stored to a non-volatile memory. The output information may also be applied to one or more output devices, such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including particular visual depictions of physical and tangible objects produced on a display.

The present invention is not limited to the above embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention as long as the technical effects of the present invention are achieved by the same means. The invention is capable of other modifications and variations in its technical solution and/or its implementation, within the scope of protection of the invention.

Claims

1. An intelligent logistics car-sharing method based on deep learning is characterized by comprising the following steps:

determining the roi distance grade of goods confusion detection according to the historical car pooling information-confusion distance table, shooting an image of the first goods after being loaded as a first image, shooting an image of the second goods after being unloaded as a second image, calculating all difference matrixes and coordinate positions of the first image and the second image, determining an image difference point set, obtaining the total height of the first goods and the total height of the second goods according to the image difference point set, calculating a goods height difference, and listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule;

2. The intelligent logistics car-sharing method based on deep learning of claim 1, wherein: the acquiring a list of cargoes waiting for carpooling, determining cargo information in the list of cargoes waiting for carpooling, acquiring truck information, selecting a first cargo to be transported by a truck owner at a client according to the list of cargoes waiting for carpooling, and determining a transportable residual volume and a transportable residual weight of the truck according to the cargo information of the first cargo and the truck information, comprises:

3. The intelligent logistics car-sharing method based on deep learning of claim 1 is characterized in that: the method comprises the steps that a server obtains historical car pooling information-accident occurrence condition table, statistics is carried out on the goods combination condition of historical car pooling, the road transportation length, the road jolt condition, the transportation acceleration condition and the type name of an accident are carried out under the goods combination condition, the historical car pooling information-accident occurrence condition table, the road transportation length and the road jolt condition are used as characteristic values, the type name of the accident is used as a predicted value, and a goods combination accident prediction model is trained by combining a deep learning or machine learning classification algorithm, and the method comprises the following steps:

the road transport length, comprising: the total road mileage for the common transportation of the two cargoes;

the road bump condition comprising: detecting the times of road jolting exceeding a threshold value by combining a vibration sensor to obtain the road jolting condition;

4. The intelligent logistics car-sharing method based on deep learning of claim 1 is characterized in that: excluding the goods which are combined with the first goods in a wrong way from the waiting car sharing goods list, screening a candidate car sharing goods list according to the total goods volume, the goods weight, the transportable residual volume of the truck, the transportable residual weight of the truck, the goods destination and the goods starting position, predicting and traversing the candidate car sharing goods list based on the goods combination accident prediction model, inputting the goods information of which the prediction result is accident-free risk into an accident-free risk goods list, and selecting the second goods based on the accident-free risk goods list, wherein the steps comprise:

traversing the candidate car pooling cargo list, temporarily combining each cargo information with the first cargo, acquiring the predicted road transport length under the condition of the cargo combination based on map software, inputting the predicted road transport length and the predicted road jolt condition into the cargo combination accident prediction model based on the historical car pooling information-accident occurrence condition table, acquiring the predicted road jolt condition under the condition of the predicted road transport length, inputting the cargo information of the temporary cargo combination, the predicted road transport length and the predicted road jolt condition into the cargo combination accident prediction model, acquiring a prediction result of the type name of an accident, inputting the traversed cargo information into an accident-risk-free cargo list if the prediction result is accident-risk-free, and acquiring the accident-risk-free cargo list after traversing;

selecting the goods with the highest freight transportation reward as the second goods based on the accident-risk-free goods list;

or

Traversing the cargo information in the accident-risk-free cargo list based on the accident-risk-free cargo list, and determining a freight price required by reissuing after the cargos are mixed up as a cargo reissue price according to the distance between the destination of the cargo and the destination of the first cargo and the internet freight price;

5. The intelligent logistics car-sharing method based on deep learning of claim 1 is characterized in that: the car sharing transportation is carried out according to the first goods and the second goods, and the placement positions of the first goods and the second goods are arranged, and the car sharing transportation system comprises: the first goods are placed on the inner side of the carriage, the second goods are placed on the outer side of the carriage, a goods gap is formed between the first goods and the second goods to distinguish the first goods from the second goods, and the goods gap must be larger than the width of the single goods of the first goods and the second goods.

6. The intelligent logistics car-sharing method based on deep learning of claim 1 is characterized in that: counting the goods combination condition of the historical carpooling, and under the goods combination condition, the road transportation length, the road jolt condition, the transportation acceleration condition and the goods confusion distance, and establishing a historical carpooling information-confusion distance table, wherein the table comprises the following steps:

7. The intelligent logistics car-sharing method based on deep learning of claim 1 is characterized in that: determining the roi distance grade of goods confusion detection according to the historical car pooling information-confusion distance table, shooting an image after the first goods are loaded as a first image, shooting an image after the second goods are unloaded as a second image, calculating all difference matrixes and coordinate positions of the first image and the second image, and determining an image difference point set, wherein the method specifically comprises the following steps of:

8. The intelligent logistics car-sharing method based on deep learning of claim 1, wherein: the step of obtaining the total height of the goods of the first goods and the total height of the goods of the second goods according to the image difference point set, calculating the height difference of the goods, and listing partial image difference points in the image difference point set into a confusion reminding list based on a preset rule, specifically comprises the steps of:

when the height difference of the goods is less than 20cm or less than the height of the single goods of the first goods or the second goods, traversing each difference matrix and the coordinate position of each difference matrix of the image difference point set, if the coordinate position of a certain difference matrix is in the goods clearance area range, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm, and if the difference matrix belongs to the second goods, listing the difference matrix into the confusion reminding list;

the method for acquiring the coordinates of the first goods placement area comprises the steps of taking the coordinates of the innermost part of the truck as initial coordinates in a shot picture, taking the total length of the goods of the first goods as the length and taking the width of a boxcar as the width;

when the height difference of the goods is higher than 20cm or higher than the height of the single goods of the first goods or the second goods, if the first goods is higher, traversing each difference matrix and the coordinate position of the difference matrix of the image set, if the coordinate position of one difference matrix is in the range of the first goods placing area, counting the number of the single goods of the second picture in the range of the first goods placing area through a target detection algorithm, judging whether the number of the single goods of the second picture in the range of the first goods placing area is the same as that of the first picture, if so, not reminding, if not, sending confusion reminding information to a background to remind the second goods that the first goods are possibly confused, and listing all the second goods in a second confusion reminding list; if the difference matrix is in the range of the goods clearance area, calculating whether the difference matrix belongs to the first goods or the second goods through an image classification algorithm, and if the difference matrix belongs to the second goods, listing the difference matrix into the confusion reminding list;

when the cargo height difference is higher than 20cm or higher than the height of a single piece of cargo of the first cargo or the second cargo, if the second cargo is higher, traversing each difference matrix and the coordinate position of the difference matrix in the image difference point set, if the coordinate position of a certain difference matrix is in the range of the first cargo placement area, calculating whether the difference matrix belongs to the first cargo or the second cargo through an image classification algorithm, if the difference matrix belongs to the second cargo, judging the roi distance grade of the difference matrix, if the difference matrix is in the first roi distance grade and not in the second roi distance grade, the confusion probability is higher, listing the difference matrix into the confusion reminding list, if the difference matrix is in the second roi distance grade, the confusion probability is lower, and according to the historical car pooling information-confusion distance list, and calculating whether data of cargo combination, road transportation length, road bump condition, transportation acceleration condition and cargo confusion distance similarity exceeding a third threshold exist in the table, and if so, listing the difference matrix in the confusion reminding list.

9. The intelligent logistics car-sharing method based on deep learning of claim 1, wherein: the sending confusion verification reminding information to the carrier according to the quasi-point transportation weight of the goods in the confusion reminding list and by combining the highest number of the goods which can be verified by the carrier comprises the following steps:

and calculating whether the quasi-point transportation weight of the goods in the confusion reminding list is higher than a preset fourth threshold, sending confusion verification reminding information to a carrier for each goods in the confusion reminding list when the quasi-point transportation weight is higher than the preset fourth threshold, presetting the highest number of the goods which can be verified by the carrier in the system when the quasi-point transportation weight is lower than the preset fourth threshold, carrying out similarity calculation on each difference matrix in the confusion reminding list and the single-piece goods image of the second goods, sorting in a descending order according to the similarity, taking the difference matrix with the highest number of the goods which can be verified by the carrier and the highest similarity, and sending the confusion verification reminding information to the carrier.

10. A computer apparatus comprising a memory for storing at least one program and a processor for loading the at least one program to perform the method of any one of claims 1-9.