CN114463697A - Loading rate calculation method based on image recognition - Google Patents
Loading rate calculation method based on image recognition Download PDFInfo
- Publication number
- CN114463697A CN114463697A CN202210084272.9A CN202210084272A CN114463697A CN 114463697 A CN114463697 A CN 114463697A CN 202210084272 A CN202210084272 A CN 202210084272A CN 114463697 A CN114463697 A CN 114463697A
- Authority
- CN
- China
- Prior art keywords
- carriage
- detection frame
- tail end
- loaded
- goods
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a loading rate calculation method based on image recognition, which comprises the following steps: (1) collecting a plurality of videos with scenes of goods loaded by trucks of different models, and performing frame dismantling processing to obtain sample pictures as training sample pictures; (2) the obtained labeling information of the training sample picture is stored as a labeling file; (3) inputting training sample pictures and marking files into a yolo-v3 model for training to obtain a goods and carriage tail end contour recognition model; (4) acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into the cargo and carriage tail end contour recognition model for recognition, and outputting a loaded cargo detection frame and a carriage tail end contour detection frame in the field images; (5) calculating the actual distance L between the loaded goods detection frame and the carriage tail end contour detection frame; (6) and acquiring the actual carriage length K of each truck, and calculating the truck loading rate. The invention improves the identification efficiency of the loading rate of the truck.
Description
Technical Field
The invention relates to the technical field of particle board surface defect collection, in particular to a loading rate calculation method based on image recognition.
Background
With the gradual popularization of on-line consumption, the domestic logistics transportation industry is developed vigorously, wherein highway freight becomes the most main logistics transportation mode by virtue of the advantages of high transportation speed, high reliability, high maneuverability, good economic benefit and the like. In the process of carrying goods by a truck, the loading rate of a carriage directly influences the scheduling of the number of cars and the unit goods carrying cost, and is a key index for determining the profit level of the whole logistics industry. At present, most logistics units are based on a manual evaluation method, the statistical accuracy and the reliability are difficult to guarantee, meanwhile, real-time statistical data cannot be obtained, and the overall scheduling operation cannot be achieved, so that the effects of cost reduction and efficiency improvement are achieved. The long-acting and efficient management of the transportation process can be realized only by utilizing scientific and technological means. The difference exists between the calculated loading rate and the actual loading rate of the boxcar, and the reliability cannot be ensured; in the prior art, the accuracy and the real-time performance are difficult to control, and the whole operation effect is not obviously improved.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a load factor calculation method based on image recognition, which is used for distance measurement sampling in a carriage through image recognition, calculating the average distance of the remaining space by combining multiple adoption points and finally calculating the carriage load factor.
A loading rate calculation method based on image recognition comprises the following steps:
(1) collecting a plurality of videos with different types of goods loaded by trucks in scenes, and performing frame splitting processing to obtain sample pictures as training sample pictures; the plurality of videos comprise the area range of goods loaded by the truck in the truck and the outline of the tail end of the carriage;
(2) respectively labeling the region range of goods loaded by the truck and the contour of the tail end of the carriage of the truck in the training sample picture to obtain labeling information of the training sample picture and storing the labeling information as a labeling file;
(3) inputting training sample pictures and marking files into a yolo-v3 model for training to obtain a contour recognition model of the tail end of the cargo and the carriage; the goods and carriage tail end contour recognition model is used for inputting pictures and outputting a loaded goods detection frame and a carriage tail end contour detection frame of the pictures;
(4) acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into the cargo and carriage tail end contour recognition model for recognition, and outputting a loaded cargo detection frame and a carriage tail end contour detection frame in the field images; the field image comprises the area range of goods loaded by the truck in the truck and the contour of the tail end of the carriage;
(5) calculating the actual distance L between the loaded goods detection frame and the carriage tail end contour detection frame through the loaded goods detection frame and the carriage tail end contour detection frame in the step (4);
(6) and (5) acquiring the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame and the carriage tail end outline detection frame in the step (5), and calculating the truck loading rate to be (K-L)/K x 100%.
In particular, the markup document is a document in the format of XML.
In particular, the yolo-v3 model includes the Darknet-53 network.
Specifically, the specific method for calculating the distance L between the loaded cargo detection frame and the car tail end contour detection frame in the step (5) is as follows:
s1, respectively calculating a center point coordinate of a left side edge as a left side center point coordinate of a loaded goods detection frame and a center point coordinate of a right side edge as a left side center point coordinate of the loaded goods detection frame according to the loaded goods detection frame;
s2, respectively calculating the coordinate of the center point of the left side edge as the coordinate of the center point of the contour detection frame of the tail end of the left carriage and the coordinate of the center point of the right side edge as the coordinate of the center point of the contour detection frame of the tail end of the right carriage according to the contour detection frame of the tail end of the carriage;
s3, calculating a pixel distance L1 from the center point coordinate of the contour detection frame at the tail end of the left carriage to the center point coordinate of the left side of the loaded cargo detection frame, and calculating a pixel distance L2 from the center point coordinate of the contour detection frame at the tail end of the right carriage to the center point coordinate of the right side of the loaded cargo detection frame;
s4, matching a relation model of the pixel distance and the actual distance by combining the pixel distance which is acquired in advance and corresponds to the center point of the outline of the tail end of the carriage from the center point of the cargo frame and the actual distance from the center point of the edge of the actually measured cargo to the center point of the outline of the tail end of the carriage, and realizing the conversion from the pixel distance to the actual distance;
s5, converting the pixel distance L1 and the pixel distance L2 in the step S3 into a corresponding actual distance L1 'and a corresponding actual distance L2' according to the fitted relation model in the step S4, and calculating an average distance L of the actual distance L1 'and the actual distance L2' to be used as the distance L between the loaded cargo detection frame and the carriage tail end contour detection frame.
A system for image recognition based load factor calculation using the above method, comprising:
the acquisition module is used for acquiring a plurality of videos of goods loaded by trucks with different types of scenes; the acquisition module is also used for performing frame splitting processing on the plurality of videos to obtain sample pictures as training sample pictures; the multiple videos can be collected to the goods loading area of the goods wagon in the wagon and the contour position of the tail end of the carriage;
the marking module is used for marking the region range of goods loaded on the truck and the contour of the tail end of the carriage of the truck in the training sample picture respectively to obtain marking information of the training sample picture and storing the marking information as a marking file;
the training module is used for inputting training sample pictures and marking files into a yolo-v3 model for training to obtain a contour recognition model of the tail end of the goods and the carriage; the goods and carriage tail end contour recognition model is used for inputting pictures and outputting a loaded goods detection frame and a carriage tail end contour detection frame of the pictures;
the identification module is used for acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into the cargo and carriage tail end contour identification model for identification, and outputting a loaded cargo detection frame and a carriage tail end contour detection frame in the field images; the field image comprises the area range of goods loaded by the truck in the truck and the contour of the tail end of the carriage;
the distance conversion module is used for calculating the actual distance L between the loaded goods detection frame and the carriage tail end contour detection frame through the loaded goods detection frame and the carriage tail end contour detection frame acquired by the identification module;
the loading rate calculation module is used for obtaining the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame and the carriage tail end outline detection frame of the distance conversion module, and calculating the loading rate of the truck to be (K-L)/K x 100%;
the acquisition module, the marking module, the training module, the recognition module, the distance conversion module and the loading rate calculation module are sequentially connected.
Particularly, the acquisition module is an IP camera arranged at the top end of the tail of the carriage of the truck or a camera fixed on a rudder port and used for shooting a truck loading scene video.
Specifically, the actual distance L between the loaded cargo detection frame and the carriage tail end contour detection frame obtained by the distance conversion module is transmitted to the loading rate calculation module by adopting json format information.
Particularly, the acquisition module further comprises a plurality of infrared ranging sensors; the infrared ranging sensors are arranged on the edges of the front end and the tail end of the boxcar in the depth direction.
In particular, the markup document is a document in the format of XML.
In particular, the yolo-v3 model includes the Darknet-53 network.
Compared with the prior art, the invention has the beneficial effects that:
the method comprises the steps of firstly obtaining a shot image of a truck loading scene, then labeling the image to obtain a labeled file, then inputting the image and the labeled file into a preset YOLO-V3 model of an extraction network by taking Darknet-53 as a characteristic, and training the image and the labeled file to obtain a cargo area identification model and a carriage tail end contour identification model. By taking the design of Darknet-19 and Resnet in yolov2 as a reference and fusing the FPNs in the front to form a Darknet-53 convolutional layer, a rapid target detection algorithm is formed, and therefore the identification efficiency of the truck loading rate is improved. In the application of the freight car loading rate statistics, the freight car loading rate condition can be accurately identified and calculated, the freight car loading rate condition is timely recorded in the system, the unqualified freight cars are judged and reminded, the waste of logistics resources is avoided to a great extent through the integral scheduling operation, and the logistics cost is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a method of an embodiment of the invention.
Fig. 2 is a schematic diagram of the internal structure of the embodiment of the invention.
FIG. 3 is a schematic block diagram of a system of an embodiment of the invention.
In the figure, 1, left side contour; 2. a right side profile; 3. a loaded cargo detection frame; 4. detecting the coordinate of the center point of a frame for detecting the tail end contour of the left carriage; 5. detecting the coordinate of the center point of a frame for detecting the tail end contour of the right carriage; 6. coordinates of the center point of the left side of the loaded cargo detection frame; 7. coordinates of the right center point of the loaded cargo detection frame; 10. a left side outline frame; 11. a right side outline frame; 12. the truck loads the goods.
Detailed Description
The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, and the scope of the present invention will be more clearly and clearly defined.
It is to be understood that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," "outer," and the like are used in a generic and descriptive sense only and not for purposes of limitation, the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," "outer," and the like are used in the generic and descriptive sense only and not for purposes of limitation, as the term is used in the generic and descriptive sense, and not for purposes of limitation, unless otherwise specified or implied, and the specific reference to a device or element is intended to be a reference to a particular element, structure, or component. Furthermore, the terms "first," "second," "third," and the like are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.
Furthermore, the terms "horizontal", "vertical", "overhang" and the like do not imply that the components are required to be absolutely horizontal or overhang, but may be slightly inclined. For example, "horizontal" merely means that the direction is more horizontal than "vertical" and does not mean that the structure must be perfectly horizontal, but may be slightly inclined.
In the description of the present invention, it should also be noted that, unless otherwise explicitly specified or limited, the terms "disposed," "mounted," "connected," and "connected" are to be construed broadly and may, for example, be fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
As shown in fig. 1 and fig. 2, a method for calculating a loading rate based on image recognition according to this embodiment includes the following steps:
(1) collecting a plurality of videos with scenes of goods 12 loaded on trucks of different models, and performing frame splitting processing to obtain sample pictures as training sample pictures; the plurality of videos includes the extent of the area of the truck in the truck loaded with cargo 12 and the contour of the rear end of the car. In the embodiment of the present invention, a specific method for acquiring multiple videos of trucks with different types of scenes for loading cargos 12 is as follows: the method comprises the steps of firstly shooting a video of a freight car loading scene through an IP camera (needing a wide-angle camera) or a rudder port fixed camera arranged at the top end of the tail part of a carriage, meanwhile shooting a freight car loading area 12 and the contour position of the tail end of the carriage, taking frames at intervals to generate pictures, and taking the pictures as training sample pictures. The tail end contour of the carriage at the tail end contour position comprises a left side contour 1 and a right side contour 2.
(2) Labeling the region range of the freight car loaded with the goods 12 and the contour of the tail end of the carriage of the freight car in the training sample picture respectively to obtain labeling information of the training sample picture and storing the labeling information as a labeling file; the markup file is a file with the format of XML. And marking the training sample pictures generated above by adopting a manual labeling mode, and framing the goods stacking area and the tail end outlines of the carriages at the left side and the right side by using a labelImg tool in a rectangular frame mode to generate a resolvable XML marking file.
(3) And inputting the training sample pictures and the labeling files into a yolo-v3 model for training to obtain a contour recognition model of the tail end of the goods and the carriage. The goods and carriage tail end contour recognition model is used for inputting pictures and outputting loaded goods detection frames and carriage tail end contour detection frames of the pictures. The yolo-v3 model includes the Darknet-53 network. The yolo-v3 model is an end-to-end target detection model, and the basic idea of the algorithm is as follows: firstly, extracting features from input features through a feature extraction network to obtain feature map output with a specific size. The input image is divided into 13 × 13 grid cells, and then if the center coordinate of an object in the real frame falls in a grid cell, the object is predicted by the grid cell. There are a fixed number of bounding boxes per object, three in the yolo-v3 model, and logistic regression is used to determine the regression box to predict. The yolo-v3 model of the present embodiment uses a 53-layer convolutional network, which is a superposition of residual units. In balance between classification accuracy and efficiency. The embodiment of the invention adopts the Darknet-53 model to perform better than ResNet-101, ResNet-152 and Darknet-19.
(4) The method comprises the steps of acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into a cargo and carriage tail end contour recognition model for recognition, and outputting a loaded cargo detection frame 3 and a carriage tail end contour detection frame in the field images, wherein the carriage tail end contour detection frame comprises a left side contour frame 10 and a right side contour frame 11, so that three rectangular frames are obtained. The live image includes the extent of the area of the wagon carrying cargo 12 and the contour of the rear end of the wagon.
(5) Calculating the actual distance L between the loaded goods detection frame 3 and the carriage tail end contour detection frame through the loaded goods detection frame 3 and the carriage tail end contour detection frame in the step (4), wherein the specific method comprises the following steps:
s1, respectively calculating a central point coordinate of a left side as a central point coordinate 6 of the left side of the loaded goods detection frame and a central point coordinate of a right side as a central point coordinate 6 of the left side of the loaded goods detection frame according to the loaded goods detection frame 3;
s2, respectively calculating a center point coordinate of a left side edge as a center point coordinate 4 of a left side carriage tail end contour detection frame and a center point coordinate of a right side edge as a center point coordinate 5 of a right side carriage tail end contour detection frame according to the carriage tail end contour detection frame;
s3, calculating a pixel distance L1 from a center point coordinate 4 of the contour detection frame at the tail end of the left carriage to a center point coordinate 6 at the left side of the loaded cargo detection frame, and calculating a pixel distance L2 from a center point coordinate 5 of the contour detection frame at the tail end of the right carriage to a center point coordinate 7 at the right side of the loaded cargo detection frame;
s4, matching a relation model of the pixel distance and the actual distance by combining the pixel distance which is acquired in advance and corresponds to the center point of the outline of the tail end of the carriage from the center point of the cargo frame and the actual distance from the center point of the edge of the actually measured cargo to the center point of the outline of the tail end of the carriage, and realizing the conversion from the pixel distance to the actual distance;
and S5, converting the pixel distance L1 and the pixel distance L2 in the step S3 into a corresponding actual distance L1 'and a corresponding actual distance L2' according to the fitted relation model in the step S4, and calculating an average distance L of the actual distance L1 'and the actual distance L2' as the distance L between the loaded cargo detection frame 3 and the contour detection frame at the tail end of the carriage. The result output by the embodiment of the invention is a very small json format.
(6) And (5) acquiring the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame 3 and the carriage tail end contour detection frame in the step (5): when a vehicle is registered, basic information such as carriage length and the like can be recorded into the system borne by the embodiment of the invention, different types of boxcars are different in length, the vehicle can be locked according to the unique ID returned by the IP camera, and then the carriage length information K of the current vehicle can be obtained by performing correlation matching. And finally, calculating the freight car loading rate to be (K-L)/K x 100% according to the actual carriage length K of each freight car recorded by the system.
As shown in fig. 3, a system for performing image recognition-based load factor calculation using the above method according to an embodiment of the present invention includes:
and the acquisition module is used for acquiring a plurality of videos with different types of goods 12 loaded on trucks. The acquisition module is also used for performing frame splitting processing on the plurality of videos to obtain sample pictures as training sample pictures; the multiple videos can be collected to the area of the freight car for loading the cargos 12 and the contour position of the tail end of the carriage; the acquisition module is an IP camera arranged at the top end of the tail of the carriage of the truck or a camera fixed on a rudder port and used for shooting a truck loading scene video.
The marking module is used for marking the region range of the goods 12 loaded on the truck in the training sample picture and the contour of the tail end of the carriage of the truck respectively to obtain marking information of the training sample picture and storing the marking information as a marking file;
and the training module is used for inputting the training sample pictures and the labeling files into the yolo-v3 model for training to obtain a contour recognition model of the tail end of the goods and the carriage. The goods and carriage tail end contour recognition model is used for inputting pictures and outputting a loaded goods detection frame and a carriage tail end contour detection frame of the pictures;
the identification module is used for acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into a cargo and carriage tail end contour identification model for identification, and outputting a loaded cargo detection frame 3 and a carriage tail end contour detection frame in the field images;
the distance conversion module is used for calculating the actual distance L between the loaded goods detection frame 3 and the carriage tail end contour detection frame through the loaded goods detection frame 3 and the carriage tail end contour detection frame acquired by the identification module; the actual distance L between the loaded goods detection frame 3 and the carriage tail end contour detection frame obtained by the distance conversion module is transmitted to the loading rate calculation module by adopting json format information and can be transmitted to the loading rate calculation module through a 4G/5G network.
The loading rate calculation module is used for obtaining the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame 3 of the distance conversion module and the carriage tail end outline detection frame, and calculating the loading rate of the truck to be (K-L)/K x 100%;
the acquisition module, the marking module, the training module, the recognition module, the distance conversion module and the loading rate calculation module are sequentially connected.
The acquisition module of this embodiment still includes a plurality of infrared ranging sensor. The infrared ranging sensor is arranged at the edge of the front end and the tail end of the boxcar in the depth direction and used for correcting the data for ranging and sampling in the boxcar.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, various changes or modifications may be made by the patentees within the scope of the appended claims, and within the scope of the invention, as long as they do not exceed the scope of the invention described in the claims. The principles and embodiments of the present invention are explained herein using specific examples, which are presented only to assist in understanding the method and its core concepts of the present invention. It should be noted that there are no specific structures but a few objective structures due to the limited character expressions, and that those skilled in the art may make various improvements, decorations or changes without departing from the principle of the invention or may combine the above technical features in a suitable manner; such modifications, variations, combinations, or adaptations of the invention using its spirit and scope, as defined by the claims, may be directed to other uses and embodiments.
Claims (10)
1. A loading rate calculation method based on image recognition is characterized in that: the method comprises the following steps:
(1) collecting a plurality of videos with different types of goods loaded by trucks in scenes, and performing frame splitting processing to obtain sample pictures as training sample pictures; the plurality of videos comprise the area range of goods loaded by the truck in the truck and the outline of the tail end of the carriage;
(2) respectively labeling the region range of goods loaded by the truck and the contour of the tail end of the carriage of the truck in the training sample picture to obtain labeling information of the training sample picture and storing the labeling information as a labeling file;
(3) inputting training sample pictures and marking files into a yolo-v3 model for training to obtain a contour recognition model of the tail end of the cargo and the carriage; the goods and carriage tail end contour recognition model is used for inputting pictures and outputting a loaded goods detection frame and a carriage tail end contour detection frame of the pictures;
(4) acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into the cargo and carriage tail end contour recognition model for recognition, and outputting a loaded cargo detection frame and a carriage tail end contour detection frame in the field images; the field image comprises the area range of goods loaded by the truck in the truck and the contour of the tail end of the carriage;
(5) calculating the actual distance L between the loaded goods detection frame and the carriage tail end contour detection frame through the loaded goods detection frame and the carriage tail end contour detection frame in the step (4);
(6) and (5) acquiring the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame and the carriage tail end outline detection frame in the step (5), and calculating the truck loading rate to be (K-L)/K x 100%.
2. The image recognition-based load factor calculation method according to claim 1, wherein: the labeled file is a file with an XML format.
3. The image recognition-based load factor calculation method according to claim 1, wherein: the yolo-v3 model includes the Darknet-53 network.
4. The image recognition-based load factor calculation method according to claim 1, wherein: the specific method for calculating the distance L between the loaded goods detection frame and the carriage tail end contour detection frame in the step (5) is as follows:
s1, respectively calculating a center point coordinate of a left side edge as a left side center point coordinate of a loaded goods detection frame and a center point coordinate of a right side edge as a left side center point coordinate of the loaded goods detection frame according to the loaded goods detection frame;
s2, respectively calculating the coordinate of the center point of the left side edge as the coordinate of the center point of the contour detection frame of the tail end of the left carriage and the coordinate of the center point of the right side edge as the coordinate of the center point of the contour detection frame of the tail end of the right carriage according to the contour detection frame of the tail end of the carriage;
s3, calculating a pixel distance L1 from the center point coordinate of the contour detection frame at the tail end of the left carriage to the center point coordinate of the left side of the loaded cargo detection frame, and calculating a pixel distance L2 from the center point coordinate of the contour detection frame at the tail end of the right carriage to the center point coordinate of the right side of the loaded cargo detection frame;
s4, matching a relation model of the pixel distance and the actual distance by combining the pixel distance which is acquired in advance and corresponds to the center point of the outline of the tail end of the carriage from the center point of the cargo frame and the actual distance from the center point of the edge of the actually measured cargo to the center point of the outline of the tail end of the carriage, and realizing the conversion from the pixel distance to the actual distance;
s5, converting the pixel distance L1 and the pixel distance L2 in the step S3 into a corresponding actual distance L1 'and a corresponding actual distance L2' according to the fitted relation model in the step S4, and calculating an average distance L of the actual distance L1 'and the actual distance L2' to be used as the distance L between the loaded cargo detection frame and the carriage tail end contour detection frame.
5. A system for image recognition based load rate calculation using the method of any of claims 1-4, comprising:
the acquisition module is used for acquiring a plurality of videos of goods loaded by trucks with different types of scenes; the acquisition module is also used for frame-splitting the plurality of videos into sample pictures serving as training sample pictures; the multiple videos can be collected to the goods loading area of the goods wagon in the wagon and the contour position of the tail end of the carriage;
the marking module is used for marking the region range of goods loaded on the truck and the contour of the tail end of the carriage of the truck in the training sample picture respectively to obtain marking information of the training sample picture and storing the marking information as a marking file;
the training module is used for inputting training sample pictures and marking files into a yolo-v3 model for training to obtain a contour recognition model of the tail end of the goods and the carriage; the goods and carriage tail end contour recognition model is used for inputting pictures and outputting a loaded goods detection frame and a carriage tail end contour detection frame of the pictures;
the identification module is used for acquiring a plurality of field images in a truck loading scene captured in real time, inputting the field images into the cargo and carriage tail end contour identification model for identification, and outputting a loaded cargo detection frame and a carriage tail end contour detection frame in the field images; the field image comprises the area range of goods loaded by the truck in the truck and the contour of the tail end of the carriage;
the distance conversion module is used for calculating the actual distance L between the loaded goods detection frame and the carriage tail end contour detection frame through the loaded goods detection frame and the carriage tail end contour detection frame acquired by the identification module;
the loading rate calculation module is used for obtaining the actual carriage length K of each truck through the actual distance L between the loaded goods detection frame and the carriage tail end outline detection frame of the distance conversion module, and calculating the loading rate of the truck to be (K-L)/K x 100%;
the acquisition module, the marking module, the training module, the recognition module, the distance conversion module and the loading rate calculation module are sequentially connected.
6. The system for image recognition-based load factor calculation according to claim 5, wherein: the acquisition module is an IP camera arranged at the top end of the tail of the carriage of the truck or a camera fixed on a rudder port and used for shooting a truck loading scene video.
7. The system for image recognition-based load factor calculation according to claim 5, wherein: and the actual distance L between the loaded cargo detection frame and the carriage tail end contour detection frame obtained by the distance conversion module is transmitted to the loading rate calculation module by adopting json format information.
8. The system for image recognition-based load factor calculation according to claim 5, wherein: the acquisition module also comprises a plurality of infrared ranging sensors; the infrared ranging sensors are arranged on the edges of the front end and the tail end of the boxcar in the depth direction.
9. The system for image recognition-based load factor calculation according to claim 5, wherein: the labeled file is a file with an XML format.
10. The system for image recognition-based load factor calculation according to claim 5, wherein: the yolo-v3 model includes the Darknet-53 network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210084272.9A CN114463697A (en) | 2022-01-25 | 2022-01-25 | Loading rate calculation method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210084272.9A CN114463697A (en) | 2022-01-25 | 2022-01-25 | Loading rate calculation method based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114463697A true CN114463697A (en) | 2022-05-10 |
Family
ID=81411282
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210084272.9A Pending CN114463697A (en) | 2022-01-25 | 2022-01-25 | Loading rate calculation method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114463697A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115759514A (en) * | 2022-11-18 | 2023-03-07 | 广东豆加壹科技有限公司 | Cold chain distribution vehicle scheduling management method and device |
| CN117078138A (en) * | 2023-10-12 | 2023-11-17 | 北京汇通天下物联科技有限公司 | Truck loading information processing method and server |
| CN117893891A (en) * | 2024-03-11 | 2024-04-16 | 深圳安行致远技术有限公司 | Space utilization rate measuring and calculating method and system based on machine learning |
-
2022
- 2022-01-25 CN CN202210084272.9A patent/CN114463697A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115759514A (en) * | 2022-11-18 | 2023-03-07 | 广东豆加壹科技有限公司 | Cold chain distribution vehicle scheduling management method and device |
| CN117078138A (en) * | 2023-10-12 | 2023-11-17 | 北京汇通天下物联科技有限公司 | Truck loading information processing method and server |
| CN117078138B (en) * | 2023-10-12 | 2024-02-06 | 北京汇通天下物联科技有限公司 | Truck loading information processing method and server |
| CN117893891A (en) * | 2024-03-11 | 2024-04-16 | 深圳安行致远技术有限公司 | Space utilization rate measuring and calculating method and system based on machine learning |
| CN117893891B (en) * | 2024-03-11 | 2024-05-17 | 深圳安行致远技术有限公司 | Space utilization rate measuring and calculating method and system based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114463697A (en) | Loading rate calculation method based on image recognition | |
| CN112233097B (en) | Road scene other vehicle detection system and method based on space-time domain multi-dimensional fusion | |
| CN107591005B (en) | Parking area management method, server and system combining dynamic and static detection | |
| CN111899515B (en) | Vehicle detection system based on wisdom road edge calculates gateway | |
| CN111080612B (en) | Truck bearing damage detection method | |
| CN104092988A (en) | Method, device and system for managing passenger flow in public place | |
| CN111814739B (en) | Method, device, equipment and storage medium for detecting express package volume | |
| CN111626277A (en) | Vehicle tracking method and device based on over-station inter-modulation index analysis | |
| CN111723854A (en) | Method and device for detecting traffic jam of highway and readable storage medium | |
| CN111079748A (en) | Method for detecting oil throwing fault of rolling bearing of railway wagon | |
| CN114022537B (en) | Method for analyzing loading rate and unbalanced loading rate of vehicle in dynamic weighing area | |
| CN113788051A (en) | Train on-station running state monitoring and analyzing system | |
| CN113762144B (en) | Deep learning-based black smoke vehicle detection method | |
| CN111966857A (en) | Method and system for detecting modified vehicle | |
| CN114066997B (en) | Offset detection method based on binocular vision and symmetry | |
| CN105095851A (en) | Steel coil position identification method | |
| CN117274967A (en) | Multi-mode fusion license plate recognition algorithm based on convolutional neural network | |
| CN117037085A (en) | Vehicle identification and quantity statistics monitoring method based on improved YOLOv5 | |
| CN116520351A (en) | Train state monitoring method, system, storage medium and terminal | |
| CN114973156B (en) | Night muck car detection method based on knowledge distillation | |
| CN115359306A (en) | Intelligent identification method and system for high-definition images of railway freight inspection | |
| CN113705549B (en) | Road rescue work node determination method and device and related equipment | |
| CN115857040A (en) | Dynamic visual detection device and method for foreign matters on locomotive roof | |
| CN101577052B (en) | Device and method for detecting vehicles by overlooking | |
| CN115343719A (en) | Truck severe overload detection method based on infrared and laser radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |