Vehicle detection method and system, scrap steel intelligent grading system

Abstract

The present invention provides a vehicle detection method and system, and a scrap steel intelligent grading system, wherein the vehicle detection method includes: collecting an image of the unloading area; confirming the vehicle to be detected in the unloading area image; detecting the overall vehicle range and the compartment range of the vehicle to be detected at each moment; correcting the compartment range at each moment based on the overall vehicle range at each moment to obtain a corrected compartment range; calculating the predicted compartment range at the next moment based on the corrected compartment range at each moment; and obtaining the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range. The present application first corrects the compartment range by the overall vehicle range, and then predicts the compartment range at the next moment, and obtains the actual compartment range at the next moment by combining the corrected compartment range and the predicted compartment range at the next moment, thereby ensuring the correctness of the compartment range image acquisition and avoiding the influence of full load, proximity to scrap steel piles, and spray devices on the compartment range detection.

Description

Technical Field

[0001] The present invention relates to the technical field of scrap steel detection, and in particular to a vehicle detection method and system, and an intelligent scrap steel grading system. Background Art

[0002] During scrap unloading, the intelligent scrap grading system focuses on and photographs the scrap within the detected vehicle compartment. By identifying the scrap in the photograph, it automatically determines the scrap grade. Failure to accurately detect the vehicle compartment during vehicle compartment detection will result in an incorrect focus range, leading to an inaccurate scrap grade.

[0003] Therefore, how to propose a solution that can accurately identify the carriage range in real time has become an urgent problem to be solved. Summary of the Invention

[0004] To solve the above technical problems, a first aspect of the present invention provides a vehicle detection method.

[0005] A second aspect of the present invention further provides a vehicle detection system.

[0006] The third aspect of the present invention further provides a readable storage medium.

[0007] The fourth aspect of the present invention further proposes an intelligent scrap steel grading system.

[0008] In view of this, the first aspect of the present invention proposes a vehicle detection method, including: collecting an image of the unloading area; confirming the vehicle to be detected in the image of the unloading area; detecting the overall vehicle range and the compartment range of the vehicle to be detected at each moment; correcting the compartment range at each moment based on the overall vehicle range at each moment to obtain a corrected compartment range; calculating the predicted compartment range at the next moment based on the corrected compartment range at each moment; and obtaining the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range.

[0009] According to the vehicle detection method provided by the present invention, by collecting images of the unloading area, the vehicle to be detected is confirmed based on the image, which is convenient for tracking the vehicle with detection, and then the vehicle to be detected at each moment in the collected images of the unloading area is detected, and the overall range of the vehicle and the range of the carriage are detected. Since the image of the carriage range is collected when the scrap steel is graded, and when the carriage is fully loaded and close to the scrap steel pile, the carriage range is easily misjudged, resulting in the collected image not completely covering the carriage, while the vehicle as a whole will not be affected by being fully loaded and close to the scrap steel pile. After all, the characteristics of the front of the vehicle are more obvious than those of the carriage carrying scrap steel, which is convenient for identification. Therefore, the car range can be corrected through the overall range of the vehicle to obtain the corrected car range, thereby avoiding the influence of car range misjudgment on car image acquisition; further, the corrected car range at each moment is predicted, and the car range at the next moment is predicted as the predicted car range, and then the corrected car range and the predicted car range can be combined to obtain the actual car range, wherein the actual car range is the actual car range at the next moment obtained based on the corrected car range at the next moment and the predicted car range at the next moment, and then the scrap steel image can be collected according to the actual car range, and the scrap steel grade can be judged, thereby improving the accuracy of car range image acquisition. The present application first corrects the car range through the overall range of the vehicle, and then predicts the car range at the next moment. The corrected car range at the next moment and the predicted car range at the next moment are combined to obtain the actual car range at the next moment, thereby ensuring the correctness of the car range image acquisition, avoiding the impact of full load and nearby scrap steel piles on the car range detection, and reducing the occurrence of inaccurate car range detection in a short period of time. At the same time, by predicting the car range at the next moment, the impact of the spray device on the car range detection in a short period of time is greatly reduced.

[0010] In addition, the vehicle detection method in the above technical solution provided by the present invention may also have the following additional technical features:

[0011] In the above technical solution, the step of confirming the vehicle to be detected in the unloading area image specifically includes: when the vehicle enters the image acquisition range of the unloading area, the license plate of the vehicle is identified as the unique code of the vehicle to confirm the vehicle to be detected.

[0012] In this technical solution, the license plate of the vehicle within the image acquisition range entering the unloading area is identified as the vehicle's code, thereby confirming the vehicle to be detected, and the data can be matched with the vehicle in subsequent calculations, avoiding data confusion caused by the presence of multiple vehicles within the image acquisition range.

[0013] In the above technical solution, the vehicle detection method further includes: when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, recording the relative position of the vehicle's entire range and the vehicle compartment range.

[0014] In this technical solution, when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative positions of the vehicle's overall range and the car body range are recorded. Since the detection results of the vehicle's overall range and the car body range are very accurate when the vehicle is far away from the scrap steel pile, the relative positions of the two are recorded at this time. Since the relative position of the vehicle as a whole and the car body is fixed, the car body range can be corrected according to the relative position of the car body and the vehicle as a whole when the vehicle approaches the scrap steel pile, thereby avoiding false detection during car body range detection.

[0015] In the above technical solution, the step of correcting the car range at each moment based on the overall range of the vehicle at each moment to obtain the corrected car range specifically includes: correcting the car range of the unloading area image at each moment based on the overall range and relative position of the vehicle to obtain the corrected car range at each moment.

[0016] In this technical solution, the cabin range can be corrected through the overall range of the vehicle and the relative position between the vehicle as a whole and the cabin. Since the relative position between the vehicle as a whole and the cabin is fixed, this correction method has high accuracy and can avoid false detection during cabin range detection.

[0017] In the above technical solution, the vehicle detection method further includes: performing calculations to predict the vehicle compartment range based on Kalman filtering.

[0018] In this technical solution, a Kalman filter is used to predict the predicted compartment range at the next moment based on the corrected compartment range at the previous moment. Combined with the corrected compartment range at the next moment, the actual compartment range at that narrow moment is obtained. This Kalman filter is then used to predict the compartment range at the next moment, combining the corrected compartment range at the previous moment with the corrected compartment range at the previous moment. The combination of the two yields the actual compartment range. During the unloading process, high dust levels can temporarily affect the accuracy of compartment target detection, requiring the spray device to be activated. Kalman filtering can predict the compartment range at the next moment, significantly reducing inaccurate compartment range detection that may occur over short periods of time.

[0019] Kalman filtering is an algorithm that uses a linear system state equation and observation data from the system's input and output to optimally estimate the system's state. Because the observation data includes the effects of noise and interference in the system, optimal estimation can also be considered a filtering process.

[0020] In the above technical solution, the vehicle detection method further includes: grading scrap steel based on an image of the actual vehicle compartment range.

[0021] In this technical solution, scrap steel grading is performed by collecting images of the actual carriage range, thereby avoiding incomplete scrap steel image acquisition caused by errors in carriage range detection and improving the accuracy of scrap steel grading.

[0022] In the above technical solution, the vehicle detection method further includes: when the vehicle enters the image acquisition range of the unloading area, tracking the vehicle until the vehicle completes unloading.

[0023] In this technical solution, the vehicle is tracked within the image acquisition range from the time the vehicle enters the unloading area until the vehicle finishes unloading, and the overall range of the vehicle and the range of the carriage are detected in real time, thereby improving the accuracy of the carriage range detection.

[0024] In the above technical solution, the vehicle detection method further includes: starting unloading when the vehicle is stopped; and ending unloading when there is no scrap steel in the vehicle compartment.

[0025] This technical solution uses visual recognition to identify the vehicle's driving status and the amount of scrap remaining in the vehicle, controlling the start and stop of unloading based on the vehicle's status and the amount of scrap remaining. This eliminates the need for drivers to swipe their cards to determine the start time of unloading, and for remote quality inspectors to manually determine the end time of grading. This improves the efficiency of intelligent scrap grading and enables an unmanned, intelligent workflow.

[0026] The second aspect of the present invention provides a vehicle detection system, comprising: an acquisition module for acquiring images of the unloading area; a confirmation module for confirming the vehicle to be detected in the unloading area image; a detection module for detecting the overall vehicle range and the compartment range of the vehicle to be detected at each moment; a correction module for correcting the compartment range at each moment based on the overall vehicle range at each moment to obtain a corrected compartment range; a calculation module for calculating the predicted compartment range at the next moment based on the corrected compartment range at each moment, and obtaining the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range.

[0027] The vehicle detection system provided by the technical solution of the present invention includes an acquisition module, a confirmation module, a detection module, a correction module and a calculation module. Among them, the acquisition module is used to acquire the unloading area image; the confirmation module is used to confirm the vehicle to be detected in the unloading area image; the detection module is used to detect the overall vehicle range and the compartment range of the vehicle to be detected at each moment; the correction module is used to correct the compartment range at each moment based on the overall vehicle range at each moment to obtain the corrected compartment range; the calculation module is used to calculate the predicted compartment range at the next moment based on the corrected compartment range at each moment, and obtain the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range. At the same time, the vehicle detection system provided by the technical solution of the present invention, because it is used to implement the steps of the vehicle detection method provided by the first aspect of the present invention, the vehicle detection system has all the technical effects of the vehicle detection method, which will not be repeated here.

[0028] In the above technical solution, the confirmation module is specifically used to: when a vehicle enters the image acquisition range of the unloading area, identify the vehicle's license plate as the vehicle's unique code to confirm the vehicle to be detected.

[0029] In this technical solution, the license plate of the vehicle within the image acquisition range entering the unloading area is identified as the vehicle's code, thereby confirming the vehicle to be detected, and the data can be matched with the vehicle in subsequent calculations, avoiding data confusion caused by the presence of multiple vehicles within the image acquisition range.

[0030] In the above technical solution, the vehicle detection system further includes: a recording module for recording the relative positions of the entire vehicle range and the vehicle compartment range when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold.

[0031] In this technical solution, when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative positions of the vehicle's overall range and the car body range are recorded. Since the detection results of the vehicle's overall range and the car body range are very accurate when the vehicle is far away from the scrap steel pile, the relative positions of the two are recorded at this time. Since the relative position of the vehicle as a whole and the car body is fixed, the car body range can be corrected according to the relative position of the car body and the vehicle as a whole when the vehicle approaches the scrap steel pile, thereby avoiding false detection during car body range detection.

[0032] In the above technical solution, the correction module is specifically used to correct the compartment range of the unloading area image at each moment based on the overall range and relative position of the vehicle to obtain the corrected compartment range at each moment.

[0033] In this technical solution, the cabin range can be corrected through the overall range of the vehicle and the relative position between the vehicle as a whole and the cabin. Since the relative position between the vehicle as a whole and the cabin is fixed, this correction method has high accuracy and can avoid false detection during cabin range detection.

[0034] In the above technical solution, the prediction module is specifically used to: calculate the predicted vehicle compartment range based on Kalman filtering.

[0035] In this technical solution, a Kalman filter is used to predict the predicted compartment range at the next moment based on the corrected compartment range at the previous moment. Combined with the corrected compartment range at the next moment, the actual compartment range at that narrow moment is obtained. This Kalman filter is then used to predict the compartment range at the next moment, combining the corrected compartment range at the previous moment with the corrected compartment range at the previous moment. The combination of the two yields the actual compartment range. During the unloading process, high dust levels can temporarily affect the accuracy of compartment target detection, requiring the spray device to be activated. Kalman filtering can predict the compartment range at the next moment, significantly reducing inaccurate compartment range detection that may occur over short periods of time.

[0036] Kalman filtering is an algorithm that uses a linear system state equation and observation data from the system's input and output to optimally estimate the system's state. Because the observation data includes the effects of noise and interference in the system, optimal estimation can also be considered a filtering process.

[0037] In the above technical solution, the vehicle detection system further includes: a scrap steel grading module, which is used to grade the scrap steel based on the image of the actual vehicle compartment range.

[0038] In this technical solution, scrap steel grading is performed by collecting images of the actual carriage range, thereby avoiding incomplete scrap steel image acquisition caused by errors in carriage range detection and improving the accuracy of scrap steel grading.

[0039] In the above technical solution, the vehicle detection system further includes: a tracking module, which is used to track the vehicle when the vehicle enters the image acquisition range of the unloading area until the vehicle finishes unloading.

[0040] In this technical solution, the vehicle is tracked within the image acquisition range from the time the vehicle enters the unloading area until the vehicle finishes unloading, and the overall range of the vehicle and the range of the carriage are detected in real time, thereby improving the accuracy of the carriage range detection.

[0041] In the above technical solution, the vehicle detection system further includes: starting unloading when the vehicle is in a stopped state; and ending unloading when there is no scrap steel in the vehicle compartment.

[0042] This technical solution uses visual recognition to identify the vehicle's driving status and the amount of scrap remaining in the vehicle, controlling the start and stop of unloading based on the vehicle's status and the amount of scrap remaining. This eliminates the need for drivers to swipe their cards to determine the start time of unloading, and for remote quality inspectors to manually determine the end time of grading. This improves the efficiency of intelligent scrap grading and enables an unmanned, intelligent workflow.

[0043] A third aspect of the present invention provides a readable storage medium having programs and / or instructions stored thereon, which implement the steps of the vehicle detection method in any of the above technical solutions when the programs and / or instructions are executed by a processor.

[0044] The readable storage medium provided by the technical solution of the present invention has all the beneficial technical effects of the above-mentioned vehicle detection method, which will not be repeated here, because the programs and / or instructions stored thereon can implement the steps of the vehicle detection method in any of the above-mentioned technical solutions when executed by the processor.

[0045] A fourth aspect of the present invention provides an intelligent scrap steel grading system, comprising the vehicle detection system as in the above technical solution; or the readable storage medium as in the above technical solution.

[0046] The intelligent scrap steel grading system provided by the technical solution of the present invention includes the vehicle detection system or the readable storage medium as described in the technical solution above. Therefore, the intelligent scrap steel grading system has all the technical effects of the vehicle detection system or the readable storage medium, and no further details are given here.

[0047] Additional aspects and advantages of the invention will become apparent from the description which follows, or may be learned by practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments with reference to the accompanying drawings, in which:

[0049] Figure 1 is a flow chart of a vehicle detection method according to an embodiment of the present invention;

[0050] Figure 2 is a block diagram of a vehicle detection system according to an embodiment of the present invention;

[0051] Figure 3 is a schematic diagram of the relationship between steps according to another embodiment of the present invention.

[0052] in, Figure 2 The corresponding relationship between the reference numerals and component names is as follows:

[0053] 200 vehicle detection system, 202 acquisition module, 204 confirmation module, 206 detection module, 208 correction module, 210 calculation module. DETAILED DESCRIPTION

[0054] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0056] Refer to the following Figures 1 to 3 The vehicle detection method and system and the scrap steel intelligent grading system in some embodiments of the present invention are described.

[0057] The first embodiment of the present invention provides a vehicle detection method, such as Figure 1 Shown, including:

[0058] S102, collecting images of the unloading area;

[0059] S104, confirming the vehicle to be detected in the unloading area image;

[0060] S106, detecting the entire vehicle range and the compartment range of the vehicle to be detected at each moment;

[0061] S108, correcting the cabin range at each moment based on the overall vehicle range at each moment to obtain a corrected cabin range;

[0062] S110, calculating the predicted car range at the next moment based on the corrected car range at each moment;

[0063] S112, obtaining the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range.

[0064] According to the vehicle detection method provided by this embodiment, by collecting images of the unloading area, the vehicle to be detected is confirmed based on the image, which is convenient for tracking the vehicle with detection, and then the vehicle to be detected at each moment in the collected images of the unloading area is detected, and the overall range of the vehicle and the range of the carriage are detected. Since the image of the carriage range is collected when the scrap steel is graded, and when the carriage is fully loaded and close to the scrap steel pile, it is easy to misjudge the range of the carriage, resulting in the collected image not completely covering the carriage, while the vehicle as a whole will not be affected by being fully loaded and close to the scrap steel pile. After all, the characteristics of the front of the vehicle are more obvious than the characteristics of the carriage carrying scrap steel, which is convenient for identification. Therefore, the car range can be corrected through the overall range of the vehicle to obtain the corrected car range, thereby avoiding the influence of car range misjudgment on car image acquisition; further, the corrected car range at each moment is predicted, and the car range at the next moment is predicted as the predicted car range, and then the corrected car range and the predicted car range can be combined to obtain the actual car range, wherein the actual car range is the actual car range at the next moment obtained based on the corrected car range at the next moment and the predicted car range at the next moment, and then the scrap steel image can be collected according to the actual car range, and the scrap steel grade can be judged, thereby improving the accuracy of car range image acquisition. The present application first corrects the car range through the overall range of the vehicle, and then predicts the car range at the next moment. The corrected car range at the next moment and the predicted car range at the next moment are combined to obtain the actual car range at the next moment, thereby ensuring the correctness of the car range image acquisition, avoiding the impact of full load and nearby scrap steel piles on the car range detection, and reducing the occurrence of inaccurate car range detection in a short period of time. At the same time, by predicting the car range at the next moment, the impact of the spray device on the car range detection in a short period of time is greatly reduced.

[0065] In the above embodiment, the step of confirming the vehicle to be detected in the unloading area image specifically includes: when the vehicle enters the image acquisition range of the unloading area, the license plate of the vehicle is identified as the unique code of the vehicle to confirm the vehicle to be detected.

[0066] In this embodiment, the license plate of the vehicle within the image acquisition range entering the unloading area is identified as the vehicle's code, thereby confirming the vehicle to be detected, and the data can be matched with the vehicle in subsequent calculations, thereby avoiding data confusion caused by the presence of multiple vehicles within the image acquisition range.

[0067] In the above embodiment, the vehicle detection method further includes: when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, recording the relative position of the entire vehicle range and the vehicle compartment range.

[0068] In this embodiment, when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative positions of the vehicle's overall range and the car body range are recorded. Since the detection results of the vehicle's overall range and the car body range are very accurate when the vehicle is far away from the scrap steel pile, the relative positions of the two are recorded at this time. Since the relative position of the vehicle as a whole and the car body is fixed, the car body range can be corrected according to the relative position of the car body and the vehicle as a whole when the vehicle approaches the scrap steel pile, thereby avoiding false detection during car body range detection.

[0069] In the above embodiment, the step of correcting the car range at each moment based on the overall range of the vehicle at each moment to obtain the corrected car range specifically includes: correcting the car range at each moment of the unloading area image based on the overall range and relative position of the vehicle to obtain the corrected car range at each moment.

[0070] In this embodiment, the vehicle compartment range can be corrected through the overall range of the vehicle and the relative position between the vehicle as a whole and the vehicle compartment. Since the relative position between the vehicle as a whole and the vehicle compartment is fixed, this correction method has high accuracy and can avoid false detection during vehicle compartment range detection.

[0071] In the above embodiment, the vehicle detection method further includes: performing calculation to predict the vehicle compartment range based on Kalman filtering.

[0072] In this embodiment, a Kalman filter is used to predict the predicted car range at the next moment based on the corrected car range at the previous moment. Combined with the corrected car range at the next moment, the actual car range at the narrow moment is obtained. The Kalman filter is used to predict the car range at the next moment in combination with the corrected car range at the previous moment, and the actual car range is obtained by combining the two. During the unloading process, high dust levels may necessitate the activation of the spray device, which can temporarily affect the accuracy of car target detection. Kalman filtering can predict the car range at the next moment, significantly reducing inaccurate car range detection that may occur over a short period of time.

[0073] Kalman filtering is an algorithm that uses a linear system state equation and observation data from the system's input and output to optimally estimate the system's state. Because the observation data includes the effects of noise and interference in the system, optimal estimation can also be considered a filtering process.

[0074] In the above embodiment, the vehicle detection method further includes: grading the scrap steel based on the image of the actual vehicle compartment range.

[0075] In this embodiment, scrap steel grading is performed by collecting images of the actual carriage range, thereby avoiding incomplete scrap steel image acquisition due to errors in carriage range detection and improving the accuracy of scrap steel grading.

[0076] In the above embodiment, the vehicle detection method further includes: when the vehicle enters the image acquisition range of the unloading area, tracking the vehicle until the vehicle completes unloading.

[0077] In this embodiment, the vehicle is tracked within the image acquisition range from the time the vehicle enters the unloading area until the vehicle finishes unloading, and the overall range of the vehicle and the range of the carriage are detected in real time, thereby improving the accuracy of the carriage range detection.

[0078] In the above embodiment, the vehicle detection method further includes: starting unloading when the vehicle is stopped; and ending unloading when there is no scrap steel in the vehicle compartment.

[0079] In this embodiment, visual recognition of the vehicle's driving status and the amount of scrap remaining in the vehicle compartment is used to control the start and stop of unloading based on the vehicle's status and the amount of scrap remaining. This eliminates the need for the driver to swipe a card to determine the start time of unloading, and for remote quality inspectors to manually determine the end time of grading. This improves the efficiency of intelligent scrap grading and implements an unmanned intelligent workflow.

[0080] A second aspect of the present invention provides a vehicle detection system 200, such as Figure 2 As shown, it includes: an acquisition module 202 for acquiring images of the unloading area; a confirmation module 204 for confirming the vehicle to be detected in the image of the unloading area; a detection module 206 for detecting the overall vehicle range and the compartment range of the vehicle to be detected at each moment; a correction module 208 for correcting the compartment range at each moment based on the overall vehicle range at each moment to obtain a corrected compartment range; a calculation module 210 for calculating the predicted compartment range at the next moment based on the corrected compartment range at each moment, and obtaining the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range.

[0081] The vehicle detection system provided according to an embodiment of the present invention includes an acquisition module 202, a confirmation module 204, a detection module 206, a correction module 208, and a calculation module 210. Among them, the acquisition module 202 is used to acquire the unloading area image; the confirmation module 204 is used to confirm the vehicle to be detected in the unloading area image; the detection module 206 is used to detect the overall vehicle range and the compartment range of the vehicle to be detected at each moment; the correction module 208 is used to correct the compartment range at each moment based on the overall vehicle range at each moment to obtain the corrected compartment range; the calculation module 210 is used to calculate the predicted compartment range at the next moment based on the corrected compartment range at each moment, and obtain the actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range. At the same time, the vehicle detection system provided according to an embodiment of the present invention, since it is used to implement the steps of the vehicle detection method provided by the first aspect of the present invention, the vehicle detection system has all the technical effects of the vehicle detection method, which will not be repeated here.

[0082] In the above embodiment, the confirmation module is specifically used to: when a vehicle enters the image acquisition range of the unloading area, identify the vehicle's license plate as the vehicle's unique code to confirm the vehicle to be detected.

[0083] In this embodiment, the license plate of the vehicle within the image acquisition range entering the unloading area is identified as the vehicle's code, thereby confirming the vehicle to be detected, and the data can be matched with the vehicle in subsequent calculations, thereby avoiding data confusion caused by the presence of multiple vehicles within the image acquisition range.

[0084] In the above embodiment, the vehicle detection system further includes: a recording module for recording the relative positions of the entire vehicle range and the vehicle compartment range when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold.

[0085] In this embodiment, when the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative positions of the vehicle's overall range and the car body range are recorded. Since the detection results of the vehicle's overall range and the car body range are very accurate when the vehicle is far away from the scrap steel pile, the relative positions of the two are recorded at this time. Since the relative position of the vehicle as a whole and the car body is fixed, the car body range can be corrected according to the relative position of the car body and the vehicle as a whole when the vehicle approaches the scrap steel pile, thereby avoiding false detection during car body range detection.

[0086] In the above embodiment, the correction module is specifically used to correct the compartment range of the unloading area image at each moment based on the overall range and relative position of the vehicle to obtain the corrected compartment range at each moment.

[0087] In this embodiment, the vehicle compartment range can be corrected through the overall range of the vehicle and the relative position between the vehicle as a whole and the vehicle compartment. Since the relative position between the vehicle as a whole and the vehicle compartment is fixed, this correction method has high accuracy and can avoid false detection during vehicle compartment range detection.

[0088] In the above embodiment, the prediction module is specifically used to: perform calculation to predict the vehicle compartment range based on Kalman filtering.

[0089] In this embodiment, a Kalman filter is used to predict the predicted car range at the next moment based on the corrected car range at the previous moment. Combined with the corrected car range at the next moment, the actual car range at the narrow moment is obtained. The Kalman filter is used to predict the car range at the next moment in combination with the corrected car range at the previous moment, and the actual car range is obtained by combining the two. During the unloading process, high dust levels may necessitate the activation of the spray device, which can temporarily affect the accuracy of car target detection. Kalman filtering can predict the car range at the next moment, significantly reducing inaccurate car range detection that may occur over a short period of time.

[0090] Kalman filtering is an algorithm that uses a linear system state equation and observation data from the system's input and output to optimally estimate the system's state. Because the observation data includes the effects of noise and interference in the system, optimal estimation can also be considered a filtering process.

[0091] In the above embodiment, the vehicle detection system further includes: a scrap steel grading module, which is used to perform scrap steel grading based on the image of the actual vehicle compartment range.

[0092] In this embodiment, scrap steel grading is performed by collecting images of the actual carriage range, thereby avoiding incomplete scrap steel image acquisition due to errors in carriage range detection and improving the accuracy of scrap steel grading.

[0093] In the above embodiment, the vehicle detection system further includes: a tracking module, which is used to track the vehicle when the vehicle enters the image acquisition range of the unloading area until the vehicle completes unloading.

[0094] In this embodiment, the vehicle is tracked within the image acquisition range from the time the vehicle enters the unloading area until the vehicle finishes unloading, and the overall range of the vehicle and the range of the carriage are detected in real time, thereby improving the accuracy of the carriage range detection.

[0095] In the above embodiment, the vehicle detection system further includes: starting unloading when the vehicle is stopped; and ending unloading when there is no scrap steel in the vehicle compartment.

[0096] In this embodiment, visual recognition of the vehicle's driving status and the amount of scrap remaining in the vehicle compartment is used to control the start and stop of unloading based on the vehicle's status and the amount of scrap remaining. This eliminates the need for the driver to swipe a card to determine the start time of unloading, and for remote quality inspectors to manually determine the end time of grading. This improves the efficiency of intelligent scrap grading and implements an unmanned intelligent workflow.

[0097] A third aspect of the present invention provides a readable storage medium having programs and / or instructions stored thereon, which implement the steps of the vehicle detection method in any of the above embodiments when the programs and / or instructions are executed by a processor.

[0098] The readable storage medium provided according to an embodiment of the present invention has all the beneficial technical effects of the above-mentioned vehicle detection method, which will not be repeated here, because the programs and / or instructions stored thereon can implement the steps of the vehicle detection method in any of the above-mentioned embodiments when executed by the processor.

[0099] A fourth aspect of the present invention provides an intelligent scrap steel grading system, comprising a vehicle detection system as in the above embodiment; or a readable storage medium as in the above embodiment.

[0100] The intelligent scrap steel grading system provided in accordance with an embodiment of the present invention includes the vehicle detection system or the readable storage medium described in the above-mentioned embodiment. Therefore, the intelligent scrap steel grading system possesses all the technical effects of the vehicle detection system or the readable storage medium, and thus will not be further described here.

[0101] The vehicle detection method provided by this application is further described below in conjunction with another specific embodiment.

[0102] This embodiment performs scrap steel vehicle detection based on target tracking, which is divided into three steps: target detection, detection result optimization, and Kalman filtering.

[0103] (1) Target detection.

[0104] By inspecting the vehicle and the carriage, the vehicle refers to the complete vehicle including the front of the vehicle, and the carriage refers to the part inside the railing of the carriage.

[0105] The vehicle reverses into the unloading point. Starting from the moment the vehicle enters the camera range, target detection is performed on the vehicle and carriage in each frame of the video. The detection results will be optimized and then sent to the Kalman filter.

[0106] Simultaneous vehicle and cabin detection is designed to improve detection accuracy. In practice, vehicle detection is more effective due to the distinct features of the vehicle's head. Detection of the cabin area can be optimized based on the vehicle detection results.

[0107] (2) Optimization of test results.

[0108] The main purpose of optimizing detection results is to correct the size of the detection results and prevent occasional erroneous detection results. When the vehicle is far from the scrap pile, the vehicle detection and cabin detection results are very accurate. The vehicle detection size, cabin size, and their relative position are recorded at this time. When the vehicle is fully loaded and near the scrap pile, the cabin detection is prone to false detection at the rear. When the vehicle is completely within the camera's range, the relative position of the vehicle detection range and cabin detection range is relatively fixed, so the cabin detection range can be corrected based on the vehicle detection range.

[0109] (3) Kalman filter.

[0110] The Kalman filter is an algorithm that uses a linear system state equation and observation data from the system's input and output to optimally estimate the system state. In this paper, the Kalman filter uses the results of the previous target detection (at time T) to predict the estimated target state at the next time (at time T+1). This is combined with the optimized results of the target detection at time T+1 to determine the actual target state at time T+1.

[0111] The relationship between the above three steps is as follows Figure 3 shown.

[0112] The implementation process of the present invention is briefly described below.

[0113] (1) License plate recognition.

[0114] When a vehicle enters the camera range, the license plate is identified and the license plate number serves as the unique code for the vehicle.

[0115] (2) Vehicle tracking.

[0116] The vehicle reverses into the unloading point. Starting from the moment the vehicle enters the camera range, target detection is performed on the vehicle and carriage in each frame of the video. The detection results are optimized and then sent to the Kalman filter to establish a tracker. The Kalman filter state is updated based on the optimized detection results of the next frame and the results predicted by the Kalman filter.

[0117] (3) Determine the time to start unloading.

[0118] When the vehicle enters the unloading area and the state of the vehicle becomes a stopped state, unloading begins.

[0119] (4) Determine the time to end unloading.

[0120] When the vehicle is in the unloading area, the grabber stops grabbing scrap steel, and there is no scrap steel in the carriage, and the unloading is completed.

[0121] (5) Stop tracking

[0122] After unloading is completed, vehicle tracking is stopped. Reason: After unloading is completed, tracking the vehicle is meaningless.

[0123] The present invention performs target detection on both the vehicle and the carriage separately, determining the relative position of the vehicle and carriage. During vehicle reversing, the relative position between the two remains largely unchanged, resulting in a more distinct vehicle signature and high vehicle detection accuracy. The carriage range can be accurately determined based on the vehicle range and the relative position between the two. This processing method accurately determines the carriage range during the entire intelligent scrap grading process, including reversing and unloading. This solves the problem of inaccurate carriage range detection when a fully loaded scrap vehicle is near a scrap pile. By optimizing the detection results, each target detection result is optimized, preserving the relative position of the vehicle and carriage and correcting the carriage's accurate position. A Kalman filter updates the state, predicting the next target state, and combining the optimized detection results at time T+1 to update the actual target state at time T+1. This makes it easy to determine when the vehicle is reversing, when unloading begins, and when unloading ends, achieving truly unmanned intelligent control. Furthermore, during unloading, the spray device can sometimes temporarily affect the accuracy of carriage target detection. Kalman filtering can effectively reduce these inaccuracies in carriage detection.

[0124] In this specification, the term "plurality" refers to two or more, unless otherwise specified. Terms such as "mounted," "connected," "connected," and "fixed" should be interpreted broadly. For example, "connected" can refer to a fixed connection, a removable connection, or an integral connection; "connected" can refer to a direct connection or an indirect connection through an intermediary. Those skilled in the art will understand the specific meanings of these terms in this disclosure based on specific circumstances.

[0125] Throughout this specification, terms such as "one embodiment" or "some embodiments" mean that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0126] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A vehicle detection method for a scrap steel intelligent grading system, characterized in that: include: Collect images of the unloading area; confirming the vehicle to be detected in the unloading area image; Detecting the entire vehicle range and the vehicle compartment range of the vehicle to be detected at each moment; Correcting the cabin range at each moment based on the vehicle overall range at each moment to obtain a corrected cabin range; Calculating the predicted carriage range at the next moment based on the corrected carriage range at each moment; Obtaining an actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range; The step of confirming the vehicle to be detected in the unloading area image specifically includes: When a vehicle enters the image acquisition range of the unloading area, the license plate of the vehicle is identified as the unique code of the vehicle to confirm the vehicle to be detected; The vehicle detection method further includes: When the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative position of the entire vehicle range and the carriage range is recorded; The step of correcting the vehicle compartment range at each moment based on the vehicle overall range at each moment to obtain a corrected vehicle compartment range specifically includes: Correcting the carriage range of the unloading area image at each moment based on the overall range of the vehicle and the relative position to obtain the corrected carriage range at each moment; The vehicle detection method further includes: Calculating the predicted compartment range based on Kalman filtering; Establish a tracker and update the Kalman filter state based on the optimized detection results of the next frame and the results of the Kalman filter prediction.

2. The vehicle detection method according to claim 1, characterized in that: Also includes: Scrap steel grading is performed based on the image of the actual carriage range.

3. The vehicle detection method according to claim 1, characterized in that: Also includes: When a vehicle enters the image acquisition range of the unloading area, the vehicle is tracked until the vehicle finishes unloading.

4. A vehicle detection system, characterized in that: include: An acquisition module, used for acquiring images of the unloading area; A confirmation module, used for confirming the vehicle to be detected in the unloading area image; A detection module, configured to detect the entire vehicle range and the vehicle compartment range of the vehicle to be detected at each moment; a correction module, configured to correct the cabin range at each moment based on the overall vehicle range at each moment to obtain a corrected cabin range; a calculation module, configured to calculate a predicted compartment range at a next moment based on the corrected compartment range at each moment, and obtain an actual compartment range of the vehicle to be detected based on the corrected compartment range and the predicted compartment range; The confirmation module is specifically used to: When a vehicle enters the image acquisition range of the unloading area, the license plate of the vehicle is identified as the unique code of the vehicle to confirm the vehicle to be detected; The vehicle detection system further includes: When the distance between the vehicle to be detected and the scrap steel pile is greater than a preset threshold, the relative position of the entire vehicle range and the carriage range is recorded; The correction module is specifically used to: Correcting the carriage range of the unloading area image at each moment based on the overall range of the vehicle and the relative position to obtain the corrected carriage range at each moment; The vehicle detection system further includes: Calculating the predicted compartment range based on Kalman filtering; Establish a tracker and update the Kalman filter state based on the optimized detection results of the next frame and the results of the Kalman filter prediction.

5. A readable storage medium, characterized in that: Programs and / or instructions are stored thereon, and when the programs and / or instructions are executed by a processor, the steps of the vehicle detection method according to any one of claims 1 to 3 are implemented.

6. An intelligent scrap steel grading system, characterized in that: comprising a vehicle detection system as claimed in claim 4; or The readable storage medium according to claim 5.

Images