Method and device for full-automatic identification of muck dumping vehicle

Abstract

The invention discloses a method and a device for fully automatically identifying a muck dumping vehicle, and the method comprises the steps: 1, installing a monitoring device on a road of a monitoring region, and capturing pictures of passing vehicles through the monitoring device, the pictures including license plates, vehicle bodies and hopper loads; step 2, identifying the vehicle picture by using an image identification model, wherein the vehicle picture comprises a license plate number and empty and full load identifiers of the muck vehicle; wherein the value of the empty and full load identification of the muck truck is full load of the muck truck, no load of the muck truck and others; 3, the vehicles with the slag car empty and full load identification values being no load in the step 2 are selected, and the step 4 is executed; step 4, for the vehicle selected in the step 3, reversely checking whether the identification result of the last snapshot of the vehicle is that the muck vehicle is fully loaded or not, if so, entering step 5, and if not, judging that the vehicle does not pour muck in the snapshot of this time; and step 5, based on the time of two snapshots of the vehicle, judging whether the vehicle stays for a short time, if so, determining that the vehicle pours the muck, and otherwise, determining that the vehicle does not pours the muck in the snapshots of this time. According to the invention, the muck dumping behavior can be identified.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence image recognition technology, specifically a method and apparatus for fully automatic identification of dump trucks. Background Technology

[0002] The illegal dumping of construction waste persists despite repeated crackdowns due to a combination of factors, including economic costs and management loopholes. The difficulty in investigating and controlling illegal dumping lies not only in its high degree of concealment but also in the practical obstacles of detection and evidence collection. Unauthorized individuals often choose to operate secretly in remote suburban areas at night, where the lack of surveillance equipment makes it difficult to determine the time of dumping and to detect the activity in a timely manner.

[0003] Currently, the most common regulatory method is to install tracking devices on dump trucks to monitor their behavior. However, this method is ineffective in controlling illegal dumping across administrative regions, which often occurs in the "border areas" where administrative regions meet. Effectively monitoring vehicles from other areas that illegally dump waste within the administrative region is crucial to combating illegal dumping. Summary of the Invention

[0004] The purpose of this invention is to provide a fully automatic method and device for identifying vehicles dumping construction waste, aiming to solve the problem of identifying illegal dumping of construction waste.

[0005] The technical solution to achieve the purpose of this invention is as follows:

[0006] A fully automated method for identifying dump trucks carrying construction waste includes:

[0007] Step 1: Install surveillance equipment on the roads in the monitored area and use the surveillance equipment to capture photos of passing vehicles, including license plates, vehicle bodies, and cargo in the truck bed;

[0008] Step 2: Use an image recognition model to identify vehicle photos, including: license plate number, time, whether it is a dump truck, and monitoring equipment number; among them, the values ​​for dump truck empty / full load are: dump truck fully loaded, dump truck empty, and others.

[0009] Step 3: Select the vehicles whose empty / full load indicator value is "empty" in Step 2, and proceed to Step 4.

[0010] Step 4: For the vehicles selected in Step 3, check whether the identification result of the last capture of the vehicle was that the dump truck was fully loaded. If so, proceed to Step 5; otherwise, the vehicle was not dumping dump trucks in this capture.

[0011] Step 5: Based on the time of the two captures of the vehicle, determine whether the vehicle made a brief stop. If so, it is considered that the vehicle dumped construction waste; otherwise, it is considered that the vehicle did not dump construction waste in this capture.

[0012] Furthermore, the monitoring device is a camera.

[0013] Furthermore: Based on historical driving big data statistics, the normal driving time matrix between any two cameras is obtained to obtain the normal passage time. The time difference between the current vehicle passage time and the normal passage time is calculated. If the time difference is within a set range, a short stop is allowed.

[0014] Furthermore, the set interval is from 15 minutes to one hour.

[0015] Furthermore, step 5 uses a learning network model to determine whether the vehicle has made a brief stop. The learning network model is trained based on historical driving big data, and the trained learning network model is used to determine whether the vehicle has made a brief stop.

[0016] A fully automatic device for identifying dump trucks carrying construction waste includes:

[0017] The vehicle image capture module is used to capture images of passing vehicles and store the images on a file server.

[0018] The vehicle image recognition module retrieves vehicle images from the file server, uses an image recognition model to identify the license plate number, time, whether it is a dump truck, and the monitoring equipment number, and saves all recognition results; among them, the dump truck empty / full load identifier has the following values: dump truck full load, dump truck empty, and others.

[0019] The dumping identification module first filters out vehicles identified as empty dump trucks from the identification results; then, for empty dump trucks, it determines whether the vehicle was identified as a fully loaded dump truck in the previous capture; finally, for vehicles that were identified as fully loaded dump trucks in the previous capture, it determines whether the vehicle made a brief stop. If it did make a brief stop, it is identified as a dumping vehicle.

[0020] The identification result output module is used to save the identification results in the dumping soil identification module and provide a data interface for other applications to call.

[0021] Furthermore, it also includes a monitoring equipment management module, which is used to manage all monitoring equipment, configure the monitoring equipment, and monitor its status.

[0022] Compared with the prior art, the beneficial effects of the present invention are as follows: Based on the characteristics of dumping construction waste, which generally involves dump trucks transporting waste, entering fully loaded, stopping in the middle, and leaving empty, the present invention designs a method to determine whether a vehicle stops to dump construction waste. The present invention identifies the dumping behavior, solves the problem of identifying illegal dumping of construction waste, and can be applied to early warning of illegal dumping of construction waste and post-event evidence collection, assisting regulatory authorities in controlling vehicles that steal construction waste. Attached Figure Description

[0023] Figure 1 This is a flowchart of a fully automatic method for identifying dump trucks according to the present invention.

[0024] Figure 2 This is a structural diagram of a fully automatic device for identifying illegally dumped construction waste. Detailed Implementation

[0025] Combination Figure 1 This embodiment provides a method for fully automatic identification of dump trucks, including the following steps:

[0026] 1. Install surveillance cameras on the roads within the monitored area to capture photos of passing vehicles. The cameras must be able to capture clear images of vehicles, including license plates, the vehicle body, and the cargo in the truck bed.

[0027] 2. Vehicle Recognition: AI image recognition technology is used to identify vehicle photos, including: license plate number and the empty / full load indicator for the dump truck. The empty / full load indicator can be set to: fully loaded, empty, or other.

[0028] 3. Select the vehicles identified as "empty dump trucks" in step 2 and proceed to step 4.

[0029] 4. For the vehicle in step 3, check back to see if the identification result of step 2 in the previous capture was "full load of construction waste truck". If it is "yes", proceed to step 5. If it is "no", the vehicle did not dump construction waste in this capture.

[0030] 5. Calculate the time between the two snapshots in steps 3 and 4 to determine if the vehicle made a brief stop. A "yes" result indicates the vehicle dumped construction waste; a "no" result indicates the vehicle did not dump construction waste in this snapshot. Historical driving data can be used to statistically analyze the normal driving time matrix between any two cameras, thus obtaining the normal passage time. The time difference between the vehicle's passage time and the normal passage time can be calculated. A time difference within a certain range indicates a brief stop. Based on experience, a time difference of 15 minutes to one hour is generally considered a brief stop. Alternatively, historical driving data can be used to train an AI model, which can then be used to directly identify whether a brief stop has occurred.

[0031] In another aspect of this embodiment, a device for fully automatic identification of dump trucks is also provided, combined with Figure 2 ,include:

[0032] Camera Management Module: Manages all surveillance cameras, allowing for camera settings and status monitoring.

[0033] Vehicle image capture module: Captures images of passing vehicles and stores the images on a file server.

[0034] Vehicle image recognition module: Retrieves vehicle images from the file server, uses a vehicle image recognition model to identify license plate numbers and the empty / full load status of dump trucks, and saves all recognition results. The empty / full load status of dump trucks can be: fully loaded, empty, or other.

[0035] The construction waste dumping identification module works as follows: First, it filters out vehicles identified as "empty construction waste trucks" from the identification results. Second, it further filters out vehicles whose identification result in the previous capture was "fully loaded construction waste trucks." Finally, it filters out vehicles that briefly stopped between the previous and current captures; these vehicles can be identified as construction waste dumping vehicles. The method for determining whether there is a brief stop is as follows: It can be based on historical driving data to statistically analyze the normal driving time matrix between any two cameras, thus obtaining the normal passage time. The time difference between the vehicle's passage time and the normal passage time is calculated; a time difference within a certain range indicates a brief stop. Alternatively, it can directly train an AI model using historical driving data and then use the AI ​​model to directly identify whether there is a brief stop.

[0036] Recognition Result Output Module: Saves the recognition results from the waste dumping recognition module and provides a data interface for other applications to use. Recognition results include: license plate number, recognition time, whether it is a waste dumping vehicle, camera number, etc.

[0037] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for fully automatic identification of dump trucks, characterized in that, include: Step 1: Install surveillance equipment on the roads in the monitored area and use the surveillance equipment to capture photos of passing vehicles, including license plates, vehicle bodies, and cargo in the truck bed; Step 2: Use an image recognition model to identify vehicle photos, including: license plate number, time, whether it is a dump truck, and monitoring equipment number; among them, the values ​​for dump truck empty / full load are: dump truck fully loaded, dump truck empty, and others. Step 3: Select the vehicles whose empty / full load indicator value is "empty" in Step 2, and proceed to Step 4. Step 4: For the vehicles selected in Step 3, check whether the identification result of the last capture of the vehicle was that the dump truck was fully loaded. If so, proceed to Step 5; otherwise, the vehicle was not dumping dump trucks in this capture. Step 5: Based on the time of the two captures of the vehicle, determine whether the vehicle made a brief stop. If so, it is considered that the vehicle dumped construction waste; otherwise, it is considered that the vehicle did not dump construction waste in this capture.

2. The method for fully automatic identification of dump trucks according to claim 1, characterized in that, The monitoring device is a camera.

3. The method for fully automatic identification of dump trucks according to claim 1, characterized in that, Step 5 specifically includes: statistically analyzing the normal driving time matrix between any two cameras based on historical driving big data to obtain the normal passage time, calculating the time difference between the current vehicle passage time and the normal passage time, and if the time difference is within a set range, then there is a brief stop.

4. The method for fully automatic identification of dump trucks according to claim 3, characterized in that, The set interval is from 15 minutes to one hour.

5. The method for fully automatic identification of dump trucks according to claim 1, characterized in that, Step 5 uses a learning network model to determine whether the vehicle has made a brief stop. The learning network model is trained based on historical driving big data, and the trained learning network model is used to determine whether the vehicle has made a brief stop.

6. A device for fully automatic identification of dump trucks implementing the method of any one of claims 1-9, characterized in that, include: The vehicle image capture module is used to capture images of passing vehicles and store the images on a file server. The vehicle image recognition module retrieves vehicle images from the file server, uses an image recognition model to identify the license plate number, time, whether it is a dump truck, and the monitoring equipment number, and saves all recognition results; among them, the dump truck empty / full load identifier has the following values: dump truck full load, dump truck empty, and others. The dumping identification module first filters out vehicles identified as empty dump trucks from the identification results; then, for empty dump trucks, it determines whether the vehicle was identified as a fully loaded dump truck in the previous capture; finally, for vehicles that were identified as fully loaded dump trucks in the previous capture, it determines whether the vehicle made a brief stop. If it did make a brief stop, it is identified as a dumping vehicle. The identification result output module is used to save the identification results in the dumping soil identification module and provide a data interface for other applications to call.

7. The fully automatic device for identifying illegal dumping of construction waste according to claim 6, characterized in that, It also includes a monitoring equipment management module, which is used to manage all monitoring equipment, configure the monitoring equipment, and monitor its status.

Images