CN109934161A - Vehicle identification and detection method and system based on convolutional neural network - Google Patents

Abstract

The invention relates to a vehicle identification and detection method and system based on a convolutional neural network, which is characterized by comprising the following steps of: extracting a vehicle picture sample and marking; carrying out region segmentation and type analysis training on the marked picture sample; extracting a group of random continuous images to be identified in a video to be identified; predicting the positions and the types of the vehicles in all the images to be recognized; outputting the motion state of the vehicle; the invention adopts a vehicle identification system to carry out intelligent and automatic management on related vehicles, controls and manages urban traffic by adopting a modern technical means, masters the running state of each vehicle in real time, can conveniently realize vehicle scheduling, can guide the rescue vehicle to track and dispose when danger occurs, and lays a solid foundation for timely leading and efficient monitoring of management departments.

Description

Vehicle recognition and detection method and system based on convolutional neural network

technical field

The invention relates to the technical field of vehicle identification and management, in particular to a method and system for vehicle identification and detection based on a convolutional neural network.

Background technique

With the rapid development of society and economy, the number of cars in major cities in various countries is increasing day by day. Illegal parking of vehicles is one of the major causes of traffic congestion. Therefore, various countries have clearly stipulated through corresponding laws and regulations that prohibition in specific places, places and roads. Parking. At present, the supervision of illegal parking by the traffic department mainly adopts the method of manual patrol. Therefore, the supervision of illegal parking by manual patrol requires a lot of manpower and material resources. There are few devices that can meet the real-time, accuracy and requirements for validity.

In addition, in recent years, the ownership of special vehicles such as oil tankers, hazardous chemical transport vehicles, dump trucks, military and police vehicles has also continued to increase. In order to determine the location of vehicles in motion, widely used A method of positioning by using radio wave signals from GPS (Global Positioning System). However, the accuracy of using GPS to locate the vehicle contains an error of about tens of meters, and it is difficult to determine the location in detail with higher accuracy. Therefore, the traditional management mode of special vehicles has been difficult to meet the actual needs, the supervision efficiency is low, and there is a safety problem. hidden danger.

As an important detection technology, traffic information collection based on video image processing has received extensive attention at home and abroad. With the development of social economy and the advancement of science and technology, video detection technology has also achieved rapid development. Video detection products are widely used in After going through two important development stages of analog and digital, it is now in the stage of high-definition development. At present, high-definition video detection products have appeared in the market. Traffic video detection sensors obtain traffic scene images through video acquisition equipment located above the road, and use computer image processing. , artificial intelligence, pattern recognition and other technologies automatically analyze and process scene image information to obtain traffic information. Because it is a non-contact traffic information collection device, it can be installed, debugged and maintained without affecting the operation of the vehicle without closing the road section. At the same time, the video detection sensor can detect multiple lanes at the same time. Traffic monitoring in scenarios has the characteristics of low cost, comprehensive and intuitive information, easy maintenance and installation, so it has a high application prospect in intelligent transportation systems.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to overcome the deficiencies in the prior art and provide a method and system for vehicle identification and detection based on a convolutional neural network.

The present invention is achieved through the following technical solutions:

A vehicle identification and detection method based on a convolutional neural network, characterized in that it comprises the following steps:

a. Extract a set of image samples of different types of vehicles with specific shapes, and mark all types of vehicles in the image samples;

b. Use mask-rcnn to perform regional segmentation and category analysis training on labeled image samples;

c. Establish a communication connection with the video monitoring terminal, and extract a group of random and continuous images to be recognized in the video to be recognized;

d. Predict the position and type of vehicles in all images to be recognized through mask-rcnn semantic segmentation;

e. Output the motion state of the vehicle through the motion tracking of the video;

f. Bind the motion state to the vehicle monitoring business logic, and output the monitoring results and instructions.

According to the above technical solution, preferably, the to-be-recognized image is a group of images continuously selected in a randomly extracted time period from the to-be-recognized video.

According to the above technical solution, preferably, step e includes: tracking all vehicle positions in the to-be-identified images of adjacent frames through an object tracking algorithm; and outputting the motion state of the vehicle according to the vehicle tracking situation.

According to the above technical solution, preferably, step f includes: when the motion state of the vehicle in the specific area is stationary, sending a reminder instruction to the monitoring terminal.

According to the above technical solution, preferably, step f includes: when the stationary time of the vehicle exceeds a preset time period, sending a reminder instruction to the monitoring terminal.

A vehicle identification and detection system based on a convolutional neural network, which is characterized by comprising: a labeling unit for extracting a group of picture samples of different types of vehicles with specific shapes, and labeling all types of vehicles in the picture samples; a training unit, It is used to use mask-rcnn to perform regional segmentation and type analysis training on the marked image samples; the extraction unit is used to establish a communication connection with the video monitoring terminal, and extract a group of random and continuous images to be recognized in the video to be recognized; vehicle information recognition The unit is used to predict the position and type of the vehicle in all the images to be recognized through the semantic segmentation of mask-rcnn; the vehicle displacement recognition unit is used to output the motion state of the vehicle through the motion tracking of the video; the monitoring unit is used to bind the motion state Vehicle monitoring business logic, output monitoring results and instructions.

According to the above technical solution, preferably, the vehicle displacement identification unit includes: a tracking module, used for tracking all vehicle positions in the to-be-identified images of adjacent frames through an object tracking algorithm; state of motion.

According to the above technical solution, preferably, the monitoring unit includes: a first judging module, configured to send a reminder instruction to the monitoring terminal when the vehicle is stationary in a motion state in a specific area.

According to the above technical solution, preferably, the monitoring unit includes: a second judging module, configured to send a reminder instruction to the monitoring terminal when the stationary time of the vehicle exceeds a preset time period.

The beneficial effects of the present invention are:

Through the image recorded by the video monitoring terminal, extract the video clips for identification and analysis, identify the motion state of the vehicle through the motion tracking of the video, and transmit the information command to the vehicle monitoring and management center through the Internet for the vehicle state that meets certain conditions, and use the vehicle identification system to identify the vehicle status. Relevant vehicles are managed intelligently and automatically. In order to use modern technical means to control and manage urban traffic, the running status of each vehicle can be controlled in real time, which can facilitate the realization of vehicle scheduling. When danger occurs, it can guide rescue vehicles to track and dispose of them. Lay a solid foundation for timely guidance and efficient monitoring by the management department.

Description of drawings

FIG. 1 is a schematic diagram of the working process of the present invention.

Detailed ways

In order to make those skilled in the art better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and the best embodiments.

As shown in the figure, the present invention discloses a vehicle identification and detection method based on a convolutional neural network, which is characterized by comprising the following steps: a. Mark all types of vehicles in the vehicle, and use the VIA image marking algorithm framework to finely mark the contours of all vehicles in the intercepted image samples containing various vehicle specific shapes, forming an outline surrounded by multiple closed polygons, and marking the vehicle type name at the same time. Export the marked information into a json file; b. Use mask-rcnn to perform regional segmentation and category analysis training on the marked image samples. In this example, the mask-rcnn algorithm under the tensorflow framework is used to train the json file sample; c. with video surveillance The terminal establishes a communication connection, and extracts a group of random and continuous images to be identified in the video to be identified. In this example, the video monitoring terminal can be a surveillance camera, which is used to monitor whether an ordinary vehicle invades a specific area, or a video shot by a satellite. It is used to monitor the motion state of special vehicles in different environments; d. Predict the position and type of vehicles in all images to be recognized through mask-rcnn semantic segmentation; e. Output the motion state of the vehicle through video motion tracking; f. Convert the motion state Bind the vehicle monitoring business logic and output monitoring results and instructions. Through the image recorded by the video monitoring terminal, the video clips are extracted for identification and analysis, the motion status of the vehicle is identified through the motion tracking of the video, and the information command is transmitted to the vehicle monitoring management center through the Internet for the vehicle status that meets certain conditions, and the vehicle identification system is used. Relevant vehicles are managed intelligently and automatically. In order to use modern technical means to control and manage urban traffic, the running status of each vehicle can be controlled in real time, which can facilitate vehicle scheduling, and can guide rescue vehicles to track and dispose of them when danger occurs. Lay a solid foundation for timely guidance and efficient monitoring by the management department.

According to the above embodiment, preferably, the to-be-recognized image is a group of images continuously selected in a time period randomly extracted from the to-be-recognized video. In this example, the image algorithm framework opencv is used to import the to-be-recognized video, and a certain number of images are extracted from the video. Time period, the extracted time period is related to the operation speed and observation time, which can be infinitely long but not too short. A group of images are randomly and continuously selected as the images to be recognized.

According to the above embodiment, preferably, step e includes: tracking all vehicle positions in the to-be-recognized images of adjacent frames through an object tracking algorithm, starting from the first frame in the to-be-recognized image, observing all vehicle positions in adjacent frame images, using The matrix algorithm calculates the centroid of the contour of each vehicle, takes the centroid of each vehicle in the first image as the center of the circle, and searches in sequence with the preset pixel as the radius to check the displacement of each vehicle in the next frame, and determine the distance between the object and the object where the centroid is located. Whether the shape difference is within a certain range, if it is, it is considered that the current object is the object after the displacement of the object where the centroid of the previous image is located, otherwise it is considered to be a different type of vehicle; output the motion state of the vehicle according to the vehicle tracking situation, if it is within a certain time. If the center of mass of the same vehicle does not move, the vehicle is considered to be stationary.

According to the above embodiment, preferably, step f includes: when the vehicle is in a stationary state in a specific area, sending a reminder instruction to the monitoring terminal; when the vehicle is in a stationary state in the prohibited parking position, the system sends a reminder instruction to the traffic management center, It enables traffic managers to quickly judge the motion state of vehicles, improves the intelligence level of road surveillance video, and provides timely and effective accident handling means and basis for traffic managers.

According to the above embodiment, preferably, step f includes: when the stationary time of the vehicle exceeds the preset duration, sending a reminder instruction to the monitoring terminal, setting a reasonable alarm threshold for the stationary duration of the vehicle, and when the special vehicle stays in some specific environments for a period of time When the alarm threshold is exceeded, the system sends a reminder command to the monitoring terminal, which is convenient for real-time control of the running status of each vehicle. When danger occurs, it can guide the rescue vehicle to track and deal with it, and lay a solid foundation for the management department to guide and monitor efficiently.

The invention also discloses a vehicle identification and detection system based on a convolutional neural network, which is characterized by comprising: a labeling unit for extracting a group of picture samples of different types of vehicles with specific shapes, and for all types of vehicles in the picture samples. Labeling; training unit, used to use mask-rcnn to perform regional segmentation and type analysis training on marked image samples; extraction unit, used to establish a communication connection with the video surveillance terminal, and extract a group of random and continuous in the video to be recognized. image; vehicle information recognition unit, used to predict the position and type of vehicles in all images to be recognized through mask-rcnn semantic segmentation; vehicle displacement recognition unit, used to output the motion state of the vehicle through video motion tracking; monitoring unit, used for Bind the motion state to the vehicle monitoring business logic, and output the monitoring results and instructions.

According to the above embodiment, preferably, the vehicle displacement identification unit includes: a tracking module for tracking all vehicle positions in the images to be identified in adjacent frames through an object tracking algorithm; state of motion.

According to the above embodiment, preferably, the monitoring unit includes: a first judging module, configured to send a reminder instruction to the monitoring terminal when the motion state of the vehicle in a specific area is stationary.

According to the above embodiment, preferably, the monitoring unit includes: a second judging module, configured to send a reminder instruction to the monitoring terminal when the stationary time of the vehicle exceeds a preset time period.

Through the image recorded by the video monitoring terminal, extract the video clips for identification and analysis, identify the motion state of the vehicle through the motion tracking of the video, and transmit the information command to the vehicle monitoring and management center through the Internet for the vehicle state that meets certain conditions, and use the vehicle identification system to identify the vehicle status. Relevant vehicles are managed intelligently and automatically. In order to use modern technical means to control and manage urban traffic, the running status of each vehicle can be controlled in real time, which can facilitate the realization of vehicle scheduling. When danger occurs, it can guide rescue vehicles to track and dispose of them. Lay a solid foundation for timely guidance and efficient monitoring by the management department.

The above are only the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principles of the present invention, several improvements and modifications can be made. It should be regarded as the protection scope of the present invention.

Claims (9)

1. a vehicle identification and detection method based on convolutional neural network, is characterized in that, comprises the following steps: a. Extract a set of image samples of different types of vehicles with specific shapes, and mark all types of vehicles in the image samples; b. Use mask-rcnn to perform regional segmentation and category analysis training on labeled image samples; c. Establish a communication connection with the video monitoring terminal, and extract a group of random and continuous images to be recognized in the video to be recognized; d. Predict the position and type of vehicles in all images to be recognized through mask-rcnn semantic segmentation; e. Output the motion state of the vehicle through the motion tracking of the video; f. Bind the motion state to the vehicle monitoring business logic, and output the monitoring results and instructions. 2 . The method for vehicle identification and detection based on a convolutional neural network according to claim 1 , wherein the to-be-recognized images are a group of images continuously selected in a time period randomly extracted from the to-be-recognized video. 3 . 3. a kind of vehicle identification and detection method based on convolutional neural network according to claim 2, is characterized in that, step e comprises: track all vehicle positions in the to-be-identified image of adjacent frame by object tracking algorithm; According to vehicle tracking Condition outputs the motion state of the vehicle. 4. The method for vehicle identification and detection based on a convolutional neural network according to claim 1 or 3, wherein step f comprises: sending a reminder instruction to the monitoring terminal when the vehicle is stationary in a motion state in a specific area. 5. The method for vehicle identification and detection based on a convolutional neural network according to claim 1 or 3, wherein step f comprises: when the stationary time of the vehicle exceeds a preset duration, sending a reminder instruction to the monitoring terminal. 6. A vehicle identification and detection system based on convolutional neural network, characterized in that, comprising: The labeling unit is used to extract a set of picture samples of different types of vehicles with specific shapes, and mark all types of vehicles in the picture samples; The training unit is used to use mask-rcnn to perform regional segmentation and category analysis training on labeled image samples; The extraction unit is used for establishing a communication connection with the video monitoring terminal, and extracting a group of random and continuous images to be recognized in the video to be recognized; The vehicle information recognition unit is used to predict the position and type of vehicles in all images to be recognized through mask-rcnn semantic segmentation; The vehicle displacement recognition unit is used to output the motion state of the vehicle through the motion tracking of the video; The monitoring unit is used to bind the motion state to the vehicle monitoring business logic, and output the monitoring results and instructions. 7. A kind of vehicle identification and detection system based on convolutional neural network according to claim 6, is characterized in that, described vehicle displacement identification unit comprises: The tracking module is used to track the positions of all vehicles in the to-be-identified images of adjacent frames through the object tracking algorithm; the output module is used to output the motion state of the vehicle according to the vehicle tracking situation. 8. A kind of vehicle identification and detection system based on convolutional neural network according to claim 7, is characterized in that, described monitoring unit comprises: the first judgment module, is used for when the vehicle motion state is static in the specific area, Send a reminder command to the monitoring terminal. 9. The vehicle identification and detection system based on a convolutional neural network according to claim 7, wherein the monitoring unit comprises: a second judgment module for when the stationary time of the vehicle exceeds a preset duration, Send a reminder command to the monitoring terminal.