TWI797468B - Control system and method for traffic signal light - Google Patents

Abstract

An control method for a traffic signal light includes: obtaining an image at a preset time point before a red light is turned off, wherein there is at least one car stopped by the red light in the image; determining a time length according to the image, wherein the time length is generated by an artificial intelligence algorithm; controlling a green light according to the time length.

Description

Traffic light control system and method

The present invention relates to a control system and method for traffic signal lamps, in particular to a control system and method for controlling traffic signal lamps using artificial intelligence.

Most of the intersections will be provided with traffic lights to control vehicles to move forward or stop. The traffic light time of traditional traffic lights is set according to experience, and the traffic light time will not be adjusted in real time according to the traffic flow. Therefore, when the traffic flow increases, the length of the green light may not be long enough to effectively relieve the traffic flow, resulting in traffic congestion. In order to solve the traffic congestion problem, the traditional way is to command the traffic police on the spot, or let the surveillance personnel remotely control the traffic lights through the images captured by the cameras installed at the intersections.

However, controlling traffic lights by manpower requires a very large number of personnel, and it is difficult for personnel to conduct uninterrupted command or monitoring for a long time.

One of the objectives of the present invention is to provide a control system and method for traffic lights.

One of the objectives of the present invention is to provide a control system and method that can instantly control the green light according to the traffic flow.

According to the present invention, a traffic signal light control method includes the following steps: acquiring an image at a preset time point before the end of the red light of the traffic signal light, wherein the image includes at least A vehicle; a time length is determined according to the image, wherein the time length is generated through an artificial intelligence algorithm; and a green light of the traffic signal light is controlled according to the time length.

According to the present invention, a traffic light control system includes a camera and an artificial intelligence device. The camera shoots an image at a preset time point before the end of the red light of the traffic signal, wherein the image includes at least one vehicle waiting for the red light. The artificial intelligence device is connected to the camera in a wired or wireless manner, and the artificial intelligence device includes an artificial intelligence algorithm for determining a time length for controlling the green light of the traffic signal light according to the image.

The control system and method of the present invention can adjust the time of the green light in real time according to the current traffic flow, so it can effectively relieve the traffic flow and solve the problem of traffic congestion.

FIG. 1 shows a schematic diagram of the application of traffic signal control of the present invention. The control system 10 of the present invention utilizes artificial intelligence (AI) to control the traffic lights 12 arranged at the intersection. The traffic signal light 12 includes a red light 122 , a yellow light 124 and a green light 126 . In an embodiment, the traffic signal light 12 may also have more than two green lights, for example, a green light indicating a left turn or a green light indicating a right turn are added. Referring to FIG. 1 , various components in the control system 10 can be arranged on the same bracket as the traffic signal light 12 , but it is not limited thereto. The control system 10 includes a camera 102 and an artificial intelligence device 104 . The camera 102 is used to shoot the intersection to obtain a video of the intersection, and the video is composed of a plurality of continuous images. In one embodiment, the camera 102 is a fisheye camera. The range photographed by the camera 102 includes at least a section of road before the traffic signal light 12 , and the road may include one or more lanes. When the red light 122 of the traffic signal lamp 12 was on, various vehicles would stop before the stop line 14 of the crossing. At this moment, the video camera 102 can capture a video of at least one vehicle (vehicle) that stops at the red light 122 . Vehicles mentioned herein include various means of transportation, such as automobiles and motorcycles. The camera 102 shoots an image A at a preset time point t before the end of the red light 122 , for example, 3 seconds before the end of the red light 122 . The camera 102 transmits the image A (as shown in FIG. 2 ) to the artificial intelligence device 104 in a wired or wireless manner. In another embodiment, the camera 102 may transmit the video to the artificial intelligence device 104, and the artificial intelligence device 104 obtains the image A shot at the preset time point t from a plurality of continuous images. The artificial intelligence device 104 is connected to the traffic signal lamp 12 in a wired or wireless manner. The artificial intelligence device 104 determines a time length T according to the image A, and the time length T is used to control the lighting time of the green light 126 . The length of time T is an estimated value. In the embodiment of FIG. 1 , the time length T represents the time required by the artificial intelligence device 104 to pass the stop line 14 for all the vehicles in at least one lane in the image A after the red light 122 changes to the green light 126 , But the time length T of the present invention is not limited to the time when all vehicles pass the stop line 14 . In one embodiment, the artificial intelligence device 104 controls the green light 126 according to the time length T. In another embodiment, the artificial intelligence control device 104 transmits the time length T to a remote server, and the server generates a timing plan according to the time length T to control the green light 126 .

It can be understood from the above description that the present invention proposes a traffic signal control method as shown in FIG. 3 . Referring to Fig. 1 and Fig. 3, the control method of the present invention is to acquire an image A at a preset time point t before the end of the red light 122, as shown in step S10, the image A includes at least one lane to stop and wait for the red light 122 at least one vehicle. In one embodiment, the camera 102 may be used to capture the image A at the preset time point t, and the camera 102 transmits the image A to the artificial intelligence device 104 in a wired or wireless manner. In another embodiment, the camera 102 shoots a plurality of continuous images to generate a movie, the camera 102 transmits the movie to the artificial intelligence device 104, and then the artificial intelligence device 104 acquires The captured image A. After the image A is obtained, proceed to step S12. The present invention determines a time length T according to the image A, wherein the time length T is generated through an artificial intelligence algorithm in the artificial intelligence device 104 . In one embodiment, the time length T represents the estimated time required by the artificial intelligence device 104 for all vehicles in at least one lane in the image A to pass the stop line 14 after the red light 122 changes to the green light 126 . Finally, the present invention controls the green light 126 according to the time length T, as shown in step S14. In one embodiment, the artificial intelligence device 104 controls the green light 126 according to the time length T. In another embodiment, the artificial intelligence control device 104 transmits the time length T to a remote server, and the server generates a timing plan according to the time length T to control the green light 126 .

FIG. 4 shows a first embodiment of the artificial intelligence device 104 in FIG. 1, which includes an artificial intelligence (AI) algorithm 1042 and a controller 1044, wherein the AI algorithm 1042 can be implemented by hardware circuits or software. FIG. 5 shows an operation example of the artificial intelligence device 104 in FIG. 4 , and steps S121 and S141 in FIG. 5 correspond to steps S12 and S14 in FIG. 3 , respectively. Referring to FIG. 4 and FIG. 5 , after the artificial intelligence device 104 obtains the image A, the AI algorithm 1042 in the artificial intelligence device 104 can generate a time length T according to the image A, as shown in step S121 . The AI algorithm 1042 can be, but not limited to, a convolutional neural network (Convolutional Neural Network; CNN). CNN includes a feature extraction (feature extraction) part and a regression (regression) part, and the feature extraction part is used to extract features from images. A characteristic information is generated, and the regression part generates a time length T according to the characteristic information. The controller 1044 controls the length of time that the green light 126 is on. In the embodiment of FIG. 4, the controller 1044 receives the time length T, and generates a control signal Sc to the traffic signal light 12 according to the time length T, so as to control the time length of the green light 126, as shown in step S141 of FIG. 5 shown.

FIG. 6 shows a second embodiment of the artificial intelligence device 104 in FIG. 1 , which includes an image processing circuit 1046 between the camera 102 and the AI algorithm 1042 in addition to the AI algorithm 1042 and the controller 1044 . FIG. 7 shows a first operational embodiment of the artificial intelligence device 104 of FIG. 6 . The combination of steps S122 and S123 in FIG. 7 can be understood as an embodiment of step S12 in FIG. 3 , and step S141 in FIG. 7 corresponds to step S14 in FIG. 3 . Referring to FIG. 6 and FIG. 7 , after the artificial intelligence device 104 obtains the image A, it first proceeds to step S122 . In step S122 , the image processing circuit 1046 performs preprocessing on the image A. The preprocessing includes classifying and labeling all the vehicles in the image A to generate a first labeled image B (as shown in FIG. 2 ). The image processing circuit 1046 can use but not limited to a computer vision algorithm or an artificial intelligence vision detection (AI vision detection) algorithm to classify and label at least one vehicle in the image A. In one embodiment, the manner of classifying the vehicles includes classifying the vehicles by vehicle types, such as buses, cars, locomotives, and trucks. In one embodiment, the vehicles in the image A can be marked with different colored marking frames. For example, as shown in FIG. Passenger cars and trucks are marked with a red marking frame 20 . Next, step S123 is executed by using the AI algorithm 1042 to analyze the first marker image B to generate a time length T. Different types of vehicles have different reaction times from when the green light 126 is turned on to when they start to move. Generally speaking, the reaction time of a locomotive is faster, while that of a bus and a truck is slower, so vehicles behind the bus or truck may need to More time to pass stop line 14 or intersection. The AI algorithm 1042 of the present invention estimates the time length T required for all vehicles to pass the stop line 14 according to the positional relationship between different types of vehicles in the marked image B. Finally, the controller 1044 generates a control signal Sc to the traffic signal light 12 according to the time length T, so as to control the time length of the green light 126 being on, as shown in step S141.

Fig. 8 shows the second operation embodiment of the artificial intelligence device 104 of Fig. 6, the combination of steps S122, S124 and S125 of Fig. 8 can be understood as another embodiment of step S12 of Fig. 3, and step S141 of Fig. 8 is corresponding Step S14 in FIG. 3 . Referring to FIG. 6 and FIG. 8 , after the artificial intelligence device 104 obtains the image A, the image processing device 1046 first performs preprocessing on the image A, as shown in steps S121 and S124 . Specifically, the image processing device 1046 first classifies and labels all the vehicles in the image A to generate a first labeled image B, as shown in step S122 . Afterwards, in order to prevent objects other than vehicles (such as road signs 22, street lamps 24, road trees 26, buildings 28, etc. shown in FIG. A second logo image C, as shown in step S124. In the embodiment of FIG. 2 , the second marking image C has a black background, and different types of vehicles are represented by color blocks of different colors, for example, cars are marked by green color blocks 30 , locomotives are marked by blue color blocks 32 , and Passenger cars and trucks are marked with red color blocks 34 . The second marker image C of the present invention is not limited to the embodiment shown in FIG. 2 . Next, step S125 is executed to use the AI algorithm 1042 to analyze the second marker image C to generate a time length T. Likewise, the AI algorithm 1042 of the present invention can estimate a time length T according to the positional relationship between different types of vehicles in the second marker image C. Finally, the controller 1044 generates a control signal Sc to the traffic signal light 12 according to the time length T, so as to control the time length of the green light 126 being on, as shown in step S141.

The AI algorithm 1042 of the artificial intelligence device 104 in FIG. 4 and FIG. 6 needs to be trained in advance to obtain the ability to estimate a time length T. The training method includes preparing many different images I, which may be taken at intersections such as those shown in FIG. 1 . These images I are shot at different time points, and each image I is shot at a preset time point t before the end of the red light 122 . On the other hand, it is also necessary to calculate the actual time Tr required for all the vehicles in each image I to pass the stationary line 14 . These images I and their corresponding actual times Tr are provided to a training program Pt. The training program Pt has the same model architecture as the AI algorithm 1042, such as CNN architecture. The training program Pt obtains a set of coefficients for estimating the time length T according to the images I and the corresponding actual time Tr. The AI algorithm 1042 uses the set of coefficients to analyze the image A and estimate the time T required for all the vehicles in the image A to pass through. The above-mentioned training process can be understood as allowing the training program Pt to learn how long it will take to let the vehicle in the image pass the stationary line 14 according to the image estimation. And through the above training process, the AI algorithm 1042 has the ability to estimate a time length T according to the image A.

In different embodiments, the traffic signal 12 may also include a right-turn signal 126 ″ and a left-turn signal 126 ′, and the road where the traffic signal 12 is located includes a left-turn lane 42 in addition to the straight lane 40 and the right-turn lane 44, as shown in Figure 9. According to the present invention, the image A shot by the camera 12 can be divided into the image Ad of the straight lane 40, the image Al of the left-turn lane 42 and the image Ar of the right-turn lane 44, that is The present invention can obtain the image Ad of the through lane, the image Al of the left-turn lane and/or the image Ar of the right-turn lane from the image A. In one embodiment, the artificial intelligence device 104 detects the road in the image A Arrows 402, 422 and 442 of the arrows 402 and 442 to identify lanes in different directions of travel. Generally speaking, the left-turn lane 42 will be marked with an arrow 422 representing a left turn, and the right-turn lane 44 will be marked with an arrow 442 representing a right turn. 40 will be marked with an arrow 402 representing going straight. The AI algorithm 1042 generates a time length T1 according to the image Ad to control the green light, generates a time length T2 according to the image Al to control the left turn light, and generates a time length T3 according to the image Ar to control the right turn light. Turn lights. In view of this situation, to train the AI algorithm 1042 to estimate the time T1, it is necessary to provide multiple images of the through lanes to the training program Pt. To train the AI algorithm 1042 to estimate the time T2, it is necessary to provide multiple images of the left and right lanes. The images of the turning lanes are provided to the training program Pt. To train the AI algorithm 1042 to estimate the time T3, it is necessary to provide multiple images of the right-turning lanes to the training program Pt. The rest of the details are similar to the aforementioned training process and will not be repeated here.

In the embodiment shown in FIG. 4 and FIG. 6 , the controller 1044 is set in the artificial intelligence device 104 , and the controller 1044 may be a computer or a hardware circuit for controlling the traffic signal lamp 12 . In different embodiments, the controller 1044 can also be set outside the control system 10, for example, the controller 1044 can be located in a remote traffic control center (Traffic Control Center) (not shown in the figure). In this case, the control system 10 transmits the time length T to the controller 1044 of the remote control center in a wired or wireless manner, and then the controller 1044 generates a control signal Sc according to the time length T to control the green light via a wired or wireless network 126 lighting hours.

The above descriptions of the preferred embodiments of the present invention are for the purpose of illustration, and are not intended to limit the present invention to the disclosed form. It is possible to modify or change based on the above teachings or learning from the embodiments of the present invention. The embodiment is selected and described in order to explain the principle of the present invention and to allow those familiar with the art to use the present invention in various embodiments for practical application. Decide.

10: Artificial intelligence control device 102: camera 104: Artificial intelligence device 1042: Artificial intelligence algorithm 1044: Controller 1046: Image processing circuit 12:Traffic lights 122: red light 124: yellow light 126: green light 126': left turn signal 126": right turn signal 14: Stop line 16: Green label box 18: blue label box 20: Red label box 22: Road sign 24: street lights 26: road tree 28: Buildings 30: green color block 32: blue color block 34: red color block 40: Straight lane 402: arrow 42: Left turn lane 422: arrow 44: Right turn lane 442: arrow A: Image B: The first logo image C: Second logo image

Fig. 1 shows a schematic diagram of the application of the control system of the present invention. FIG. 2 shows an embodiment of image processing in the present invention. Fig. 3 shows an embodiment of the control method of the present invention. FIG. 4 shows a first embodiment of the artificial intelligence device in FIG. 1 . FIG. 5 shows the operation steps of the artificial intelligence device in FIG. 4 . FIG. 6 shows a second embodiment of the artificial intelligence device in FIG. 1 . FIG. 7 shows a first operational embodiment of the artificial intelligence device of FIG. 6 . FIG. 8 shows a second operational embodiment of the artificial intelligence device of FIG. 6 . FIG. 9 shows a schematic diagram of the control system of the present invention applied to a road with multiple lanes.

10: Artificial intelligence control device

102: camera

104: Artificial intelligence device

12:Traffic lights

122: red light

124: yellow light

126: green light

14: Stop line

16: Green label box

18: blue label box

20: Red label box

22: Road sign

24: street lights

26: road tree

28: Buildings

Claims (18)

A control method of a traffic signal light, the traffic signal light includes a red light and a green light, the method comprises the following steps: A. acquiring an image at a preset time point before the end of the red light, wherein the image including at least one vehicle waiting for the red light; B. determining a length of time according to the positional relationship between different types of vehicles in the image, wherein the length of time is generated through an artificial intelligence algorithm; and C. according to the The length of time controls this green light. As the control method of claim 1, wherein the artificial intelligence algorithm is realized by a convolutional neural network. The control method according to claim 1, wherein the step B further includes preprocessing the image, and the preprocessing includes classifying and labeling the at least one vehicle in the image. The control method according to claim 3, wherein the preprocessing further includes representing marked vehicles of different types with color blocks of different colors. The control method according to claim 3 or 4, wherein the step of classifying and marking the at least one vehicle in the image includes using a computer vision algorithm or an artificial intelligence visual detection algorithm. The control method according to claim 1, wherein the step A includes: acquiring a video of an intersection; and acquiring the image shot at the preset time point according to a plurality of consecutive images in the video. The control method according to claim 1, wherein the image includes images of a through lane, a left-turn lane and/or a right-turn lane. As the control method of claim item 7, it further includes detecting the arrow on the road in the image from the The image of the through lane, the image of the right-turn lane or the image of the left-turn lane is obtained from the image. The control method according to claim 8, wherein the step B further includes determining the time length according to the image of the through lane, the image of the right-turn lane or the image of the left-turn lane. A traffic signal light control system, the traffic signal light includes a red light and a green light, the control system includes: a camera, at a preset time point before the end of the red light, shoots an image, wherein the image Including at least one vehicle waiting for the red light; and an artificial intelligence device connected to the camera in a wired or wireless manner, the artificial intelligence device includes an artificial intelligence algorithm for determining the position between different types of vehicles in the image The relationship determines a length of time for controlling the green light. The control system according to claim 10, wherein the artificial intelligence algorithm is realized by a convolutional neural network. The control system according to claim 10, wherein the artificial intelligence device further includes an image processing circuit connected to the camera for preprocessing the image, and the preprocessing includes classifying and labeling the at least one vehicle in the image. The control system according to claim 12, wherein the image processing circuit further uses color blocks of different colors to indicate different types of marked vehicles. The control system according to claim 12 or 13, wherein the image processing circuit includes a computer vision algorithm or an artificial intelligence visual detection algorithm for classifying and marking the at least one vehicle. The control system according to claim 10, wherein the camera obtains a video of an intersection, and the artificial intelligence device obtains the image at the preset time point from the video. Such as the control system of claim item 10, wherein the image includes the through lane, the left turn lane and /or right turn lane. The control system according to claim 16, wherein the artificial intelligence device detects the arrow on the road in the image to obtain the image of the through lane, the image of the right-turn lane or the image of the left-turn lane from the image. The control system according to claim 17, wherein the artificial intelligence algorithm determines the time length according to the image of the through lane, the image of the right-turn lane or the image of the left-turn lane.

Images