Artificial intelligence-based truck height measurement system and method

Abstract

The present invention relates to a truck height measurement system and method capable of measuring the height of a moving truck using a CCTV camera installed on a road, by combining Python, a computer vision library OpenCV, and a library for 3D object tracking when utilizing 3D object detection and tracking technology. The present invention comprises the steps of: photographing a moving vehicle through a camera device (100) (S100); extracting truck-related object information from the captured data by a truck height measurement device (200) (S110); extracting truck object information from the truck-related object information by the truck height measurement device (200) (S120); extracting a truck image from the truck object information by the truck height measurement device (200) (S130); and performing preprocessing to reduce saturation in the extracted truck image by the truck height measurement device (200) (S140). The present invention provides a truck height measurement method characterized by including the step (S150) of the truck height measuring device (200) measuring the truck height in the preprocessed truck image.

Description

Technology Field

[0001] The present invention relates to artificial intelligence-based truck height measurement, and more specifically, to a truck height measurement system and method capable of measuring the height of a truck in operation by utilizing 3D object detection and tracking technology, by combining Python and a computer vision library, OpenCV, with a library for 3D object tracking. Background Technology

[0003] With the rapid increase in grade-separated roads today, the installation of various structures such as overpasses, tunnels, bridge piers, and elevated roads is on the roads where vehicles travel; furthermore, general buildings have entry height restrictions based on the height of vehicles attempting to enter, and facilities crossing roads are also increasing in tandem with the growth of electrical and communication facilities.

[0004] Accordingly, structures such as overpasses, tunnels, and underpasses generally installed on roads have height restrictions to prevent damage to vehicles and cargo, and in most cases, these height restrictions are indicated at the entrance of the structure.

[0005] Therefore, the driver of the vehicle enters after comparing the height limit indicated on the structure with the vehicle's

[0005] height.

[0006] Structures with such height limits are marked with height restrictions in accordance with road traffic regulations. In general, the Road Traffic Act applies to all types of vehicles traveling on the road. The Road Traffic Act is a law enacted to prescribe necessary matters regarding road traffic in order to ensure safe and smooth traffic by preventing and eliminating all traffic hazards and obstacles occurring on the road.

[0007] In accordance with Article 59 of the current Road Traffic Act and Article 55 of the Enforcement Decree thereof, in order to preserve the structure of the road and prevent operational hazards, restrictions on operation are stipulated for vehicles exceeding an axle load of 10 tons, a gross weight of 40 tons, and other specifications (height, length, width). The Ministry of Land, Transport and Maritime Affairs has established the "Regulations on Enforcement of Vehicles with Operation Restrictions," which specifies enforcement guidelines for the above laws and the Enforcement Decree (applying 10% of the measured amount considering measuring instruments and estimation errors, etc.) to enforce the operation of vehicles with operation restrictions.

[0008] In addition to these structures, the clearance height of vehicles entering highways is also restricted. Particularly as the usage of Hi-Pass increases and Hi-Pass entrances become accessible to all vehicles in the future, there is a need for a system to enforce these height restrictions.

[0009] This is particularly problematic because while drivers generally know the approximate height of their vehicles, in the case of trucks, there is a possibility of accidents occurring when they are unaware of various structures such as overpasses, tunnels, bridges, and elevated roads that are too high to enter depending on the type of goods loaded onto the vehicle, and there was a problem in that drivers could not be accurately informed of this.

[0010] In addition, there was a problem where the vehicle and cargo were damaged due to miscalculation, as the driver visually compares the vehicle's height with the height limit marked on the structure. Prior art literature

[0012] Patent Document 1: Korean Published Patent No. 10-2014-0065037 (System for limiting vehicle passage height) Patent Document 2: Korean Published Patent No. 10-2012-0065558 (Method for detecting vehicles on a road using region of interest setting) The problem to be solved

[0013] Accordingly, the present invention aims to solve the various disadvantages and problems of the prior art as described above, and provides a truck height measurement system and method capable of measuring the height of a truck by combining Python, a computer vision library OpenCV, and a library for 3D object tracking when utilizing 3D object detection and tracking technology to measure the height of a truck in operation using a CCTV camera installed on a road. means of solving the problem

[0015] To achieve the above-mentioned purpose, the present invention provides a truck height measuring system characterized by comprising: a camera device (100) that captures one or more of the front, rear, and side of a vehicle in operation using a CCTV camera; and a truck height measuring device (200) that detects a truck among the vehicles in operation by utilizing 3D object detection and tracking technology on the image captured by the camera device (100), and measures the truck height by combining Python, a computer vision library OpenCV, and a library for 3D object tracking when measuring the height of the truck.

[0016] Here, the truck height measuring device (200) comprises a communication unit (210) that receives image information of a vehicle in operation from the camera unit (100), a program storage unit (220) that stores Python, a computer vision library OpenCV, and a library for 3D object tracking, a truck height measuring module (230) that loads the Python, a computer vision library OpenCV, and a library for 3D object tracking stored in the program storage unit (220) to preprocess the image of the vehicle in operation transmitted through the communication unit (210) to extract truck-related object information, extracts a truck image according to the extraction of truck object information from the truck-related object information, preprocesses the extracted truck image, and measures the truck height, and a control unit (250) that controls the communication unit (210), the program storage unit (220), and the truck height measuring module (230).

[0017] To achieve the above-mentioned purpose, the present invention provides a truck height measurement method characterized by comprising: a step (S100) of photographing a vehicle in operation through a camera device (100); a step (S110) of extracting truck-related object information from the photographed data by a truck height measuring device (200); a step (S120) of extracting truck object information from the truck-related object information by the truck height measuring device (200); a step (S130) of extracting a truck image from the truck object information by the truck height measuring device (200); a step (S140) of performing preprocessing to lower the saturation of the extracted truck image by the truck height measuring device (200); and a step (S150) of measuring the truck height by the truck height measuring device (200) from the preprocessed truck image. Effects of the invention

[0019] According to the present invention, the following effects are achieved.

[0020] First, a vehicle height measurement system can be provided that can detect a vehicle through a camera installed on a road or building, and to measure the height of the truck including cargo and loads for the detected vehicle, by combining Python and a computer vision library, OpenCV, with a library for 3D object tracking to measure the height of the truck, thereby warning the driver about the entrance / exit route the truck intends to pass through or prohibiting the vehicle from entering.

[0021] Second, highly reliable truck height extraction is possible by utilizing a method that extracts vehicles through cameras installed on roads or buildings, and verifies reliability based on the extracted scores obtained by extracting scores and classes from the detected objects. Brief explanation of the drawing

[0023] FIG. 1 is a drawing for explaining an embodiment of a truck height measuring system according to the present invention, FIG. 2 is a drawing for explaining an example of a camera configuration for measuring truck height in the truck height measuring system shown in FIG. 1. FIG. 3 is a flowchart illustrating an embodiment of a truck height measurement method according to the present invention, FIGS. 4 to 8 are flowcharts illustrating embodiments of a truck height measurement method according to the present invention. Specific details for implementing the invention

[0024] A preferred embodiment of the present invention will be described in detail with reference to the attached drawings as follows.

[0025] Furthermore, while the terms used in this invention have been selected to be as widely used as possible, there are also terms arbitrarily chosen by the applicant in specific cases. Since the meanings of these terms are described in detail in the relevant description of the invention, it should be noted that the invention should be understood based on the meaning of the terms rather than their mere names. Additionally, in describing the embodiments, explanations regarding technical content that is well known in the technical field to which this invention belongs and is not directly related to this invention are omitted. This is intended to convey the essence of the invention more clearly without obscuring it by omitting unnecessary explanations.

[0027] FIG. 1 is a drawing for explaining an embodiment of a truck height measuring system according to the present invention, and FIG. 2 is a drawing for explaining an embodiment of a camera configuration for measuring truck height in the truck height measuring system shown in FIG. 1.

[0028] An embodiment of the truck height measuring system according to the present invention is composed of a camera device (100) and a truck height measuring device (200), as shown in FIG. 1. In addition, it may further include a safety device (300).

[0029] Here, the camera device (100) photographs a vehicle in operation using a CCTV camera and may be composed of a plurality of first to N cameras (110, 120, 130), and as shown in FIG. 2, these plurality of first to N cameras (110, 120, 130) may be composed of a plurality of cameras that photograph the front, rear, and side of the vehicle (truck).

[0030] Meanwhile, the truck height measuring device (200) detects a truck among vehicles in motion and measures the height of the truck by utilizing 3D object detection and tracking technology. In the present invention, the truck height is measured by combining Python, the computer vision library OpenCV, and a library for 3D object tracking.

[0031] To this end, the truck height measuring device (200) is configured to include a communication unit (210), a program storage unit (220), a truck height measuring module (230), and a control unit (250). Additionally, if necessary, it may further include an entrance / exit road height data storage unit (240).

[0032] The communication unit (210) receives video information of a vehicle in operation from the camera device (100). And, if necessary, transmits control information for operating the safety device to the safety device (300).

[0033] The program storage unit (220) stores Python, the computer vision library OpenCV, and a library for 3D object tracking. This program storage unit (220) can be composed of a hard disk or memory.

[0034] The truck height measurement module (230) loads Python, a computer vision library OpenCV, and a 3D object tracking library stored in the program storage unit (220), preprocesses an image of a moving vehicle transmitted through the communication unit (210) to extract truck-related object information, extracts a truck image based on the extraction of truck object information from the truck-related object information, and measures the truck height after preprocessing the extracted truck image.

[0035] In this case, for truck image preprocessing, a method can be used to detect the truck's boundary lines by converting from BGR (Blue, Green, Red) to grayscale, applying a Gaussian blur to the grayscale image to blur it—that is, to lower the saturation—and then detecting edges in the lowered-saturation image. This is because the height of the truck must include the cargo loaded on the truck, not just the height of the truck itself. For example, there are roofed trucks and roofless trucks; in the case of roofless trucks, the height of the truck can vary depending on the cargo, and in such cases, the truck driver may sometimes be unaware of or mistaken about the truck's height.

[0036] And as an embodiment for implementing this, the truck height measurement module (230) of the present invention may be composed of a truck-related object information extraction unit (231) for extracting truck-related object information from an image of a vehicle in operation transmitted through a communication unit (210), a truck object information extraction unit (232) for extracting truck object information from the truck-related object information extracted by the truck-related object information extraction unit (231), a truck image extraction unit (233) for extracting a truck image from the truck object information extracted by the truck object information extraction unit (232), a truck image preprocessing unit (234) for performing preprocessing on the truck image extracted by the truck image extraction unit (233), and a truck height measurement unit (235) for measuring the height of the truck from the truck image preprocessed by the truck image preprocessing unit (234).

[0037] The entrance / exit height data storage unit (240) stores height data of the entrance / exit road installed in the direction in which a vehicle or truck is traveling.

[0038] The control unit (250) controls the communication unit (210), the program storage unit (220), the truck height measurement module (230), and the entrance / exit road height data storage unit (240), and compares the truck height measured by the truck height measurement module (230) with the height of the entrance / exit road in the direction of vehicle travel stored in the entrance / exit road height data storage unit (240). Meanwhile, when connected to the safety device (300) described later, it transmits data indicating that the truck cannot pass through the entrance / exit road to the safety device (300) of the entrance / exit road with a height restriction in the direction of truck travel, according to the measured truck height.

[0039] The truck height measurement system may further include a safety device (300), which is intended to give a warning to the driver of the truck if the height of the truck is higher than the height of the entrance / exit road (tunnel, underpass, bridge, etc.) in the direction of the truck's travel according to the measurement result of the truck height measurement device (200). It may be configured to include a communication unit (310), a warning light (320), a barrier bar (330), a display (340), and a control unit (350). When data is received indicating that the height of the truck is higher than the entrance / exit road, the control unit (350) controls the warning light (320), the barrier bar (330), and the display (340) to prevent the truck driver from entering the entrance / exit road. At this time, the warning light may be made to flash, for example, a red light, the barrier bar (330) may directly prohibit entry in front of the truck, and the display (340) may display a notice prohibiting entry.

[0041] FIG. 3 is a flowchart for explaining an embodiment of a truck height measurement method according to the present invention, and FIGS. 4 to 8 are flowcharts for explaining an embodiment of a truck height measurement method according to the present invention.

[0042] An embodiment of the truck height measurement method according to the present invention is shown in FIG. 3, wherein a vehicle in operation is photographed through a camera device (100) (S100). At this time, the front, rear, and side of the vehicle are photographed.

[0043] The captured video data is transmitted to a truck height measuring device (200), and the truck height measuring device (200) extracts truck-related object information from the captured data (S110).

[0044] An example of a method for extracting such truck-related object information is described in more detail with reference to FIG. 4. First, in order to measure the height of a truck in operation, the present invention utilizes 3D object detection and tracking technology. To this end, Python and a combination of the computer vision library OpenCV and a library for 3D object tracking are used.

[0045] To this end, OpenCV and necessary libraries are installed in the program storage unit (220), and code for 3D object detection and tracking is input.

[0046] An example of such code is as follows.

[0047] 1. Install OpenCV and necessary libraries:

[0048] pip install opencv-python

[0049] pip install opencv-contrib-python

[0050] pip install numpy

[0052] 2. Code for 3D Object Detection and Tracking:

[0053] import cv2

[0054] import numpy as np

[0056] # YOLOv4 model and weight file paths

[0057] model_path = 'path / to / yolov4.weights'

[0058] config_path = 'path / to / yolov4.cfg'

[0060] # YOLOv4 Model Load

[0061] net = cv2.dnn.readNetFromDarknet(config_path, model_path)

[0063] # Load class name

[0064] with open('path / to / coco.names', 'r') as f:

[0065] classes = [line.strip() for line in f.readlines()]

[0067] # Settings when using GPU

[0068] net.setPreferableBackend(cv2.dnn.DNN_BACKEND_CUDA)

[0069] net.setPreferableTarget(cv2.dnn.DNN_TARGET_CUDA)

[0071] # Open video file or real-time video stream

[0072] cap = cv2.VideoCapture('path / to / video.mp4')

[0074] while True:

[0075] ret, frame = cap.read() # Read frame from video

[0077] if not ret:

[0078] break

[0080] Object Detection Using YOLOv4

[0081] blob = cv2.dnn.blobFromImage(frame, 1 / 255.0, (608, 608), swapRB=True, crop=False)

[0082] net.setInput(blob)

[0083] layer_names = net.getLayerNames()

[0084] output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]

[0085] outputs = net.forward(output_layers)

[0087] boxes = []

[0088] confidences = []

[0089] class_ids = []

[0090] When this code is executed (S111), the captured image taken by the camera device (100: 110, 120, 130) is preprocessed to prepare the input to the neural network (S112), and when the preprocessed captured image is input to the neural network (S113), the layers and output settings of the neural network are configured (S114). Then, the neural network is executed to perform object detection (S115), and the execution result is processed to extract only truck-related object information (S116).

[0091] The learning required for such neural networks may be utilized from the applicant RC Co., Ltd.’s registered patent No. 10-2296471 (AI-based vehicle loading defect detection system).

[0092] Next, the truck height measuring device (200) extracts truck object information from truck-related object information (S120).

[0093] An embodiment of the method for extracting such truck object information is described in more detail with reference to FIG. 5. As described above, the code is executed (S121), and the output list of extracted truck-related object information is iterated to perform processing for each output (S122). Then, the detection list, i.e., the extraction list, is iterated to perform processing for each output detection (S123). Subsequently, a score and a class are extracted from each detected object (S124), and the reliability is checked based on the extracted score (S125). Then, if a preset level of reliability is satisfied, object information (truck, etc.) is extracted (S126), and the extracted object information is stored (S127). An embodiment of the code for extracting the score and class from each object is as follows:

[0094] # Extract detected object information

[0095] for output in outputs:

[0096] for detection in output:

[0097] scores = detection[5:]

[0098] class_id = np.argmax(scores)

[0099] confidence = scores[class_id]

[0101] if confidence > 0.5: # Extract only objects with a confidence level greater than 0.5

[0102] center_x = int(detection[0] * frame.shape[1])

[0103] center_y = int(detection[1] * frame.shape[0])

[0104] width = int(detection[2] * frame.shape[1])

[0105] height = int(detection[3] * frame.shape[0])

[0107] x = int(center_x - width / 2)

[0108] y = int(center_y - height / 2)

[0110] boxes.append([x, y, width, height])

[0111] confidences.append(float(confidence))

[0112] class_ids.append(class_id)

[0113] It can be implemented as follows.

[0115] And the truck height measuring device (200) extracts a truck image from the truck object information (S130).

[0116] An example of a method for extracting a truck image from such truck object information is described in more detail with reference to FIG. 6. After executing the code (S131), object information is received as input and extracted (S132). Then, non-maximum suppression is performed on the extracted objects to remove overlapping objects (S133), and iterative processing is performed on the results of non-maximum suppression (S134). Then, an index is extracted for each result (S135), truck image information is extracted using the extracted index (S136), and a label is generated based on the extracted information (S137).

[0117] An example of code for extracting a truck image from such truck object information is,

[0118] # Tracking trucks using extracted object information

[0119] indices = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4)

[0121] for i in indices:

[0122] i = i[0]

[0123] x, y, width, height = boxes[i]

[0124] label = str(classes[class_ids[i]

[0125] It can be implemented as follows.

[0127] Meanwhile, the truck height measuring device (200) preprocesses the extracted truck image to make it easier to measure the truck height (S140).

[0128] An example of a preprocessing method to facilitate measuring the truck height is described in more detail with reference to FIG. 7. After executing the code (S141), the input image is converted from BGR to grayscale (S141). Then, a Gaussian blur is applied to blur the image (S143), and edges are detected in the blurred image (S144). That is, after lowering the saturation, edges are detected in the image with lowered saturation and used to detect the boundary lines of the truck.

[0129] An example of code for extracting a truck image from such truck object information is,

[0130] When opening a video file or a real-time video stream,

[0131] cap = cv2.VideoCapture(0). (Here, # 0 is the index for the default webcam; you must change this index to use a different video source.)

[0132] Then, read the video frame and measure the height of the truck.

[0133] These truck height measurement codes are,

[0134] def measure_truck_height():

[0135] _, frame = cap.read() (# Read frame from video)

[0136] # The image preprocessing code for truck detection is,

[0137] gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

[0138] blurred = cv2.GaussianBlur(gray, (5, 5), 0)

[0139] edges = cv2.Canny(blurred, 50, 150)

[0140] It can be implemented as follows.

[0142] Next, the truck height measuring device (200) measures the truck height from the preprocessed truck image (S150).

[0143] An embodiment of the method for measuring the truck height is described in more detail with reference to FIG. 8. After executing the code (S151), a boundary detection code is executed in the edge image (S152). Then, a code to select the largest boundary among the detected boundaries is executed (S153), and a code to draw a rectangle on the selected boundary is executed (S154). After that, the height of the truck is measured in pixels based on the rectangular area of ​​the boundary (S155). Subsequently, a code to return the measured truck height is executed (S156), and the returned truck height is checked and the result is output (S157).

[0144] An example of a code for measuring such truck height is,

[0145] # Finding boundaries for truck detection

[0146] contours, _ = cv2.findContours(edges, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

[0148] # Select the largest boundary line (truck)

[0149] if len(contours) > 0:

[0150] truck_contour = max(contours, key=cv2.contourArea)

[0152] # Draw a rectangle on the border

[0153] x, y, w, h = cv2.boundingRect(truck_contour)

[0154] cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)

[0156] # Measure height in pixels

[0157] truck_height = h

[0158] return truck_height

[0160] return None

[0162] # Start measuring truck height

[0163] truck_height = measure_truck_height()

[0164] if truck_height is not None:

[0165] print("Truck height:", truck_height)

[0166] else:

[0167] print("Cannot find the truck.")

[0168] It can be implemented as follows.

[0169] The code above detects trucks in a real-time video stream and measures their height in pixels.

[0171] After measuring the truck height in this way, the truck height measuring device (200) compares the measured truck height with the height of the entrance / exit road in the driving direction (S160).

[0172] If, as a result of comparison, it is determined that it is dangerous for the truck to enter the access road (S170), the truck height measuring device (200) requests the safety device (300) to prohibit the truck from entering, and the safety device (300) performs an event to prohibit the truck from entering (S180). This entry prohibition event can be performed through one or more of a warning light (320), a barrier bar (330), and a display (340).

[0174] Although the present invention has been described with the examples above, the present invention is not necessarily limited to these examples and can be implemented with various modifications within the scope of the technical concept of the present invention. Accordingly, the examples disclosed in the present invention are intended to explain, not limit, the technical concept of the present invention, and the scope of the technical concept of the present invention is not limited by these examples. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention. Explanation of the symbols

[0176] 100 : Camera device 110, 120, 130 : Camera 200: Truck height measuring device 210: Communication unit 220: Program storage unit 230: Truck height measurement module 240: Access / Exit Route Height Data Storage Unit 250: Control Unit 300 : Safety device 310 : Communication unit 320 : Warning light 330 : Barrier 340 : Display 350 : Control unit

Claims

Claim 1 A camera device (100) that films one or more of the front, rear, and side of a moving vehicle using a CCTV camera; A truck height measuring device (200) is configured to detect a truck among vehicles in motion by utilizing 3D object detection and tracking technology on an image captured by the camera device (100), and to measure the truck height by combining Python, a computer vision library OpenCV, and a 3D object tracking library when measuring the height of the truck; wherein the camera device (100) is configured to photograph a truck, which is a vehicle in motion, using a CCTV camera, and is configured with a plurality of cameras (110, 120, 130) that photograph the front, rear, and side of the truck, which is a vehicle; and the truck height measuring device (200) comprises a communication unit (210) that receives image information of a vehicle in motion from the camera device (100) and transmits control information for operating the safety device to a safety device (300), a program storage unit (220) in which Python, a computer vision library OpenCV, and a 3D object tracking library are stored, and Python and a computer vision library OpenCV stored in the program storage unit (220) and The system is configured to include a truck height measurement module (230) that loads a library for 3D object tracking, preprocesses an image of a moving vehicle transmitted through the communication unit (210) to extract truck-related object information, extracts a truck image based on the extraction of truck object information from the truck-related object information, and measures the truck height after preprocessing the extracted truck image, an entrance / exit road height data storage unit (240) in which height data of an entrance / exit road installed in the direction in which a vehicle or truck is traveling is stored, and a control unit (250) that controls the communication unit (210), the program storage unit (220), and the truck height measurement module (230). The preprocessing of the truck image in the truck height measurement module (230) is performed by converting from BGR (Blue, Green, Red) to grayscale, andThe truck height measurement module (230) processes the truck including the cargo loaded on the truck by using a method of detecting the boundary lines of the truck through edge detection in the reduced image after applying a Gaussian blur to the image converted to grayscale to lower the saturation of the image, and the truck height measurement module (230) comprises a truck-related object information extraction unit (231) for extracting truck-related object information from an image of a moving vehicle transmitted through the communication unit (210), a truck object information extraction unit (232) for extracting truck object information from the truck-related object information extracted by the truck-related object information extraction unit (231), a truck image extraction unit (233) for extracting a truck image from the truck object information extracted by the truck object information extraction unit (232), a truck image preprocessing unit (234) for performing preprocessing on the truck image extracted by the truck image extraction unit (233), and a truck height measurement unit (235) for measuring the height of the truck from the truck image preprocessed by the truck image preprocessing unit (234). It is configured such that the control unit (250) compares the truck height measured by the truck height measuring module (230) with the height of the entrance / exit road in the direction of vehicle travel stored in the entrance / exit road height data storage unit (240), and transmits data indicating that the truck cannot pass through the entrance / exit road to the safety device (300) of the entrance / exit road with a height restriction in the direction of truck travel according to the truck height measured by the truck height measuring unit (235). The safety device (300) is configured to give a warning to the driver operating the truck according to the measurement result of the truck height measuring device (200), and includes a communication unit (310), a warning light (320), a barrier bar (330), a display (340), and a control unit (350). When data is received indicating that the height of the truck is higher than the entrance / exit road, the control unit (350) controls the warning light (320), barrier bar (330), and display (340) to warn the truck driver not to enter the entrance / exit road. The warning light is set to flash red, andA truck height measuring system characterized by directly prohibiting entry in front of the truck through the above-mentioned barrier (330) and displaying a notice prohibiting entry through the above-mentioned display (340). Claim 2 delete Claim 3 A method for measuring truck height using a truck height measuring system described in claim 1, comprising: a step (S100) of photographing the front, rear, and side of a vehicle in operation through a camera device (100) composed of a plurality of first to N cameras (110, 120, 130); a step (S110) of extracting truck-related object information from the photographed data in a truck height measuring device (200); a step (S120) of extracting truck object information from the truck-related object information in the truck height measuring device (200); a step (S130) of extracting a truck image from the truck object information in the truck height measuring device (200); and a step (S140) of performing preprocessing to reduce saturation in the truck image extracted in the truck height measuring device (200). In the above-mentioned preprocessed truck image, the truck height measuring device (200) measures the truck height (S150); the truck height measuring device (200) compares the measured truck height with the height of the entry / exit road in the driving direction (S160); if, as a result of the comparison, it is determined that it is dangerous for the truck to enter the entry / exit road (S170), the truck height measuring device (200) requests the safety device (300) to prohibit the truck from entering, and the safety device (300) proceeds with an event to prohibit the truck from entering (S180), wherein the step of extracting truck-related object information from the captured data in the truck height measuring device (200) (S110) includes a step (S111) in which code for 3D object detection and tracking is executed in a program storage unit (220) in which Python, a computer vision library OpenCV, and a library for 3D object tracking are stored, and the camera device (100 : 110, 120, The method comprises a step (S112) of preparing neural network input by preprocessing a captured image taken in 130), a step (S114) of configuring the layers and output settings of the neural network when the preprocessed captured image is input to the neural network (S113), a step (S115) of executing the neural network to perform object detection, and a step (S116) of processing the result of the execution to extract truck object information.The step (S120) of extracting truck object information from truck-related object information in the truck height measuring device (200) comprises: a step (S122) of repeating the output list of extracted truck-related object information and performing processing for each output; a step (S123) of repeating the extraction list and performing processing for each output detection; a step (S124) of extracting a score and a class from each object; a step (S125) of checking reliability based on the extracted score; a step (S126) of extracting object information when a preset level of reliability is satisfied; and a step (S127) of storing the extracted object information. The step (S130) of extracting a truck image from truck object information in the truck height measuring device (200) comprises: a step (S132) of receiving the truck object information as input and extracting it; a step (S133) of removing overlapping objects by performing non-maximum suppression on the extracted objects; and a step of performing iterative processing on the non-maximum suppression results. The method comprises a step (S134), a step of extracting an index for each result (S135), a step of extracting truck image information using the extracted index (S136), and a step of generating a label based on the extracted information (S137), and the step (S140) of performing preprocessing to lower the saturation of the truck image extracted from the truck height measuring device (200) includes a step (S142) of converting the image from BGR to grayscale. The method comprises a step of converting an image to be blurred by applying a Gaussian blur (S143) and a step of detecting edges to detect the boundary line of a truck in the blurred image (S144); and a step (S150) in which the truck height measuring device (200) measures the truck height in the preprocessed truck image comprises a step of executing a boundary line detection code in the edge image (S152), a step of executing a code to select the largest boundary line among the detected boundary lines (S153), and a step of executing a display code by drawing a rectangle on the selected boundary line (S154).A method for measuring truck height, characterized by comprising the steps of: measuring the height of the truck in pixels based on a rectangular area of ​​a boundary line (S155); executing a code to return the measured truck height (S156); and verifying the returned truck height and outputting the result (S157).

Images