CN102750544B - Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition - Google Patents

Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition Download PDF

Info

Publication number
CN102750544B
CN102750544B CN201210177961.0A CN201210177961A CN102750544B CN 102750544 B CN102750544 B CN 102750544B CN 201210177961 A CN201210177961 A CN 201210177961A CN 102750544 B CN102750544 B CN 102750544B
Authority
CN
China
Prior art keywords
module
seat belt
license plate
detection
plate
Prior art date
Application number
CN201210177961.0A
Other languages
Chinese (zh)
Other versions
CN102750544A (en
Inventor
尚凌辉
郑晓隆
于晓静
郑永宏
王弘玥
蒋宗杰
高勇
Original Assignee
浙江捷尚视觉科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江捷尚视觉科技股份有限公司 filed Critical 浙江捷尚视觉科技股份有限公司
Priority to CN201210177961.0A priority Critical patent/CN102750544B/en
Publication of CN102750544A publication Critical patent/CN102750544A/en
Application granted granted Critical
Publication of CN102750544B publication Critical patent/CN102750544B/en

Links

Abstract

一种基于车牌识别的未扣紧安全带违章驾驶检测系统,包括:车牌定位模块,车牌分割模块,车牌字符识别模块,人脸检测模块和安全带检测模块,所述人脸检测模块检测得到人脸后,往图像下方采用所述安全带检测模块对是否扣紧安全带进行检测。 Detection system based on one kind of illegal driving license plate recognition unfastened seat belt, comprising: positioning module plate, plate segmentation module, a license plate character recognition module, and seat belt face detection module detection module, a face detection module detects human give after the face, to the bottom of the belt using the image detection module detects whether the seat belt fastened. 在检测到人脸后,再检测是否扣紧安全带,提高了检测的精确性。 After a face is detected, and then detect whether the seat belt buckle, improve the accuracy of detection.

Description

基于车牌识别的未扣紧安全带违章驾驶检测系统及方法 Illegal driving belt license plate recognition based detection systems and methods of unfastened

技术领域 FIELD

[0001] 本发明属于交通视频检测领域,具体地是一种基于车牌识别的未扣紧安全带违章驾驶检测系统及其方法。 [0001] The present invention belongs to the field of traffic video detection, in particular a license plate recognition based in the unfastened seat belt system and method for detecting illegal driving.

背景技术 Background technique

[0002] 目前行车安全带检测主要依靠交警肉眼判别,而人工判断的准确性和有效性因人而异。 [0002] detecting current driving belt rely mainly on traffic visually identified, and human judgment accuracy and effectiveness vary. 采用视频或图像处理的方法,并且基于目前高清卡口的抓拍相机,可以有效的定位未系紧安全带的车主,帮助执法人员有效检索信息。 Or methods using video image processing, and based on the current high-definition camera to capture the bayonet, can effectively locate the owners did not fasten your seat belts, help law enforcement officers to effectively retrieve information.

[0003] 安全带检测主要依赖于车牌定位来找到车辆位置,并以此推算驾驶员的位置。 [0003] The belt detector depends on the positioning plate to locate the vehicle position, and thus the estimated position of the driver. 这种定位车辆的方法受环境影响小,一般能达到98%左右的识别率。 Such positioning of the vehicle by the method of small environmental impact, generally achieved recognition rate of about 98%. 但是安全带检测受相机架设角度和车内遮挡物影响,检测安全带准确率和人眼判断的准确率相当。 However, the camera set up by the belt angle and the detected vehicle impact shield, the detection of belt eye judgment accuracy and considerable accuracy. 只有人能够看请安全带佩戴的情况下,才能通过视频或图像进行检测。 Only people can look at your seat belt wearing situation can be detected by a video or image.

发明内容 SUMMARY

[0004] 本发明的基于车牌识别的未扣紧安全带违章驾驶检测系统主要分为两步,首先基于车牌识别定位车主,然后对车主是否佩戴安全带进行检测。 [0004] The license plate recognition based unfastened seat belt drivebadly detection system of the present invention is mainly divided into two steps, firstly the positioning plate recognition based on the owner, then the owner is detected whether or not wearing a seat belt. 主要运用的技术有车牌识别和安全带检测技术。 The main use of license plate recognition techniques and detection techniques belt.

[0005] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,包括:车牌定位模块,车牌分割模块,车牌字符识别模块,人脸检测模块和安全带检测模块,其特征在于:所述人脸检测模块检测得到人脸后,往图像下方采用所述安全带检测模块对是否扣紧安全带进行检测。 [0005] fastening the seat belt is not illegal driving license plate recognition based detection system, comprising: a positioning module plate, plate segmentation module, a license plate character recognition module, and seat belt face detection module detection module, wherein: the later the face to give the face detection module, to use the belt below the image detection module detects whether the seat belt fastened.

[0006] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌定位模块,车牌分割模块和车牌字符识别模块对车牌进行识别。 [0006] The license plate recognition based illegal driving unfastened seat belt detection system, wherein said plate positioning module, and a plate segmentation module License Plate Recognition module identification plate.

[0007] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌定位模块和人脸检测模块采用Adaboost分类算法,对图像中的车牌和人脸进行定位和检测。 [0007] The license plate recognition based illegal driving unfastened seat belt detection system, wherein said plate positioning unit and face detection module using Adaboost classification algorithm, the image of the face plate and the positioning and testing.

[0008] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌分割模块使用车牌定位模块的结果,采用基于字符二值化投影和车牌字符模板匹配方法对车牌进行分割。 [0008] The license plate recognition based illegal driving unfastened seat belt detection system, wherein the license plate segmentation module uses the results of the positioning module, using the divided character on the license plate and the plate projection binary character template matching based on .

[0009] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌字符识别模块采用车牌分割模块的结果,使用卷积神经网络算法。 [0009] The license plate recognition based illegal driving unfastened seat belt detection system, wherein the license plate character recognition module uses the results of the segmentation module, a convolutional neural network algorithm.

[0010] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述安全带检测模块采用基于Hough变换的直线检测算法。 [0010] The license plate recognition based illegal driving unfastened seat belt detection system, wherein said seat belt use detection module line detection algorithm based on the Hough transform.

[0011] 所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,还包括报警模块,根据所述安全带检测模块输出的结果,若有驾驶员或乘客未按规定佩戴安全带,则报警。 [0011] The license plate recognition based illegal driving unfastened seat belt detection system further comprises an alarm module, in accordance with the detection result of the belt module output, if the driver or passenger failing to wear a seat belt, then Call the police.

[0012] 本发明还提供了一种基于车牌识别的未扣紧安全带违章驾驶检测方法,包括如下步骤:对图像中的车辆车牌进行识别;对图像中的人脸进行检测;检测到人脸后,往图像下方对是否扣紧安全带进行检测。 [0012] The present invention further provides a license plate recognition based in the seat belt unfastened illegal driving detection method, comprising the steps of: vehicle license plate image is identified; face in the image is detected; detects faces after, below the image to detect whether or not fastening a seat belt.

附图说明 BRIEF DESCRIPTION

[0013] 图1是本基于车牌识别的未扣紧安全带违章驾驶检测系统的工作流程图; [0013] FIG. 1 is a flowchart of the license plate recognition based unfastened seat belt detection system illegal driving;

[0014] 图2是字符分割模块的字符分割模板示意图; [0014] FIG. 2 is a character segmentation character template at a division module;

[0015] 图3是本基于车牌识别的未扣紧安全带违章驾驶检测系统的检测示意图。 [0015] FIG. 3 is a schematic diagram of the detection of the license plate recognition based illegal driving unfastened seat belt detection system.

具体实施方式 Detailed ways

[0016] 如图1所示的安全带的检测流程,本发明的基于车牌识别的未扣紧安全带违章驾驶检测系统以高清相机拍摄的图像为基础,其包括车牌定位模块,车牌分割模块,车牌字符识别模块,人脸检测模块,安全带检测模块,报警模块。 [0016] inspection process belt shown in Figure 1, the present invention is based on the license plate image recognition system to detect illegal driving belt HD camera unfastened based positioning module which comprises a plate, plate segmentation module, Vehicle license plate recognition module, a face detection module, a belt detector module, alarm module.

[0017] 车牌定位模块和人脸检测模块采用了常用的Adaboost分类算法,对图像中的车牌和人脸进行定位和检测。 [0017] The location module plate and face detection module uses a common classification Adaboost algorithm on the image of the face plate and the positioning and testing.

[0018] 车牌分割模块使用车牌定位模块的结果,采用基于字符二值化投影和车牌字符模板匹配方法对车牌进行分割。 [0018] Results using the license plate location module segmentation module, based character binary character plate and a projection of the plate template matching method divides. 如图2所示的字符分割模板,采用以下步骤分割字符: Character segmentation template shown in FIG. 2, the divided character following steps:

[0019] 1)对车牌定位结果进行字符二值化。 [0019] 1) on the plate positioning result can be binarized character.

[0020] 2)对字符二值化结果进行水平坐标轴投影。 [0020] 2) character binarization result for the horizontal coordinate axis projection.

[0021] 3)不断在水平坐标轴上平移字符分割模板,并分别计算落入(al,bl),(a2,b2), (a3,b3), (a4,b4), (a5,b5), (a6,b6), (a7,b7)的二值化像素点个数,和落入(bl〜a2), (bl〜a2), (b2〜a3),(b3〜a4),(b4〜a5),(b5〜a6),(b6〜a7)的二值化像素点个数,计算两者比值最大的位置作为车牌字符分割位置。 [0021] 3) continuously translated along the horizontal axis divided character template, and calculates falls (al, bl), (a2, b2), (a3, b3), (a4, b4), (a5, b5) , (a6, b6), (a7, b7) the number of binarized pixels, and fall (bl~a2), (bl~a2), (b2~a3), (b3~a4), (b4 ~a5), (b5~a6), (b6~a7) binary number of pixels, calculated as the ratio of the two position of maximum plate character segmentation positions.

[0022] 车牌字符识别模块采用车牌分割模块的结果,使用了卷积神经网络算法,该卷积神经网络算法通过如下方式得到:采用反向传播(BP)算法,也就是通过理想输出值和实际输出值之间的差值来后向传递误差,并通过随机梯度法来修正权重。 [0022] Vehicle license plate recognition module uses the results of plate segmentation module, using a convolutional neural network algorithm, the convolutional neural network algorithm by: using the back-propagation (BP) algorithm, which is output by the ideal value and the actual the difference between the output value to the transmission error, and the corrected weight by stochastic gradient method. 具体的更新过程如下: Specific update process is as follows:

[0023] 对每个样本要优化的目标函数是其均方误差MSE : [0023] For each sample to be optimized objective function is the mean square error MSE:

[0024] [0024]

Figure CN102750544BD00051

[0025] 其中/是理想输出,其值和这个样本的类别有关,通常假定为只在向量的类别处输出为1,其他输出都是-1。 [0025] where / is the ideal output, and its category value for this sample, but is usually assumed to only one category at the output of the vector, the other output is -1. f是CNN的实际输出。 f is the actual output of CNN. Ei是每个样本的误差。 Ei is the error for each sample.

[0026] 卷积神经网络的权重更新采用的是梯度下降法,也就是按下式进行: [0026] Right convolutional neural network weights are updated using the gradient descent method, the following formula is carried out:

[0027] 其中W是卷积神经网络每个神经元的权重。 [0027] where W is the weight of each neuron convolutional neural network weights.

[0028] [0028]

Figure CN102750544BD00052

[0029] 所以要求出 [0029] Therefore, the requirements

Figure CN102750544BD00053

,也就是最终的误差对每个权重的导数,另外更新因子CT和二阶导 , I.e. the final error for each of the derivative weight, and further updates the second derivative factor CT

[0030] 数 [0030] number

Figure CN102750544BD00062

有关,所以关键是要求出各个层中权重和偏置的这两个导数。 For, it is a key requirement in a number of two conductive layers weight and weight of the individual bias.

Figure CN102750544BD00061

[0031] [0031]

[0032] 其中yi是单个神经元节点的输出结果,h是单个神经元关于权重1的偏置系数。 [0032] where yi is the output node of individual neurons, h is the individual neurons on the weight of the offset coefficient.

[0033] 对于最后一层,由公式(1)得到: [0033] For the final layer, obtained by the formula (1):

[0034] [0034]

Figure CN102750544BD00063

[0035] 其中卷积神经网络目标输出屯。 [0035] The convolutional neural network in which the target output Tun.

[0036] 由公式(3)得到: [0036] (3) obtained by the formula:

[0037] [0037]

Figure CN102750544BD00064

[0038] 由公式(2)得到: [0038] (2) obtained by the formula:

Figure CN102750544BD00065

[0039] [0039]

[0040] [0040]

[0041] 这样就可以由公式(4)导出公式(5),(6),(7)。 [0041] This can be derived by equation (4) Equation (5), (6), (7).

[0042] 同理对二阶导数: [0042] Similarly to the second derivative:

Figure CN102750544BD00066

[0043] [0043]

[0044] [0044]

[0045] 由于 [0045] Since the

Figure CN102750544BD00067

原函数、一阶导数和二阶导数波形如下,为了方便计算,令 Primitive, the first derivative and second derivative waveform as follows, for convenience of calculation, so that

Figure CN102750544BD00068

在理想的情况下,原函数的中间一条曲线应该是一条斜直线,所以这个假设是合理的。 In the ideal case, the original function of an intermediate curve should be a straight line inclined, so this is a reasonable assumption.

[0046] 车牌字符识别模块使用了卷积神经网络算法,这样可以对分割后的车牌进行识另IJ,得到图像中的车的车牌号码。 [0046] Vehicle license plate recognition module uses a convolutional neural network algorithm, which can identify the license plate of the divided IJ another, to obtain the license plate number of the vehicle in the image.

[0047] 基于人脸检测模块的检测结果,若检测成功得到人脸,则往图像下方继续采用安全带检测模块对是否扣紧安全带进行检测。 [0047] Based on the detection result of the face detection module, if the obtained face detection is successful, then continue to the belt below the image detection module to detect whether the seat belt fastened. 安全带检测模块采用基于Hough变换的直线检测算法。 Seatbelt detection module with linear detection algorithm based on the Hough transform. Hough变换是图像处理中从图像中识别几何形状的方法,Hough变换的基本原理在于利用点与线的对偶性,将原始图像空间的给定的曲线通过曲线表达形式变为参数空间的一个点。 Hough transform is a method of identifying an image processing from the image geometry, the basic principle is to use the Hough transform and the line point duality, given the curve form of the original image space into a curve expressing the point parameter space. 这样就把原始图像中给定曲线的检测问题转化为寻找参数空间中的峰值问题,也即把检测整体特性转化为检测局部特性,比如直线。 This put the original image given curve detection problem into finding the peak parameters of the problem space, i.e. the overall characteristics of the detector to detect a partial conversion characteristic, such as a straight line. 这样就可以把是否佩戴安全带给检测出来。 So that you can wear whether security to be detected.

[0048] 报警模块,根据安全带检测模块输出的结果,若有驾驶员或乘客未按规定佩戴安全带,则报警。 [0048] The alarm module, in accordance with the detection result of the belt module output, if the driver or passenger failing to wear a seat belt, the alarm.

[0049] 图3示出了该检测系统的检测实例。 [0049] FIG. 3 illustrates an example of detection of the detection system.

[0050] 1.首先由车牌定位模块,车牌分割模块和车牌字符识别模块确认车辆信息,如上图中车牌2的车牌号为"川BBW861"; [0050] 1. First, a plate positioning module, and a plate segmentation module confirm the vehicle license plate character recognition module information, such as the figure above the license plate No. 2 is "plain BBW861";

[0051] 2.识别成功时,人脸检测模块推算出车辆的大致位置,并以此检测车内人脸1 : [0051] 2. If a successful recognition, face detection module calculate the approximate location of the vehicle, and thereby detecting the car 1 Face:

[0052] 1)若检测不到人脸1 :则可能是图像质量不够清晰、反光或者相机架设角度不正确,放弃检测安全带。 [0052] 1) If a face is not detected 1: it may be the image quality is not clear, reflective or camera angle is not set up correctly, giving up the seat belt is detected.

[0053] 2)若检测到人脸1 :进入安全带检测模块。 [0053] 2) If a face is detected: a seat belt into the detection module.

[0054] 3.安全带检测模块在人脸正下方检测两条平行的斜线,并且两条斜线和车窗边缘相连。 [0054] 3. A seatbelt detection module detects the face just below the two parallel oblique lines, and the edge of the window and is connected to two oblique lines. 若检测到,则驾驶员佩戴了安全带3,反之作为未佩戴安全带3。 If detected, the driver is wearing the seat belt 3, otherwise not wearing a seat belt 3.

[0055] 4.若安全带检测模块检测到未佩戴安全带3,则报警模块报警。 [0055] 4. If the seatbelt detection module does not wear the seat belt 3, the warning alarm module.

[0056] 目前车牌识别在交通违法检测中应用很广泛,并受到较大关注。 [0056] In the present license plate recognition detection of illegal traffic in very broad application, and subject to greater attention. 加入安全带检测之后,能够有效的捕捉驾驶员是否系安全带的状态。 After the addition the seat belt detection, whether the driver is able to effectively capture the state of wearing a seatbelt. 特别是在检测到人脸后,再检测是否扣紧安全带,提高了检测的精确性。 Especially after a face is detected, and then detect whether the seat belt buckle, improve the accuracy of detection.

[0057] 在卡口的高清相机中添加车牌识别和安全带检测模块,可以有效的确认违法车辆,定位车主状态,并保存证据和报警。 [0057] Adding license plate recognition module, and seat belt detection bayonet HD camera can be effectively confirmed illegal vehicles, locating the owner status, and alarm and preserve evidence.

[0058] 以上所述及图中所示的仅是本发明的优选实施方式。 [0058] are only preferred embodiments of the invention described above and illustrated in FIG. 应当指出,对于本领域的普通技术人员来说,在不脱离本发明的前提下,还可以作出若干变型和改进,这些也应视为属于本发明的保护范围。 It should be noted that those of ordinary skill in the art, without departing from the present invention in the premise, but also may be made to a number of modifications and improvements, which should be considered as being within the scope of the present invention.

Claims (5)

1. 一种基于车牌识别的未扣紧安全带违章驾驶检测系统,包括:车牌定位模块,车牌分割模块,车牌字符识别模块,人脸检测模块和安全带检测模块,其特征在于:所述人脸检测模块检测得到人脸(1)后,往图像下方采用所述安全带检测模块对是否扣紧安全带(3) 进行检测,其中所述车牌定位模块,车牌分割模块和车牌字符识别模块对车牌(2)进行识另IJ;所述车牌字符识别模块采用车牌分割模块的结果,使用卷积神经网络算法,该卷积神经网络算法通过如下方式得到:采用反向传播(BP)算法,也就是通过理想输出值和实际输出值之间的差值来后向传递误差,并通过随机梯度法来修正权重,具体的更新过程如下: 对毎个样本要优化的目标函数是其询方误差MSE: 1. Based on an unfastened seat belt driving detecting illegal license plate recognition system, comprising: a positioning module plate, plate segmentation module, a license plate character recognition module, and seat belt face detection module detection module, wherein: the human after the obtained face detection module detects a face (1), to the use of the belt below the image detection module determines whether the seat belt fastening (3) is detected, wherein said plate positioning module, and a license plate character recognition module segmentation module plate (2) for identifying another IJ; the license plate character recognition module uses the results of the segmentation module, a convolutional neural network algorithm, the convolutional neural network algorithm by: using the back-propagation (BP) algorithm, also that it is, by the difference between the desired output and the actual output value to a transmission error, and the corrected weight by a stochastic gradient method, and the specific update process is as follows: every sample of the objective function to be optimized is the inquiry square error MSE :
Figure CN102750544BC00021
其中V是理想输出,其值和这个样本的类别有关,通常假定为只在向量的类别处输出为1,其他输出都是-1,Y是CNN的实际输出,Ei是每个样本的误差; 卷积神经网络的权重更新采用的是梯度下降法,也就是按下式进行: 其中W是卷积神经网络每个神经元的权重; Wherein V is the ideal output, and its category value for this sample, but is usually assumed to only one category at the output of the vector, the other output is -1, Y is CNN's actual output, the error Ei of each sample; right convolutional neural network weights are updated using the gradient descent method, is performed following formula: wherein W is a convolutional neural network weights of each neuron weight;
Figure CN102750544BC00022
所以要求出 So we requested a
Figure CN102750544BC00023
,也就是最终的误差对每个权重的导数,另外更新因子《和二阶导数 , I.e. the final error derivative of each weight, in addition to updating factor "and the second derivative
Figure CN102750544BC00024
有关,所以关键是要求出各个层中权重和偏置的这两个导数; For, the key requirement is that a number of guiding two respective layers of the weights and bias;
Figure CN102750544BC00025
其中yi是单个神经元节点的输出结果,h是单个神经元关于权重I的偏置系数; 对于最后一层,由公式(1)得到: Wherein yi is the output node of individual neurons, h is the individual neurons on the weight coefficients I bias; for the last layer, obtained by the formula (1):
Figure CN102750544BC00026
其中卷积神经网络目标输出屯; 由公式(3)得到: Wherein the target output Tun convolutional neural network; (3) obtained by the formula:
Figure CN102750544BC00027
由公式(2)得到: (2) obtained by the formula:
Figure CN102750544BC00031
这样就可以由公式(4)导出公式(5),(6),(7); 同理对二阶导数: This can be derived from Equation (4) Equation (5), (6), (7); Similarly to the second derivative:
Figure CN102750544BC00032
由于 due to
Figure CN102750544BC00033
令,0 « 〇,在理想的情况下,原函数的中间一条曲线应该是一条斜直线,所以这个假设是合理的。 Order, 0 «square, in the ideal case, the original function of an intermediate curve should be a straight line inclined, so this is a reasonable assumption.
2. 如权利要求1所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌定位模块和人脸检测模块采用Adaboost分类算法,对图像中的车牌和人脸进行定位和检测。 As claimed in claim 1 based unfastened seat belt driving detecting illegal license plate recognition system, wherein said plate positioning unit and face detection module using Adaboost classification algorithm, and the image of the face plate and positioning requirements detection.
3. 如权利要求2所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述车牌分割模块使用车牌定位模块的结果,采用基于字符二值化投影和车牌字符模板匹配方法对车牌进行分割。 As claimed in claim 2 based on the license plate recognition illegal driving unfastened seat belt detection system, wherein the license plate segmentation module using the result of the positioning module, based character binary character template matching projection and plate method license plate segmentation.
4. 如权利要求1所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,其中所述安全带检测模块采用基于Hough变换的直线检测算法。 As claimed in claim 1, said license plate recognition based illegal driving unfastened seat belt detection system, wherein said seat belt use detection module line detection algorithm based on the Hough transform.
5. 如权利要求1所述的基于车牌识别的未扣紧安全带违章驾驶检测系统,还包括报警模块,根据所述安全带检测模块输出的结果,若有驾驶员或乘客未按规定佩戴安全带(3), 则报警;若检测到未扣紧安全带,则报警。 Alarm module as claimed in claim 5, according to the detection result of the belt module output, if the driver or passenger fails to wear safety regulations according to a license plate recognition based illegal driving unfastened seat belt detection system, further comprising band (3), the alarm; unfastened seat belt if it is detected, the alarm.
CN201210177961.0A 2012-06-01 2012-06-01 Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition CN102750544B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210177961.0A CN102750544B (en) 2012-06-01 2012-06-01 Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210177961.0A CN102750544B (en) 2012-06-01 2012-06-01 Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition

Publications (2)

Publication Number Publication Date
CN102750544A CN102750544A (en) 2012-10-24
CN102750544B true CN102750544B (en) 2015-01-14

Family

ID=47030709

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210177961.0A CN102750544B (en) 2012-06-01 2012-06-01 Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition

Country Status (1)

Country Link
CN (1) CN102750544B (en)

Families Citing this family (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102915640B (en) * 2012-10-30 2015-06-17 武汉烽火众智数字技术有限责任公司 Safety belt detecting method based on Hough transform
CN103021179B (en) * 2012-12-28 2016-03-30 佛山市华电智能通信科技有限公司 Seatbelt detection method based on real-time monitoring video
CN103077611A (en) * 2013-01-06 2013-05-01 山东鼎讯智能交通科技有限公司 Road vehicle recording system with face identification and safety belt detection functions
CN103150556B (en) * 2013-02-20 2015-09-30 西安理工大学 The belt detection methods for road traffic monitoring
CN104182960B (en) * 2013-05-22 2017-03-29 浙江大华技术股份有限公司 One kind of a driver seat belt wearing detecting method and device
CN103279759B (en) * 2013-06-09 2016-06-01 大连理工大学 One kind of analysis can pass in front of the vehicle based on convolutional neural network
CN103522982B (en) * 2013-10-25 2016-09-07 公安部第三研究所 Image analysis based on three-point vehicle seat belt apparatus and a method for detecting
CN103679208A (en) * 2013-11-27 2014-03-26 北京中科模识科技有限公司 Broadcast and television caption recognition based automatic training data generation and deep learning method
CN104700066B (en) * 2013-12-11 2018-02-16 杭州海康威视数字技术股份有限公司 It is a kind of detect driver whether wear safety belt method and apparatus
CN103914841B (en) * 2014-04-03 2018-03-09 深圳大学 Based on the segmentation of the vaginal bacteria of super-pixel and deep learning and categorizing system
CN103955704B (en) * 2014-04-26 2018-08-24 合肥工业大学 A kind of Safe belt detection method based on Adaboost
EP3149611A4 (en) * 2014-05-27 2017-08-09 Beijing Kuangshi Technology Co., Ltd. Learning deep face representation
CN104036323B (en) * 2014-06-26 2016-11-09 叶茂 Vehicle detection method based on convolutional neural network
CN104063719B (en) * 2014-06-27 2018-01-26 深圳市赛为智能股份有限公司 Pedestrian detection method and device based on depth convolutional network
CN104156717A (en) * 2014-08-31 2014-11-19 王好贤 Method for recognizing rule breaking of phoning of driver during driving based on image processing technology
CN104700068A (en) * 2014-12-17 2015-06-10 安徽清新互联信息科技有限公司 SVM based detection method of safety belt of driver
CN105488453B (en) * 2015-11-30 2019-03-26 杭州全实鹰科技有限公司 A kind of driver based on image procossing does not fasten the safety belt detection recognition method
CN105354572B (en) * 2015-12-10 2018-10-12 苏州大学 A kind of automatic license plate identification system based on simplified convolutional neural networks
CN105718912B (en) * 2016-01-26 2018-12-07 浙江捷尚视觉科技股份有限公司 A kind of vehicle characteristics object detecting method based on deep learning
CN105740910A (en) * 2016-02-02 2016-07-06 北京格灵深瞳信息技术有限公司 Vehicle object detection method and device
CN107292222A (en) * 2016-04-01 2017-10-24 杭州海康威视数字技术股份有限公司 A kind of vehicle peccancy detection method and device

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101398894B (en) * 2008-06-17 2011-12-07 浙江师范大学 License Plate Recognition Method and apparatus for implementing automatic
CN101436348B (en) * 2008-12-22 2014-04-23 北京中星微电子有限公司 Electronic policeman system and information identification method applied to the same
CN101645172A (en) * 2009-09-09 2010-02-10 北京理工大学 Rapid detection method for straight line in digital image

Also Published As

Publication number Publication date
CN102750544A (en) 2012-10-24

Similar Documents

Publication Publication Date Title
Grubb et al. 3D vision sensing for improved pedestrian safety
US8379924B2 (en) Real time environment model generation system
Zhou et al. A novel lane detection based on geometrical model and gabor filter
Keller et al. The benefits of dense stereo for pedestrian detection
US8812226B2 (en) Multiobject fusion module for collision preparation system
US7639841B2 (en) System and method for on-road detection of a vehicle using knowledge fusion
US20040252863A1 (en) Stereo-vision based imminent collision detection
Yuan Video-based smoke detection with histogram sequence of LBP and LBPV pyramids
TWI384408B (en) Method and system for identifying image and outputting identification result
US20050232463A1 (en) Method and apparatus for detecting a presence prior to collision
CN101023436B (en) Method for traffic sign detection
Faro et al. Adaptive background modeling integrated with luminosity sensors and occlusion processing for reliable vehicle detection
Sivaraman et al. A review of recent developments in vision-based vehicle detection
US8885929B2 (en) Abnormal behavior detection system and method using automatic classification of multiple features
US9824277B2 (en) Pedestrian right of way monitoring and reporting system and method
Guo et al. Pedestrian detection for intelligent transportation systems combining AdaBoost algorithm and support vector machine
Gerónimo et al. 2D–3D-based on-board pedestrian detection system
CN101303735B (en) Method for detecting moving objects in a blind spot region of a vehicle and blind spot detection device
John et al. Traffic light recognition in varying illumination using deep learning and saliency map
CN101630361A (en) Plate number, body color and mark identification-based equipment and plate number, body color and mark identification-based method for identifying fake plate vehicles
US9429650B2 (en) Fusion of obstacle detection using radar and camera
TWI381321B (en) Method for image recombination of multiple images and identifying image and system for identifying image and outputting identification result
CN101540105A (en) Fake-licensed car detection method based on number-plate identification and gridding supervision
CN102096803A (en) Safe state recognition system for people on basis of machine vision
CN102426786B (en) Intelligent video analyzing system and method for automatically identifying fake plate vehicle

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
CB02 Change of applicant information

Address after: Hangzhou City, Zhejiang province Yuhang District 311121 West Street Wuchang No. 998 building 7 East

Applicant after: Zhejiang iCare Vision Technology Co., Ltd.

Address before: 310013, Zhejiang, Xihu District, Hangzhou, Tian Shan Road, No. 398, Kun building, 4 floor, South Block

Applicant before: Zhejiang iCare Vision Technology Co., Ltd.

C53 Correction of patent for invention or patent application
CB02 Change of applicant information

Address after: Hangzhou City, Zhejiang province Yuhang District 311121 West Street Wuchang No. 998 building 7 East

Applicant after: ZHEJIANG ICARE VISION TECHNOLOGY CO., LTD.

Address before: Hangzhou City, Zhejiang province Yuhang District 311121 West Street Wuchang No. 998 building 7 East

Applicant before: Zhejiang iCare Vision Technology Co., Ltd.

C53 Correction of patent for invention or patent application
COR Change of bibliographic data

Free format text: CORRECT: APPLICANT; FROM: HANGZHOU ICARE VISION TECHNOLOGY CO., LTD. TO: ZHEJIANG ICARE VISION TECHNOLOGY CO., LTD.

GR01 Patent grant
C14 Grant of patent or utility model
PE01 Entry into force of the registration of the contract for pledge of patent right
PE01 Entry into force of the registration of the contract for pledge of patent right

Denomination of invention: Detection system and detection method of rule-breaking driving that safety belt is not fastened and based on plate number recognition

Effective date of registration: 20190821

Granted publication date: 20150114

Pledgee: Hangzhou Yuhang Small and Medium-sized Enterprise Transfer Service Co., Ltd.

Pledgor: ZHEJIANG ICARE VISION TECHNOLOGY CO., LTD.

Registration number: Y2019330000020