WO2020133983A1 - 一种信号灯的识别方法及装置、电子设备 - Google Patents

一种信号灯的识别方法及装置、电子设备 Download PDF

Info

Publication number
WO2020133983A1
WO2020133983A1 PCT/CN2019/093201 CN2019093201W WO2020133983A1 WO 2020133983 A1 WO2020133983 A1 WO 2020133983A1 CN 2019093201 W CN2019093201 W CN 2019093201W WO 2020133983 A1 WO2020133983 A1 WO 2020133983A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
reference image
images
information
signal light
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2019/093201
Other languages
English (en)
French (fr)
Inventor
杨洁
何东杰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Unionpay Co Ltd
Original Assignee
China Unionpay Co Ltd
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 China Unionpay Co Ltd filed Critical China Unionpay Co Ltd
Publication of WO2020133983A1 publication Critical patent/WO2020133983A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/10—Image acquisition

Definitions

  • the invention relates to the technical field of data processing, in particular to a signal light identification method and device, and electronic equipment.
  • Supplementing the image information mainly involves methods such as matching or repair.
  • the repair method mainly extracts information such as frequency domain or gray scale based on the correlation between pixels, and analyzes the continuity to make predictions.
  • the problem of supplementary information in discrete scenes of signal lights cannot be solved.
  • the algorithm represented by SIFT Scale-invariant feature transform
  • the algorithm represented by SIFT directly performs Gaussian scale space processing on the image, extracts the main direction of the relevant points to form the feature descriptor, and then passes the European The distance calculation is used for matching.
  • This kind of method can match the pixels in the image with strong correlation and can be matched well.
  • the present application provides a signal light recognition method and device, and electronic equipment, to solve the problem that the image supplement method in the prior art cannot supplement signal image information in a discrete scene.
  • the reference image is any one of the N images, and the reference image is the N images except the reference image Any one; the signal lamp to be supplemented is any signal lamp other than the reference signal lamp in the reference image;
  • the signal light information of the fourth position is determined according to the signal light information of at least one third position.
  • the number of the reference signal lamps is at least two;
  • the determination of the reference signal lamp includes:
  • the signal lights that are all in the lit state in the N images are selected as the reference signal lights, and the two reference signal lights in the same image and the signal lights to be supplemented are not located on the same straight line.
  • the at least a first position of the reference signal lamp in the reference image, a second position of the reference signal lamp in the reference image, and a signal lamp to be supplemented in the reference image including:
  • the fourth position is determined according to the third position and the position mapping relationship.
  • the determining the signal light information of the fourth position according to the signal light information of at least one third position includes:
  • the number of sub-information with the same content among the N-1 sub-information corresponding to the N-1 reference images is counted; the sub-information with the largest number is used as the sub-information of the signal light information in the fourth position.
  • the method further includes:
  • the N images are N frames of images in the same video stream, and the reference image is the first frame of the video stream.
  • An embodiment of the present invention also provides a signal light recognition device, including:
  • An obtaining unit used to obtain N images taken for the same group of signal lights within a preset period; each image is used to identify the signal lights in the lit state;
  • a determining unit configured to determine a reference signal light, all of which are in the N images
  • the calculation unit is configured to determine, for each reference image, at least the first position of the reference signal lamp in the reference image and the second position of the reference signal lamp in the reference image for the supplementary signal lamp in the reference image.
  • the reference image is any one of the N images, and the reference image is the N Any one of the images except the reference image;
  • the supplementary signal lamp is any signal lamp other than the reference signal lamp in the reference image;
  • the supplementary unit is configured to determine the signal light information of the fourth position according to the signal light information of at least one third position.
  • the number of the reference signal lamps is at least two;
  • the determining unit is specifically used for:
  • the signal lights that are all in the lit state in the N images are selected as the reference signal lights, and the two reference signal lights and the supplementary signal lights in the same image are not located on the same straight line.
  • the calculation unit is specifically used to:
  • the fourth position is determined according to the third position and the position mapping relationship.
  • the supplementary unit is specifically used for:
  • the number of sub-information with the same content among the N-1 sub-information corresponding to the N-1 reference images is counted; the sub-information with the largest number is used as the sub-information of the signal light information in the fourth position.
  • it further includes a processing unit, which is used to:
  • the N images are N frames of images in the same video stream, and the reference image is the first frame of the video stream.
  • An embodiment of the present invention also provides an electronic device, including:
  • At least one processor and,
  • a memory communicatively connected to the at least one processor; wherein,
  • the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method as described above.
  • An embodiment of the present invention further provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute the method as described above.
  • N images captured for the same group of signal lights within a preset period of time are acquired; each image is used to identify a signal light in a lit state.
  • the signal lights existing in all N images are used as reference signal lights.
  • One of the N images is used as a reference image, and any one of the N images other than the reference image is used as a reference image.
  • For each reference image based on the first position of the reference signal in the reference image, the second position of the reference signal in the reference image, and the third position of the supplementary signal in the reference image, determine the first position of the signal to be supplemented in the reference image Four positions.
  • the signal lamp to be supplemented is any signal lamp other than the reference signal lamp in the reference image.
  • the signal information of the fourth position is determined according to the signal information of the at least one third position.
  • the embodiments of the present invention compare images taken at different angles and different time points, and effectively use the positional characteristics of the distribution of signal lights to associate the positions of the same signal lights in different images, so that the signal light information in different images can complement each other and facilitate Identify all the signal lights that are lit, the calculation is small, the method is simple and efficient.
  • FIG. 1 is a schematic flowchart of a method for identifying a signal light provided by an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a binary image provided by an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of the arrangement of signal lights of a cabinet in an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of the mapping relationship between the position of the signal lamp in the reference image and the signal lamp in the reference image in an embodiment of the present invention
  • FIG. 5 is a schematic flowchart of a method for identifying a signal light provided by a specific embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a signal light identification device according to an embodiment of the present invention.
  • FIG. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
  • the intelligent inspection robot in the computer room is a key frontier innovative technology that major financial companies need to conduct in-depth research and application. After understanding, at present, large domestic banks have started the research of intelligent inspection robots in the computer room.
  • the intelligent inspection robot recognizes the status of the server's signal light by shooting or taking pictures of the server, thereby judging the status of the server.
  • an image may not accurately reflect the information of all signal lights. For example, during the preset time, some of the signal lights are actually on, but because the time point of capturing an image coincides with the time point of the flashing of the signal light, the state of the signal light in some images will appear as Not bright. Therefore, multiple images need to complement each other.
  • an embodiment of the present invention provides a signal light recognition method.
  • the signal light recognition method provided by the embodiment of the present invention includes the following steps:
  • Step 101 Acquire N images captured for the same group of signal lights within a preset period of time; each image is used to identify a signal light in a lit state.
  • Step 102 Determine a reference signal light, which is included in the N images.
  • Step 103 For each reference image, at least according to the first position of the reference signal lamp in the reference image, the second position of the reference signal lamp in the reference image, and the first position of the signal lamp to be supplemented in the reference image Three positions to determine the fourth position of the signal light to be supplemented in the reference image; the reference image is any one of the N images, and the reference image is the N images except the reference Any one other than the image; the supplementary signal light is any signal light other than the reference signal light in the reference image.
  • Step 104 Determine the signal light information of the fourth position according to the signal light information of at least one third position.
  • N images captured for the same group of signal lights within a preset period of time are acquired; each image is used to identify a signal light in a lit state.
  • the signal lights existing in all N images are used as reference signal lights.
  • One of the N images is used as a reference image, and any one of the N images other than the reference image is used as a reference image.
  • the third position of the supplementary signal in the reference image is determined based on the first position of the reference signal in the reference image and the second position of the reference signal in the reference image, and the reference image is the same as the supplementary signal in the reference image
  • the fourth position of the signal lamp to be supplemented is any signal lamp other than the reference signal lamp in the reference image.
  • the signal information of the fourth position is determined according to the signal information of the at least one third position.
  • the embodiments of the present invention compare images taken at different angles and different time points, and effectively use the location characteristics of the signal lights to associate the positions of the same signal lights in different images, so that the signal light information in different images can complement each other, which is convenient Identify all the signal lights that are lit, the calculation is small, the method is simple and efficient.
  • the embodiment of the invention shoots the signal lights on the server cabinet, and shoots N images for the same group of signal lights.
  • the above step 101 after acquiring N images captured for the same group of signal lights within a preset period, further includes:
  • the signal lights in each image are initially identified, and the initial position of each signal light is determined.
  • Binary processing is performed on the image, that is, only the position information of the signal light is retained on the blackboard to form a binary image with black dots and white dots, as shown in FIG. 2.
  • Using the binarized image for identification can effectively enhance the characteristics of the signal light and facilitate noise reduction. Because the signal light is easily confused with impurities, no noise reduction process will reduce the accuracy rate. Therefore, after the binarization process, the Gaussian window is also used to remove noise from the image.
  • gradient calculation is used to determine the exact position of the signal lights in the binarized image.
  • each pixel point with the center of gravity of the signal lights is used to replace the corresponding signal lights for analysis and processing.
  • the number of reference signal lamps is at least 2.
  • the determination of the reference signal lamp includes:
  • the signal lights that are all in the lit state in the N images are selected as the reference signal lights, and the two reference signal lights in the same image and the signal lights to be supplemented are not located on the same straight line.
  • the signal lights of the cabinet are arranged as shown in Figure 3. Due to the shooting angle and other reasons, the position of the same signal light in different images is different, and due to the deformation of the image, the positional relationship between the signal lights is also different in different images. Therefore, it is necessary to establish the positional mapping relationship of the signal lights between different images. It is convenient to determine the exact position of the signal lamp to be supplemented, so as to find the signal lamp to be supplemented in the reference image.
  • the function of the reference signal lamp is to establish the position mapping relationship between the signal lamp in the reference image and the signal lamp in the reference image.
  • the embodiment of the present invention selects the signal lights that are all lit in the N images as the reference signal lights, and the two reference signal lights and the signal lights to be supplemented in the same image are not located on the same straight line. This is because if the reference signal lamp and the signal lamp to be supplemented are located on the same straight line, the deformation in different directions on the image cannot be reflected.
  • Determining the fourth position of the signal light to be supplemented in the reference image includes:
  • the fourth position is determined according to the third position and the position mapping relationship.
  • one of the N images is selected as the reference image, and the remaining N-1 images are all reference images.
  • a simpler way is to establish a coordinate system in each image to determine the position of each signal lamp.
  • Position coordinates use the coordinates to identify the position of the corresponding signal light.
  • FIG. 4 shows the mapping relationship between the signal lights in the reference image and the signal lights in the reference image.
  • Signal lamp A and signal lamp B are reference signal lamps
  • signal lamp C is a signal lamp to be supplemented.
  • a 1 , B 1 and C 1 are the position of the semaphore in the reference image
  • a 2 , B 2 and C 2 are the position of the semaphore in the reference image.
  • a 1 corresponds to A 2 and corresponds to signal lamp A
  • B 1 corresponds to B 2 and corresponds to signal lamp B
  • C 1 corresponds to C 2 and corresponds to signal lamp C.
  • the position of the signal lamp to be supplemented in the reference image is determined, that is, C 2 is determined according to A 1 , B 1 , C 1 , A 2 and B 2 .
  • C 2 is determined according to A 1 , B 1 , C 1 , A 2 and B 2 .
  • the signal lights A, B, and C in the same image form a triangle.
  • mapping relation between A 1 A 2 by the coordinates A A 1 and 2 the mapping between 1 B 2 B given by the coordinates B B 1 and 2 relationship.
  • the mapping relationship between A 1 A 2 and the mapping relationship between B 1 B 2 the mapping relationship between C 1 C 2 can be obtained, and then knowing the coordinates of C 1 , the C 2 in the reference image can be determined coordinate.
  • step 104 the determining the signal light information of the fourth position based on the signal light information of at least one third position includes:
  • the number of sub-information with the same content among the N-1 sub-information corresponding to the N-1 reference images is counted; the sub-information with the largest number is used as the sub-information of the signal light information in the fourth position.
  • each semaphore information includes M sub-information, for each sub-information, count the number of each seed information content in N-1 reference images, and sort them, and select the largest number of sub-information content as the signal to be added in the reference image Content of the child information.
  • the sub-information is the status, color, and server of the semaphore. According to the status of the semaphore, the sub-information content is bright and unlit.
  • the sub-information content of the semaphore to be added is not bright, and the remaining 39 reference images have the sub-signal of the semaphore to be added.
  • the information content is bright.
  • the content of the status of the sub-information signal lamp is selected to be bright, that is, the signal lamp state of the signal lamp to be supplemented in the reference image is marked as bright.
  • the sub-information content has three types: red, yellow, and green.
  • the sub-information content of the signal lamp to be supplemented is red in 9 reference images
  • the sub-information content of the signal lamp to be supplemented is green in 15 reference images
  • the information of the to-be-supplied signal lamp in the remaining 25 reference images is If the content of the sub-information is yellow, the content of the color of the sub-information signal lamp is selected as yellow, that is, the color of the signal lamp of the signal lamp to be supplemented in the reference image is marked as yellow. In this way, each signal light to be supplemented in the reference image is traversed, and the signal light information to be supplemented in the reference image is supplemented with the signal information of the remaining N-1 reference images.
  • the above N images are N frames of images in the same video stream, and the reference image is the first frame of the video stream. Except for the first frame, the remaining frame images serve as reference images.
  • Step 401 Acquire N images captured for the same set of signal lights within a preset period of time.
  • Step 402 Select one of the N images as the reference image, and the remaining N-1 images as the reference image.
  • Step 403 Perform binarization processing on each image and perform denoising.
  • Step 404 Use gradient calculation to find the position of the signal light in the binarized image, and use the center of gravity of the signal light to replace the signal light for analysis and processing to determine the position of the center of gravity of the signal light.
  • Step 405 Use two signal lights that can be recognized in all images as reference signal lights.
  • Step 406 For a signal light to be identified, establish a mapping relationship between the signal light to be identified in the reference image and the signal light to be identified in any reference image.
  • the position mapping relationship is established based on the position coordinates of the reference signal lamp in the reference image, the position coordinates of the reference signal lamp in the reference image, and the position coordinates of the reference signal lamp in the reference image.
  • Step 407 Count the number of sub-information contents of the signal lamp to be identified in all reference images, and select the sub-information contents with the largest number as the sub-information contents of the signal lamp to be identified in the reference image.
  • Step 408 Traverse each signal light to be identified in the reference image, and determine the position and signal information of each signal light to be identified.
  • An embodiment of the present invention also provides a signal light recognition device, as shown in FIG. 6, including:
  • the obtaining unit 501 is used to obtain N images taken for the same group of signal lights within a preset period; each image is used to identify the signal lights in the lit state;
  • the determining unit 502 is configured to determine a reference signal lamp, and the reference signal lamps are included in the N images;
  • the calculation unit 503 is configured to determine, for each reference image, a supplementary signal light in the reference image according to at least the first position of the reference signal light in the reference image and the second position of the reference signal light in the reference image The third position of the reference image, and the fourth position of the signal light to be supplemented that is the same signal light as the supplementary signal light in the reference image;
  • the reference image is any one of the N images, and the reference image is the Any one of the N images except the reference image;
  • the supplementary signal lamp is any signal lamp other than the reference signal lamp in the reference image;
  • the supplementing unit 504 is configured to determine the signal light information of the fourth position according to the signal light information of at least one third position.
  • the number of the reference signal lights is at least two;
  • the determining unit 502 is specifically used for:
  • the signal lights that are all in the lit state in the N images are selected as the reference signal lights, and the two reference signal lights and the supplementary signal lights in the same image are not located on the same straight line.
  • calculation unit 503 is specifically used to:
  • the fourth position is determined according to the third position and the position mapping relationship.
  • supplement unit 504 is specifically used to:
  • the number of sub-information with the same content among the N-1 sub-information corresponding to the N-1 reference images is counted; the sub-information with the largest number is used as the sub-information of the signal light information in the fourth position.
  • processing unit 505 for:
  • the N images are N frames of images in the same video stream, and the reference image is the first frame of the video stream.
  • the present invention also provides an electronic device, as shown in FIG. 7, including:
  • It includes a processor 601, a memory 602, a transceiver 603, and a bus interface 604, where the processor 601, the memory 602, and the transceiver 603 are connected through a bus interface 604;
  • the processor 601 is configured to read the program in the memory 602 and execute the following methods:
  • the reference image is any one of the N images, and the reference image is the N images except the reference image Any one; the signal lamp to be supplemented is any signal lamp other than the reference signal lamp in the reference image;
  • the signal light information of the fourth position is determined according to the signal light information of at least one third position.
  • the number of the reference signal lights is at least two; the processor 601 is specifically used to:
  • the signal lights that are all in the lit state in the N images are selected as the reference signal lights, and the two reference signal lights in the same image and the signal lights to be supplemented are not located on the same straight line.
  • processor 601 is specifically used to:
  • the fourth position is determined according to the third position and the position mapping relationship.
  • processor 601 is specifically used to:
  • the number of sub-information with the same content among the N-1 sub-information corresponding to the N-1 reference images is counted; the sub-information with the largest number is used as the sub-information of the signal light information in the fourth position.
  • processor 601 is also used to:
  • each flow and/or block in the flowchart and/or block diagram and a combination of the flow and/or block in the flowchart and/or block diagram may be implemented by computer program instructions.
  • These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processing machine, or other programmable data processing device to produce a machine that enables the generation of instructions executed by the processor of the computer or other programmable data processing device
  • These computer program instructions may also be stored in a computer readable memory that can guide a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, the instructions
  • the device implements the functions specified in one block or multiple blocks of the flowchart one flow or multiple flows and/or block diagrams.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operating steps are performed on the computer or other programmable device to produce computer-implemented processing, which is executed on the computer or other programmable device
  • the instructions provide steps for implementing the functions specified in one block or multiple blocks of the flowchart one flow or multiple flows and/or block diagrams.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

本发明实施例涉及数据处理技术领域，尤其涉及一种信号灯的识别方法及装置、电子设备，用以解决现有技术中的图像补充方法无法进行离散场景中信号灯图像信息补充的问题。本发明实施例包括：获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；确定参考信号灯，所述N张图像中均有所述参考信号灯；针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置；根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。

Description

一种信号灯的识别方法及装置、电子设备
相关申请的交叉引用
本申请要求在2018年12月29日提交中国专利局、申请号为201811647448.7、申请名称为“一种信号灯的识别方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及数据处理技术领域，尤其涉及一种信号灯的识别方法及装置、电子设备。
背景技术
在识别服务器信号灯时，由于信号灯闪烁和拍摄角度的原因，一张图像不能够获取和反应所有的信息，或者有些图像中显示的信息不够清晰，需通过多张图像进行相互补充，将一张图像中有的信息补充到另外一张图像中相应的位置上。
对图像信息进行补充主要涉及匹配或修复等方法，现有技术中修复方法主要依据像素之间的关联性，提取频域或者灰度等信息，分析连续性进行预测。无法解决信号灯离散场景中的信息补充问题。而图像匹配方法中，以SIFT(尺度不变特征变换，Scale-invariant feature transform)为代表的算法，直接对图像进行高斯尺度空间处理，通过对相关点进行主方向提取形成特征描述子再通过欧式距离计算进行匹配，该类方法对图像中像素间关联性较强能进行较好的匹配，而在信号灯识别场景中信号灯较小、形状特征单一的情况下，关键点主方向信息很难提取。
发明内容
本申请提供一种信号灯的识别方法及装置、电子设备，用以解决现有技 术中的图像补充方法无法进行离散场景中信号灯图像信息补充的问题。
本发明实施例提供的一种信号灯的识别方法，包括：
获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
确定参考信号灯，所述N张图像中均有所述参考信号灯；
针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述待补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
一种可选的实施例中，所述参考信号灯的个数至少为2个；
所述确定参考信号灯，包括：
选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述待补充信号灯不位于同一直线上。
一种可选的实施例中，所述至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置，包括：
根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
根据所述第三位置和所述位置映射关系，确定所述第四位置。
一种可选的实施例中，所述根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息，包括：
针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
一种可选的实施例中，所述获取在预设时段内针对同一组信号灯拍摄的N张图像之后，还包括：
对所述N张图像进行二值化处理，并去除噪声；
利用梯度计算确定所述N张图像上所有信号灯的位置。
一种可选的实施例中，所述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。
本发明实施例还提供一种信号灯的识别装置，包括：
获取单元，用于获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
确定单元，用于确定参考信号灯，所述N张图像中均有所述参考信号灯；
计算单元，用于针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置和所述参考信号灯在基准图像中的第二位置，确定所述参照图像中的补充信号灯的第三位置，以及所述基准图像中与所述补充信号灯为同一信号灯的待补充信号灯的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
补充单元，用于根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
一种可选的实施例中，所述参考信号灯的个数至少为2个；
所述确定单元，具体用于：
选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述补充信号灯不位于同一直线上。
一种可选的实施例中，所述计算单元，具体用于：
根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像 间的位置映射关系；
根据所述第三位置和所述位置映射关系，确定所述第四位置。
一种可选的实施例中，所述补充单元，具体用于：
针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
一种可选的实施例中，还包括处理单元，用于：
对所述N张图像进行二值化处理，并去除噪声；
利用梯度计算确定所述N张图像上所有信号灯的位置。
一种可选的实施例中，所述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。
本发明实施例还提供一种电子设备，包括：
至少一个处理器；以及，
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如上所述的方法。
本发明实施例还提供一种非暂态计算机可读存储介质，所述非暂态计算机可读存储介质存储计算机指令，所述计算机指令用于使所述计算机执行如上所述的方法。
本发明实施例中，获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯。将所有N张图像中均存在的信号灯作为参考信号灯。将N张图像中的任一张图像作为基准图像，将N张图像中除了基准图像之外的任一图像作为参照图像。针对每个参照图像，根据参考信号灯在参照图像中的第一位置、参考信号灯在基准图像中的第二位置，以及补充信号灯在参照图像中的第三位置，确定待补充信号灯在基准图像的 第四位置。其中，待补充信号灯为参照图像中除参考信号灯外的任一信号灯。根据至少一个第三位置的信号灯信息确定第四位置的信号灯信息。本发明实施例将不同角度、不同时间点拍摄的图像进行对比，有效利用信号灯分布的位置特点，将同一信号灯在不同图像中的位置建立关联，从而使不同图像中的信号灯信息可以相互补充，便于识别出所有处于亮起状态的信号灯，计算量较小，方法简单高效。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简要介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域的普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本发明实施例提供的一种信号灯的识别方法的流程示意图；
图2为本发明实施例提供的二值化图像的示意图；
图3为本发明实施例中机柜的信号灯排列的示意图；
图4为本发明实施例中参照图像中信号灯与基准图像中信号灯的位置映射关系的示意图；
图5为本发明具体实施例提供的一种信号灯的识别方法的流程示意图；
图6为本发明实施例提供的一种信号灯的识别装置的结构示意图；
图7为本发明实施例提供的电子设备的结构示意图。
具体实施方式
为了使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明作进一步地详细描述，显然，所描述的实施例仅仅是本发明一部份实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
机房智能巡检机器人是各大金融企业需要深入研究和应用的重点前沿创新技术。经过了解，目前国内大型银行已经开始了机房智能巡检机器人的研究。智能巡检机器人通过对服务器摄像或拍照，识别服务器的信号灯情况，从而判断服务器的状态。但是由于信号灯会闪烁，以及拍摄角度的问题，一张图像未必能准确反应出所有信号灯的信息。例如，在预设时间内，有的信号灯实际是处于亮起状态，但由于拍摄某张图像的时间点与该信号灯闪烁的时间点相重合，因此，会出现有的图像中该信号灯的状态为不亮状态。因此需要多张图像相互补充。
为了解决上述问题，本发明实施例提供了一种信号灯的识别方法，如图1所示，本发明实施例提供的信号灯的识别方法包括以下步骤：
步骤101、获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯。
步骤102、确定参考信号灯，所述N张图像中均有所述参考信号灯。
步骤103、针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯。
步骤104、根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
本发明实施例中，获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯。将所有N张图像中均存在的信号灯作为参考信号灯。将N张图像中的任一张图像作为基准图像，将N张图像中除了基准图像之外的任一图像作为参照图像。针对每个参照图像，根据参考信号灯在参照图像中的第一位置和参考信号灯在基准图像中的第二位置，确定参照图像中的补充信号灯的第三位置，以及基准图像中与补充信号灯为 同一信号灯的待补充信号灯的第四位置。其中，补充信号灯为参照图像中除参考信号灯外的任一信号灯。根据至少一个第三位置的信号灯信息确定第四位置的信号灯信息。本发明实施例将不同角度、不同时间点拍摄的图像进行对比，有效利用信号灯分布的位置特点，将同一信号灯在不同图像中的位置建立关联，从而使不同图像中的信号灯信息可以相互补充，便于识别出所有处于亮起状态的信号灯，计算量较小，方法简单高效。
本发明实施例对服务器机柜上的信号灯进行拍摄，针对同一组信号灯拍摄N张图像。为了对图像进行初步处理，上述步骤101、所述获取在预设时段内针对同一组信号灯拍摄的N张图像之后，还包括：
对所述N张图像进行二值化处理，并去除噪声；
利用梯度计算确定所述N张图像上所有信号灯的位置。
具体实施过程中，首先对每张图像中的信号灯进行初步识别，确定每一个信号灯的初步位置。对图像进行二值化处理，即仅把信号灯的位置信息保留到黑板上，形成黑底白点的二值化图像，如图2所示。把每个信号灯的信息，包括信号灯的状态、颜色、所属服务器等信息，另外对应保存。利用二值化后的图像进行识别可以有效增强信号灯的特点，便于降噪。由于信号灯容易与杂质混淆，无降噪处理会降低正确率，因此，在二值化处理后，还利用高斯窗口对图像去除噪声。
之后，利用梯度计算确定二值化图像中信号灯的准确位置，为了便于计算，利用每个用信号灯重心的像素点代替对应的信号灯，进行分析处理。
本发明实施例中参考信号灯的个数至少为2个。所述确定参考信号灯，包括：
选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述待补充信号灯不位于同一直线上。
机柜的信号灯排列如图3所示。由于拍摄角度等原因，同一个信号灯在不同图像中的位置有差别，且由于图像的形变，信号灯之间的位置关系在不同图像中也不同，因此，需要建立不同图像间信号灯的位置映射关系，便于 确定待补充信号灯的准确位置，从而找到基准图像中的待补充信号灯。参考信号灯的作用即是为了建立基准图像中信号灯与参照图像中信号灯之间的位置映射关系。为了便于识别，本发明实施例选取N张图像中均处于亮起状态的信号灯作为参考信号灯，且，同一张图像中的两个参考信号灯以及待补充信号灯不位于同一直线上。这是因为若参考信号灯与待补充信号灯位于同一直线上，则无法反映出图像上不同方向的形变。
上述步骤103、所述至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置，包括：
根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
根据所述第三位置和所述位置映射关系，确定所述第四位置。
具体地，从N张图像中任选一张作为基准图像，其余的N-1张图像均为参照图像。针对任一参照图像，均需建立该参照图像与基准图像之间的位置映射关系，从而将参照图像中的补充信号灯与基准图像的待补充信号灯建立位置联系。具体实施过程中，可以利用同一张图像中两个参考信号灯位置之间的映射关系，标识每个参考信号灯的位置，更为简便的方式为在每个图像中建立坐标系，确定每个信号灯的位置坐标，利用坐标标识对应信号灯的位置。
举例来说，图4示出了参照图像中信号灯与基准图像中信号灯的位置映射关系。信号灯A与信号灯B为参考信号灯，信号灯C为待补充信号灯。如图4所示，A 1、B 1与C 1为参照图像中的信号灯位置，A 2、B 2与C 2为基准图像中的信号灯位置。其中，A 1与A 2相对应，且对应于信号灯A；B 1与B 2相对应，且对应于信号灯B；C 1与C 2相对应，且对应于信号灯C。根据参考信号灯的位置以及待补充信号灯在参照图像中的位置，确定待补充信号灯在基准图像中的位置，即为根据A 1、B 1、C 1、A 2和B 2，确定C 2。由图4可以看 出，根据信号灯的空间分布特点，同一张图像中信号灯A、信号灯B与信号灯C形成三角形。针对不同的两张图像，可以由A 1的坐标和A 2的坐标得出A 1A 2之间的映射关系、由B 1的坐标和B 2的坐标得出B 1B 2之间的映射关系。根据A 1A 2之间的映射关系以及B 1B 2之间的映射关系，可以得出C 1C 2之间的映射关系，再知道C 1的坐标，则可确定基准图像中C 2的坐标。
进一步地，步骤104、所述根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息，包括：
针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
具体来说，确定了基准图像中待补充信号灯的位置之后，利用其余N-1张参照图像中待补充信号灯的信息，对基准图像中待补充信号灯信息进行信息补充。每个信号灯信息包括M个子信息，针对每一个子信息，统计N-1张参照图像中每一种子信息内容的数量，并进行排序，选取数量最多的子信息内容为基准图像中该待补充信号灯的子信息内容。例如，子信息为信号灯的状态、颜色、所属服务器。针对信号灯的状态，子信息内容有亮和不亮，49个参照图像中，有10个参照图像中该待补充信号灯的子信息内容为不亮，其余39个参照图像中该待补充信号灯的子信息内容为亮。则选取子信息信号灯状态的内容为亮，即将基准图像中该待补充信号灯的信号灯状态标记为亮。又例如，针对信号灯的颜色，子信息内容有红、黄和绿三种。49个参照图像中，有9个参照图像中该待补充信号灯的子信息内容为红，15个参照图像中该待补充信号灯的子信息内容为绿，其余25个参照图像中该待补充信号灯的子信息内容为黄，则选取子信息信号灯颜色的内容为黄，即将基准图像中该待补充信号灯的信号灯颜色标记为黄。如此，遍历基准图像中的每一个待补充信号灯，利用其余N-1张参照图像的信号灯信息，对基准图像中的待补充 信号灯进行信息补充。
除了对机柜的信号灯进行拍照外，还可以拍摄视频。则上述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。除了第一帧之外其余帧图像作为参照图像。
为了更清楚地理解本发明，下面以具体实施例对上述流程进行详细描述，具体实施例的步骤如图5所示，包括：
步骤401：获取在预设时段内针对同一组信号灯拍摄的N张图像。
步骤402：从N张图像中任选一张作为基准图像，其余N-1张图像作为参照图像。
步骤403：对每一张图像进行二值化处理，并进行去噪。
步骤404：利用梯度计算找到二值化图像中信号灯的位置，用信号灯的重心代替信号灯进行分析处理，确定信号灯的重心位置。
步骤405：将所有图像中都能识别的两个信号灯作为参照信号灯。
步骤406：针对一个待识别信号灯，建立该待识别信号灯在基准图像中与该待识别信号灯在任一参照图像中的位置映射关系。位置映射关系的建立为依据基准图像中参照信号灯的位置坐标、参照图像中参照信号灯的位置坐标，以及参照图像中基准信号灯的位置坐标进行确定。
步骤407：统计所有参照图像中该待识别信号灯的子信息内容的数量，选取数量最多的子信息内容作为基准图像中待识别信号灯的子信息内容。
步骤408：遍历基准图像中每一个待识别信号灯，确定每一个待识别信号灯的位置和信号灯信息。
本发明实施例还提供了一种信号灯的识别装置，如图6所示，包括：
获取单元501，用于获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
确定单元502，用于确定参考信号灯，所述N张图像中均有所述参考信号灯；
计算单元503，用于针对每个参照图像，至少根据所述参考信号灯在所述 参照图像的第一位置和所述参考信号灯在基准图像中的第二位置，确定所述参照图像中的补充信号灯的第三位置，以及所述基准图像中与所述补充信号灯为同一信号灯的待补充信号灯的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
补充单元504，用于根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
进一步地，所述参考信号灯的个数至少为2个；
所述确定单元502，具体用于：
选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述补充信号灯不位于同一直线上。
进一步地，所述计算单元503，具体用于：
根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
根据所述第三位置和所述位置映射关系，确定所述第四位置。
进一步地，所述补充单元504，具体用于：
针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
进一步地，还包括处理单元505，用于：
对所述N张图像进行二值化处理，并去除噪声；
利用梯度计算确定所述N张图像上所有信号灯的位置。
进一步地，所述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。
基于相同的原理，本发明还提供一种电子设备，如图7所示，包括：
包括处理器601、存储器602、收发机603、总线接口604，其中处理器601、存储器602与收发机603之间通过总线接口604连接；
所述处理器601，用于读取所述存储器602中的程序，执行下列方法：
获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
确定参考信号灯，所述N张图像中均有所述参考信号灯；
针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述待补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
进一步地，所述参考信号灯的个数至少为2个；所述处理器601具体用于：
选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述待补充信号灯不位于同一直线上。
进一步地，所述处理器601具体用于：
根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
根据所述第三位置和所述位置映射关系，确定所述第四位置。
进一步地，所述处理器601具体用于：
针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
进一步地，所述处理器601还用于：
对所述N张图像进行二值化处理，并去除噪声；
利用梯度计算确定所述N张图像上所有信号灯的位置。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本发明的优选实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例作出另外的变更和修改。所以，所附权利要求意欲解释为包括优选实施例以及落入本发明范围的所有变更和修改。
显然，本领域的技术人员可以对本发明进行各种改动和变型而不脱离本发明的精神和范围。这样，倘若本发明的这些修改和变型属于本发明权利要求及其等同技术的范围之内，则本发明也意图包括这些改动和变型在内。

Claims (14)

  1. 一种信号灯的识别方法，其特征在于，包括：
    获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
    确定参考信号灯，所述N张图像中均有所述参考信号灯；
    针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述待补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
    根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
  2. 如权利要求1所述的方法，其特征在于，所述参考信号灯的个数至少为2个；
    所述确定参考信号灯，包括：
    选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且，同一张图像中的两个参考信号灯与所述待补充信号灯不位于同一直线上。
  3. 如权利要求1所述的方法，其特征在于，所述至少根据所述参考信号灯在所述参照图像的第一位置、所述参考信号灯在基准图像中的第二位置，以及待补充信号灯在所述参照图像中的第三位置，确定所述待补充信号灯在所述基准图像中的第四位置，包括：
    根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
    根据所述第三位置和所述位置映射关系，确定所述第四位置。
  4. 如权利要求1所述的方法，其特征在于，所述根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息，包括：
    针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
    针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
  5. 如权利要求1至4任一项所述的方法，其特征在于，所述获取在预设时段内针对同一组信号灯拍摄的N张图像之后，还包括：
    对所述N张图像进行二值化处理，并去除噪声；
    利用梯度计算确定所述N张图像上所有信号灯的位置。
  6. 如权利要求1至4任一项所述的方法，其特征在于，所述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。
  7. 一种信号灯的识别装置，其特征在于，包括：
    获取单元，用于获取在预设时段内针对同一组信号灯拍摄的N张图像；每张图像用于识别处于亮起状态的信号灯；
    确定单元，用于确定参考信号灯，所述N张图像中均有所述参考信号灯；
    计算单元，用于针对每个参照图像，至少根据所述参考信号灯在所述参照图像的第一位置和所述参考信号灯在基准图像中的第二位置，确定所述参照图像中的补充信号灯的第三位置，以及所述基准图像中与所述补充信号灯为同一信号灯的待补充信号灯的第四位置；所述基准图像为所述N张图像中的任一个，所述参照图像为所述N张图像中除所述基准图像外的任一个；所述补充信号灯为所述参照图像中除所述参考信号灯外的任一信号灯；
    补充单元，用于根据至少一个第三位置的信号灯信息确定所述第四位置的信号灯信息。
  8. 如权利要求7所述的装置，其特征在于，所述参考信号灯的个数至少为2个；
    所述确定单元，具体用于：
    选取所述N张图像中均处于亮起状态的信号灯作为所述参考信号灯，且， 同一张图像中的两个参考信号灯与所述补充信号灯不位于同一直线上。
  9. 如权利要求7所述的装置，其特征在于，所述计算单元，具体用于：
    根据所述第一位置和所述第二位置，确定所述参照图像与所述基准图像间的位置映射关系；
    根据所述第三位置和所述位置映射关系，确定所述第四位置。
  10. 如权利要求7所述的装置，其特征在于，所述补充单元，具体用于：
    针对N-1张参照图像中任一张参照图像，获取所述参照图像在第三位置的信号灯信息，所述信号灯信息包括M个子信息；
    针对每一个子信息，统计所述N-1张参照图像对应的N-1个子信息中内容相同的子信息的数量；将数量最多的子信息作为所述第四位置的信号灯信息的子信息。
  11. 如权利要求7至10任一项所述的装置，其特征在于，还包括处理单元，用于：
    对所述N张图像进行二值化处理，并去除噪声；
    利用梯度计算确定所述N张图像上所有信号灯的位置。
  12. 如权利要求7至10任一项所述的装置，其特征在于，所述N张图像为同一视频流中的N帧图像，所述基准图像为所述视频流的第一帧图像。
  13. 一种电子设备，其特征在于，包括：
    至少一个处理器；以及，
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1-6任一所述的方法。
  14. 一种非暂态计算机可读存储介质，其特征在于，所述非暂态计算机可读存储介质存储计算机指令，所述计算机指令用于使所述计算机执行权利要求1～6任一所述方法。
PCT/CN2019/093201 2018-12-29 2019-06-27 一种信号灯的识别方法及装置、电子设备 Ceased WO2020133983A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811647448.7 2018-12-29
CN201811647448.7A CN109711375B (zh) 2018-12-29 2018-12-29 一种信号灯的识别方法及装置

Publications (1)

Publication Number Publication Date
WO2020133983A1 true WO2020133983A1 (zh) 2020-07-02

Family

ID=66260287

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/093201 Ceased WO2020133983A1 (zh) 2018-12-29 2019-06-27 一种信号灯的识别方法及装置、电子设备

Country Status (2)

Country Link
CN (1) CN109711375B (zh)
WO (1) WO2020133983A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112364740A (zh) * 2020-10-30 2021-02-12 交控科技股份有限公司 一种基于计算机视觉的无人机房监控方法及系统
CN113129375A (zh) * 2021-04-21 2021-07-16 阿波罗智联(北京)科技有限公司 数据处理方法、装置、设备及存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109711375B (zh) * 2018-12-29 2021-03-16 中国银联股份有限公司 一种信号灯的识别方法及装置
CN111192377B (zh) * 2020-01-08 2021-08-27 中国银联股份有限公司 一种图像处理方法及装置
CN114092717B (zh) * 2020-09-23 2025-02-14 阿波罗智联(北京)科技有限公司 图像数据挖掘方法、装置、设备、路侧设备、边缘计算平台

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103531022A (zh) * 2013-10-15 2014-01-22 浙江宇视科技有限公司 一种红绿灯颜色周期统计方法
US20170353642A1 (en) * 2016-06-01 2017-12-07 Hyundai Motor Company Driver state monitoring system
CN108447091A (zh) * 2018-03-27 2018-08-24 北京颂泽科技有限公司 目标定位方法、装置、电子设备及存储介质
CN109711375A (zh) * 2018-12-29 2019-05-03 中国银联股份有限公司 一种信号灯的识别方法及装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107743200A (zh) * 2017-10-31 2018-02-27 广东欧珀移动通信有限公司 拍照的方法、装置、计算机可读存储介质和电子设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103531022A (zh) * 2013-10-15 2014-01-22 浙江宇视科技有限公司 一种红绿灯颜色周期统计方法
US20170353642A1 (en) * 2016-06-01 2017-12-07 Hyundai Motor Company Driver state monitoring system
CN108447091A (zh) * 2018-03-27 2018-08-24 北京颂泽科技有限公司 目标定位方法、装置、电子设备及存储介质
CN109711375A (zh) * 2018-12-29 2019-05-03 中国银联股份有限公司 一种信号灯的识别方法及装置

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112364740A (zh) * 2020-10-30 2021-02-12 交控科技股份有限公司 一种基于计算机视觉的无人机房监控方法及系统
CN112364740B (zh) * 2020-10-30 2024-04-19 交控科技股份有限公司 一种基于计算机视觉的无人机房监控方法及系统
CN113129375A (zh) * 2021-04-21 2021-07-16 阿波罗智联(北京)科技有限公司 数据处理方法、装置、设备及存储介质
CN113129375B (zh) * 2021-04-21 2023-12-01 阿波罗智联(北京)科技有限公司 数据处理方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN109711375B (zh) 2021-03-16
CN109711375A (zh) 2019-05-03

Similar Documents

Publication Publication Date Title
WO2020133983A1 (zh) 一种信号灯的识别方法及装置、电子设备
US10740912B2 (en) Detection of humans in images using depth information
CN106650740B (zh) 一种车牌识别方法及终端
CN110751079A (zh) 物品检测方法、装置、系统和计算机可读存储介质
CN109308447A (zh) 在电力远程监控中自动提取设备运行参数和运行状态的方法
KR20210006276A (ko) 플리커를 약화시키기 위한 이미지 처리 방법
CN106096603A (zh) 一种融合多特征的动态火焰检测方法及装置
CN112863194B (zh) 一种图像处理方法、装置、终端及介质
CN117934897A (zh) 一种设备异常检测方法、装置、设备以及存储介质
Prabhakar et al. A novel design for vehicle license plate detection and recognition
CN113221603A (zh) 一种检测监控设备被异物遮挡的方法及装置
CN106529556B (zh) 一种仪表指示灯的视觉检测系统
CN108318773A (zh) 一种输电导线断股检测方法及系统
CN111259763A (zh) 目标检测方法、装置、电子设备及可读存储介质
CN113408409A (zh) 交通信号灯识别方法、设备、云控平台和车路协同系统
US11216905B2 (en) Automatic detection, counting, and measurement of lumber boards using a handheld device
CN112883762A (zh) 一种活体检测方法、装置、系统及存储介质
CN111325773A (zh) 运动目标的检测方法、装置、设备及可读存储介质
CN103793923A (zh) 一种图像中运动目标的获取方法及设备
CN115410109A (zh) 火焰检测方法、装置、电子设备及存储介质
CN115620185B (zh) 一种指示灯发光状态的识别方法、识别装置及终端设备
CN106503740A (zh) 图片分类方法及装置
CN112949423A (zh) 物体识别方法、物体识别装置及机器人
CN116612088B (zh) 一种废旧服务端拆解零部件缺陷检测方法和系统
CN112183232A (zh) 基于深度学习的船舷号位置定位方法及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19904982

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19904982

Country of ref document: EP

Kind code of ref document: A1