WO2020187311A1 - 一种图像识别方法及装置 - Google Patents
一种图像识别方法及装置 Download PDFInfo
- Publication number
- WO2020187311A1 WO2020187311A1 PCT/CN2020/080321 CN2020080321W WO2020187311A1 WO 2020187311 A1 WO2020187311 A1 WO 2020187311A1 CN 2020080321 W CN2020080321 W CN 2020080321W WO 2020187311 A1 WO2020187311 A1 WO 2020187311A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- target vehicle
- image
- vehicle
- recognized
- edge
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
Definitions
- the present invention relates to the computer field, in particular to an image recognition method and device.
- the advanced driving assistance system is a system that uses sensors installed in the speed of the car to perceive the surrounding environment while the car is running, and through calculation and analysis to allow the driver to perceive possible dangers in advance, which can effectively increase the comfort and comfort of car driving. safety.
- the position of the vehicle ahead is important data.
- the distance measurement of the preceding vehicle can be carried out by using radar.
- the cost of radar is relatively high, and there is no distinction between target types, and accurate determination of the position of the preceding vehicle cannot be achieved.
- the image of the vehicle in front can be acquired through the image acquisition device, and the position of the vehicle in front can be obtained through image analysis. Specifically, the position of the target vehicle in the image can be identified, and then calculated according to the position of the target vehicle in the image The relative position of the target vehicle and the device that acquired the image.
- the image recognition in the prior art is not accurate enough, so the obtained position of the vehicle in front is not accurate enough.
- the embodiments of the present application provide an image recognition method and device to improve the accuracy of image recognition, thereby improving the accuracy of vehicle positioning.
- the embodiment of the application provides an image recognition method, the method includes:
- the bottom edge of the target vehicle is determined from the horizontal edge according to the relative position of the bottom feature of the target vehicle and the horizontal edge.
- the determining the bottom edge of the target vehicle from the horizontal edge according to the relative position of the bottom feature of the target vehicle and the horizontal edge includes:
- the horizontal side with the score greater than or equal to the preset value is used as the bottom side of the target vehicle.
- the identifying a shadow of the vehicle bottom on the target vehicle in the image to be identified includes:
- the identifying wheels on the target vehicle in the image to be identified includes:
- the method further includes:
- the relative position of the target vehicle and the image acquisition device that acquires the image to be identified is calculated.
- the method further includes:
- the relative position of the target vehicle and the image acquisition device that acquires the image to be recognized is calculated.
- An embodiment of the present application also provides an image recognition device, which includes:
- An image acquisition unit for acquiring an image to be identified including the target vehicle
- a horizontal edge acquiring unit configured to identify a horizontal edge on the target vehicle in the image to be recognized
- a feature acquisition unit configured to identify vehicle bottom features on the target vehicle in the to-be-recognized image, where the vehicle bottom features include vehicle bottom shadows and/or wheels;
- the bottom edge determining unit is used to determine the bottom edge of the target vehicle from the horizontal edges according to the relative positions of the bottom features of the target vehicle and the horizontal edges.
- the vehicle bottom edge determination unit includes:
- the upper boundary determination unit is used to determine the boundary of the vehicle bottom feature
- the score determining unit is configured to determine the score for each horizontal edge according to the distance between the boundary of the vehicle bottom feature and each of the horizontal edges;
- the vehicle bottom edge determination subunit is configured to use the horizontal edge with the score greater than or equal to a preset value as the vehicle bottom edge of the target vehicle.
- the feature acquisition unit is specifically configured to:
- the device further includes:
- the first position calculation unit is configured to calculate the relative position of the target vehicle and the image acquisition device that acquires the image to be identified according to the position of the bottom edge of the target vehicle in the image to be identified.
- the device further includes:
- a vertical line acquiring unit configured to recognize a vertical line on the target vehicle in the image to be recognized
- the vertical boundary line determining unit is configured to determine the vertical boundary line of the target vehicle from the vertical line according to the bottom edge end point feature of the target vehicle and the symmetry feature of the target vehicle, the The vertical boundary line includes the left boundary line and/or the right boundary line;
- the second position calculation unit is configured to calculate the relative position of the target vehicle and the image acquisition device that acquires the image to be recognized based on the bottom edge of the vehicle and the vertical boundary line.
- the embodiments of the present application provide an image recognition method and device, which acquire a to-be-recognized image that includes a target vehicle, and identify horizontal edges and under-vehicle features on the target vehicle in the to-be-recognized image.
- the under-vehicle features may include under-vehicle shadows and/ Or wheels, the bottom edge of the target vehicle is determined from the identified horizontal edges according to the characteristics of the bottom of the target vehicle and the relative position of the horizontal edge.
- the bottom edge of the vehicle since the bottom edge of the vehicle is usually the most protruding edge on the target vehicle, it can be used to characterize the position of the target vehicle, and the bottom edge of the vehicle is close in position due to the characteristics of the bottom of the vehicle.
- the relative position of the horizontal edge to identify the bottom edge of the vehicle can improve the accuracy of the recognition of the bottom edge of the vehicle, thereby improving the accuracy of the actual positioning of the vehicle.
- Fig. 1 is a schematic diagram of an image of a target vehicle in the prior art
- FIG. 2 is a flowchart of an image recognition method provided by an embodiment of the application
- FIG. 3 is a schematic diagram of an image to be recognized according to an embodiment of the application.
- FIG. 4 is a schematic diagram of another image to be recognized provided by an embodiment of this application.
- Fig. 5 is a structural block diagram of an image recognition device provided by an embodiment of the application.
- the image of the vehicle in front can be acquired through the image acquisition device and the position of the vehicle in front can be obtained through image analysis. Specifically, the position of the target vehicle in the image can be identified, and then the target vehicle and the acquired image can be calculated according to the position of the vehicle in the image. The relative position of the device, so that the relative position of the target vehicle and the vehicle where the image acquisition device is located can be obtained.
- FIG. 1 is a schematic diagram of the image of the target vehicle in the prior art.
- the area where the target vehicle is located (refer to the black rectangular frame), and then the relative position of the target vehicle and the device that obtains the image is calculated according to the location of the area where the target vehicle is located in the image.
- the inventor found through research that the area of the identified target vehicle is usually larger than the actual area of the target vehicle. Therefore, it is usually not accurate enough to calculate the relative position of the target vehicle and the device that acquires the image based on the identified area.
- the embodiment of the present application provides an image recognition method and device, which acquires an image to be recognized that includes a target vehicle, and recognizes horizontal edges and vehicle bottom features on the target vehicle in the to be recognized image.
- the vehicle bottom features may include the bottom of the vehicle.
- the bottom edge of the target vehicle is determined from the identified horizontal edges based on the characteristics of the bottom of the target vehicle and the relative position of the horizontal edge.
- since the bottom edge of the vehicle is usually the most protruding edge on the target vehicle, it can be used to characterize the position of the target vehicle, and the bottom edge of the vehicle is close in position due to the characteristics of the bottom of the vehicle.
- the relative position of the horizontal edge to identify the bottom edge of the vehicle can improve the accuracy of the recognition of the bottom edge of the vehicle, thereby improving the accuracy of the actual positioning of the vehicle.
- FIG. 2 shows a flowchart of an image recognition method provided by an embodiment of this application, which may include the following steps.
- S101 Acquire an image to be recognized that includes a target vehicle.
- the target vehicle is a vehicle in front of the image acquisition device, and the image to be recognized is obtained by the image acquisition device photographing the target vehicle.
- the image acquisition device may be a camera, a camera, etc., or other image acquisition devices.
- the image to be recognized can include the head or tail of the target vehicle, or the side of the target vehicle, as shown in FIG. 1.
- the image to be recognized may include one target vehicle or multiple target vehicles.
- a single target vehicle in the image to be recognized may be analyzed separately.
- the image to be recognized can also be obtained by intercepting the original image.
- the candidate area including the target vehicle is recognized as the image to be recognized in the original image, so that the image to be recognized can include only one target vehicle.
- the interception of the original image can be achieved through a machine learning model.
- S102 Recognizing horizontal edge and vehicle bottom features on the target vehicle in the image to be recognized.
- the horizontal edge can be identified by the edge detection algorithm, or the horizontal edge can be identified by calculating the pixel gradient value in the horizontal direction.
- the pixel gradient value in the horizontal direction It is the difference between the pixel value of each pixel in the left column and the pixel value of each pixel in the right column.
- the pixel gradient value on the horizontal side is small.
- the white line segments are the horizontal edges on the target vehicle.
- the horizontal edge may include at least one of the following line segments: the edge where the roof is located, the upper boundary of the license plate, the lower boundary of the license plate, the boundary of the tailgate, the connection of the tail light, the connection of the lower boundary of the wheel, and the horizontal line with similar pixel values at the shadow, etc. .
- the under-vehicle features of the target vehicle can include under-vehicle shadows.
- Under-vehicle shadows are shadow areas outside the body of the target vehicle, usually with low pixel values. Refer to Figure 3, where the larger gray ellipse is the under-vehicle shadow of the target vehicle. area.
- the bottom shadow of the target vehicle can be determined according to the pixel gradient value of the pixel in the image to be recognized, where the pixel gradient value can be in the horizontal direction, or in the vertical direction, and the pixel gradient in the vertical direction.
- the value is the difference between the pixel value of each pixel in the previous row and the pixel value of each pixel in the next row.
- the pixel value of the shadow of the car bottom is low, usually higher than other positions in the image to be recognized.
- the calculated pixel gradient value can be obtained from the bottom up.
- the determined under-vehicle shadow area only includes the position above the lowest pixel value of the under-vehicle.
- a complete under-vehicle shadow area can also be determined, which will not be described in detail here.
- the bottom features of the target vehicle can also include wheels.
- the wheels are also features located on the bottom of the target vehicle.
- the color of the wheels is usually darker at the edges and shallower in the middle when viewed from the rear. It is deeper than other areas, as shown in Figure 3, where the two smaller gray ellipses are the wheel areas of the target vehicle.
- the wheel area of the target vehicle can be determined according to the pixel gradient value and the pixel value of the pixel in the image to be identified.
- the pixel gradient value may be in the horizontal direction or in the vertical direction.
- the pixel gradient value of the wheel area will have maximum and minimum values at the left and right edge positions, and the pixel value will have a lower value at the middle position, thereby determining the wheel area.
- the target vehicle bottom feature may include a vehicle bottom shadow, may also include wheels, and may also include both vehicle bottom shadows and wheels.
- the elliptical undercarriage shadow and the elliptical wheel shown above are an example, which are only for clear description. In fact, for the convenience of calculation, the undercarriage shadow area and the wheel area can also be approximated as rectangles.
- S103 Determine the bottom edge of the target vehicle from the horizontal edge according to the relative position of the bottom feature of the target vehicle and the horizontal edge.
- At least one horizontal edge is identified on the target vehicle.
- these horizontal edges are not necessarily useful. Therefore, the horizontal edges need to be filtered.
- the horizontal edges can be selected from the characteristics of the target vehicle.
- the bottom edge of the target vehicle is determined in the edge.
- the bottom edge of the vehicle is usually the most protruding edge on the target vehicle. Therefore, the relative position of the target vehicle and the image acquisition device calculated based on the bottom edge of the vehicle is the closest to the actual distance.
- the boundary of the vehicle bottom feature can be determined. According to the distance between the boundary of the vehicle bottom feature and the horizontal edge, the score is determined for the horizontal edge, and the horizontal edge with the score greater than or equal to the preset value is regarded as the bottom of the target vehicle. side.
- the boundary of the vehicle bottom feature may include at least one of the following boundaries: an upper boundary of a vehicle bottom shadow, a lower boundary of a vehicle bottom shadow, an upper boundary of a wheel, and a lower boundary of the wheel.
- the upper or lower boundary of the shadow of the vehicle bottom can be a straight line or a curve of other shapes.
- the upper or lower boundary of a wheel can be a straight line or a curve of other shapes, and it can be a wheel.
- the upper boundary or the lower boundary may also be a line connecting the upper boundary of two wheels or a line connecting the lower boundary of two wheels.
- the distance between the boundary of the vehicle bottom feature and each horizontal edge can be calculated to calculate the score of the horizontal edge.
- the distance between the lower boundary of the vehicle bottom shadow and the lower boundary of the wheel and the bottom edge of the vehicle is close, or even coincides, so the distance between the horizontal edge and the lower boundary of the vehicle bottom shadow can be calculated, and/or the horizontal edge
- the score of the horizontal edge is inversely proportional to the calculated distance. In this way, the farther the lower boundary of the shadow of the vehicle or the lower boundary of the vehicle, the lower the score of the horizontal line, the less likely it is It is the bottom side of the car.
- the horizontal side with a score greater than or equal to the preset value can be used as the bottom side of the target vehicle, or the horizontal side with the highest score can be used as the bottom side of the target vehicle.
- the bottom white line segment can be used as the bottom edge of the car.
- the relative position of the target vehicle and the image capturing device that obtains the image to be recognized can be calculated according to the position of the bottom edge of the target vehicle in the image to be recognized.
- Figure 4 is a schematic diagram of another image to be recognized provided in this embodiment of the application, in which the white line segment is the determined bottom edge of each target vehicle, combined with the shooting parameters such as the focal length of the image acquisition device, The relative position of the bottom edge of each target vehicle and the image acquisition device is calculated, and accurate positioning of the target vehicle is realized.
- the vertical line on the target vehicle can also be recognized in the image to be recognized, and the recognition method of the vertical line can refer to the recognition method of the horizontal edge, which will not be repeated here.
- the vertical black line segments on the left and right sides are vertical lines.
- the vertical boundary line of the target vehicle can be determined from the vertical line according to the end point characteristics of the bottom edge of the target vehicle and the symmetry characteristics of the target vehicle.
- the vertical boundary line may include the left boundary line And/or the right boundary line. Specifically, a score is determined for each vertical line according to the distance between the left or right boundary of the vehicle bottom feature and the vertical line, and the vertical line with the score greater than or equal to the preset value is used as the vertical boundary line of the target vehicle .
- the left boundary line and the right boundary line are symmetrical about the center axis. Therefore, the center axis of the target vehicle can be identified in the image to be recognized based on the symmetry characteristics of the target vehicle. Refer to Figure 3 for the black line segment at the center position. Is the center axis of the target vehicle, and the target vehicle is symmetrical about the center axis.
- the sum of the scores of the two vertical lines symmetric about the central axis can also be calculated, and the pair of vertical lines with the largest score sum is used as the left boundary line and the right boundary line.
- the vertical boundary line of the target vehicle may be determined according to the end point of the bottom edge of the target vehicle, and the target vehicle may be calculated according to the bottom edge and the vertical boundary line.
- the relative position of the target vehicle and the image acquisition device that obtains the image to be recognized can be calculated according to the position of the bottom edge of the target vehicle and the vertical boundary line in the image to be recognized.
- the white line segment is the determined bottom edge of each target vehicle
- the black line segment is the determined vertical boundary line of each target vehicle.
- the embodiment of the present application provides an image recognition method, which obtains an image to be recognized including a target vehicle, and recognizes horizontal edges and vehicle bottom features on the target vehicle in the to be recognized image.
- the vehicle bottom features may include vehicle bottom shadows and/or wheels , According to the characteristics of the bottom of the target vehicle and the relative position of the horizontal edge, determine the bottom edge of the target vehicle from the identified horizontal edges.
- the bottom edge of the vehicle is usually the most protruding edge on the target vehicle, it can be used to characterize the position of the target vehicle, and the bottom edge of the vehicle is close to the features of the bottom of the vehicle, so according to the characteristics of the bottom Recognizing the bottom edge of the vehicle by relative position can improve the accuracy of the recognition of the bottom edge of the vehicle, thereby improving the accuracy of the actual positioning of the vehicle.
- an embodiment of the present application further provides an image recognition device.
- FIG. 5 a structural block diagram of an image recognition device provided in an embodiment of this application, the device includes:
- the image acquisition unit 110 is configured to acquire an image to be identified including the target vehicle;
- the horizontal edge acquiring unit 120 is configured to identify the horizontal edge on the target vehicle in the image to be recognized;
- the feature acquisition unit 130 is configured to identify vehicle bottom features on the target vehicle in the image to be recognized, where the vehicle bottom features include vehicle bottom shadows and/or wheels;
- the bottom edge determining unit 140 is configured to determine the bottom edge of the target vehicle from the horizontal edges according to the relative position of the bottom feature of the target vehicle and the horizontal edge.
- the vehicle bottom edge determination unit includes:
- the upper boundary determination unit is used to determine the boundary of the vehicle bottom feature
- the score determining unit is configured to determine the score for each horizontal edge according to the distance between the boundary of the vehicle bottom feature and each of the horizontal edges;
- the vehicle bottom edge determination subunit is configured to use the horizontal edge with the score greater than or equal to a preset value as the vehicle bottom edge of the target vehicle.
- the feature acquisition unit is specifically configured to:
- the device further includes:
- the first position calculation unit is configured to calculate the relative position of the target vehicle and the image acquisition device that acquires the image to be identified according to the position of the bottom edge of the target vehicle in the image to be identified.
- the device further includes:
- a vertical line acquiring unit configured to recognize a vertical line on the target vehicle in the image to be recognized
- the vertical boundary line determining unit is configured to determine the vertical boundary line of the target vehicle from the vertical line according to the characteristics of the left and right end points of the bottom edge of the target vehicle and the symmetry characteristics of the target vehicle, so The vertical boundary line includes a left boundary line and/or a right boundary line;
- the second position calculation unit is configured to calculate the relative position of the target vehicle and the image acquisition device that acquires the image to be recognized based on the bottom edge of the vehicle and the vertical boundary line.
- An embodiment of the present application provides an image recognition device that acquires an image to be recognized that includes a target vehicle, and recognizes horizontal edges and vehicle bottom features on the target vehicle in the to be recognized image.
- the vehicle bottom features may include vehicle bottom shadows and/or wheels , According to the characteristics of the bottom of the target vehicle and the relative position of the horizontal edge, determine the bottom edge of the target vehicle from the identified horizontal edges.
- the bottom edge of the vehicle since the bottom edge of the vehicle is usually the most protruding edge on the target vehicle, it can be used to characterize the position of the target vehicle, and the bottom edge of the vehicle is close in position due to the characteristics of the bottom of the vehicle.
- the relative position of the horizontal edge to identify the bottom edge of the vehicle can improve the accuracy of the recognition of the bottom edge of the vehicle, thereby improving the accuracy of the actual positioning of the vehicle.
- the computer software product can be stored in a storage medium, such as read-only memory (English: read-only memory, ROM)/RAM, magnetic disk, An optical disc, etc., includes a number of instructions to enable a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the method described in each embodiment of the application or some parts of the embodiment.
- a computer device which may be a personal computer, a server, or a network communication device such as a router
- the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments.
- the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment.
- the above-described device and system embodiments are only illustrative.
- the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in One place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the objectives of the solutions of the embodiments. Those of ordinary skill in the art can understand and implement it without creative work.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
- Image Processing (AREA)
Abstract
本申请实施例公开了一种图像识别方法及装置，获取包括目标车辆的待识别图像，在待识别图像中识别目标车辆上的水平边和车底特征，车底特征可以包括车底阴影和/或车轮，根据目标车辆的车底特征和水平边的相对位置，从识别出的水平边中确定目标车辆的车底边。在本申请实施例中，由于车底边通常是目标车辆上最为凸出的边，可以用来表征目标车辆的位置，而车底边由于车底特征接近，因此根据车底特征和水平边的相对位置来识别车底边，可以提高车底边的识别的准确性，从而提高车辆的实际定位的准确性。
Description
本申请要求于2019年03月20日提交中国专利局、申请号为201910213120.2、申请名称为“一种图像识别方法及装置”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。
本发明涉及计算机领域，特别是涉及一种图像识别方法及装置。
高级驾驶辅助系统是利用安装在车速的传感器，在汽车行驶过程中来感知周围环境，并通过运算和分析从而预先让驾驶者察觉到可能的危险的系统，这样可以有效增加汽车驾驶的舒适性和安全性。
在高级驾驶辅助系统中，前方车辆的位置是重要的数据。在一种方式中，可以通过雷达来进行对前方车辆进行测距，然而雷达成本较高，同时没有目标类型的区分，无法实现前方车辆的位置的准确判定。在另一种方式中，可以通过图像获取设备获取前方车辆的图像，通过图像分析得到前方车辆的位置，具体的，可以识别出图像中目标车辆的位置，然后根据目标车辆在图像中的位置计算目标车辆与获取图像的设备的相对位置。然而，现有技术中对图像的识别不够准确，因此得到的前方车辆的位置也不够准确。
发明内容
为解决上述技术问题，本申请实施例提供一种图像识别方法及装置，提高图像识别的准确性，从而提高车辆定位的准确性。
本申请实施例提供了一种图像识别方法，所述方法包括：
获取包括目标车辆的待识别图像；
在所述待识别图像中识别所述目标车辆上的水平边和车底特征，所述车底特征包括车底阴影和/或车轮；
根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边。
可选的，所述根据所述目标车辆的车底特征与所述水平边的相对位置，从 所述水平边中确定所述目标车辆的车底边，包括：
确定所述车底特征的边界；
根据所述车底特征的边界与各个所述水平边的距离，为各个所述水平边确定分值；
将所述分值大于或等于预设值的水平边作为所述目标车辆的车底边。
可选的，所述在所述待识别图像中识别所述目标车辆上的车底阴影，包括：
根据所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车底阴影区域；
所述在所述待识别图像中识别所述目标车辆上的车轮，包括：
根据所述待识别图像中的像素点的像素值以及所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车轮区域。
可选的，所述方法还包括：
根据所述目标车辆的车底边在所述待识别图像中的位置，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
可选的，所述方法还包括：
在所述待识别图像中识别所述目标车辆上的竖直线；
根据所述目标车辆的车底边端点特征以及所述目标车辆的对称性特征，从所述竖直线中确定所述目标车辆的竖直边界线，所述竖直边界线包括左边界线和/或右边界线；
根据所述车底边和所述竖直边界线，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
本申请实施例还提供了一种图像识别装置，所述装置包括：
图像获取单元，用于获取包括目标车辆的待识别图像；
水平边获取单元，用于在所述待识别图像中识别所述目标车辆上的水平边；
特征获取单元，用于在所述待识别图像中识别所述目标车辆上的车底特征，所述车底特征包括车底阴影和/或车轮；
车底边确定单元，用于根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边。
可选的，所述车底边确定单元，包括：
上边界确定单元，用于确定所述车底特征的边界；
分值确定单元，用于根据所述车底特征的边界与各个所述水平边的距离，为各个所述水平边确定分值；
车底边确定子单元，用于将所述分值大于或等于预设值的水平边作为所述目标车辆的车底边。
可选的，所述特征获取单元具体用于：
根据所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车底阴影区域；和/或，
根据所述待识别图像中的像素点的像素值以及所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车轮区域。
可选的，所述装置还包括：
第一位置计算单元，用于根据所述目标车辆的车底边在所述待识别图像中的位置，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
可选的，所述装置还包括：
竖直线获取单元，用于在所述待识别图像中识别所述目标车辆上的竖直线；
竖直边界线确定单元，用于根据所述目标车辆的车底边端点特征以及所述目标车辆的对称性特征，从所述竖直线中确定所述目标车辆的竖直边界线，所述竖直边界线包括左边界线和/或右边界线；
第二位置计算单元，用于根据所述车底边和所述竖直边界线，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
本申请实施例提供了一种图像识别方法及装置，获取包括目标车辆的待识别图像，在待识别图像中识别目标车辆上的水平边和车底特征，车底特征可以包括车底阴影和/或车轮，根据目标车辆的车底特征和水平边的相对位置，从识别出的水平边中确定目标车辆的车底边。在本申请实施例中，由于车底边通常是目标车辆上最为凸出的边，可以用来表征目标车辆的位置，而车底边由于车底特征在位置上接近，因此根据车底特征和水平边的相对位置来识别车底边，可以提高车底边的识别的准确性，从而提高车辆的实际定位的准确性。
为了更清楚地说明本申请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单地介绍，显而易见地，下面描述中的附图仅仅是本申请中记载的一些实施例，对于本领域普通技术人员来讲，还可以根据这些附图获得其他的附图。
图1为现有技术中的一种目标车辆的图像示意图；
图2为本申请实施例提供的一种图像识别方法的流程图；
图3为本申请实施例提供的一种待识别图像的示意图；
图4为本申请实施例提供的另一种待识别图像的示意图；
图5为本申请实施例提供的一种图像识别装置的结构框图。
为了使本技术领域的人员更好地理解本申请方案，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，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他实施例，都属于本申请保护的范围。
目前，可以通过图像获取设备获取前方车辆的图像通过图像分析得到前方车辆的位置，具体的，可以识别出图像中目标车辆的位置，然后根据车辆在图像中的位置来计算目标车辆与获取图像的设备的相对位置，这样就可以获取目标车辆与图像获取设备所在车辆的相对位置。
现有的从图像中识别目标车辆的位置的技术，通常是对图像中的目标车辆整体进行识别，参考图1所示，为现有技术中的一种目标车辆的图像示意图，首先可以识别出目标车辆所在的区域(参考黑色矩形框)，然后根据目标车辆所在区域在图像中的位置，来计算目标车辆与获取图像的设备的相对位置。然而，发明人经过研究发现，识别出的目标车辆所在的区域通常是大于目标车辆的实际区域，这样根据识别出的区域来计算目标车辆与获取图像的设备的相对位置，通常是不够准确的。
基于此，本申请实施例提供了一种图像识别方法及装置，获取包括目标车辆的待识别图像，在待识别图像中识别目标车辆上的水平边和车底特征，车底 特征可以包括车底阴影和/或车轮，根据目标车辆的车底特征和水平边的相对位置，从识别出的水平边中确定目标车辆的车底边。在本申请实施例中，由于车底边通常是目标车辆上最为凸出的边，可以用来表征目标车辆的位置，而车底边由于车底特征在位置上接近，因此根据车底特征和水平边的相对位置来识别车底边，可以提高车底边的识别的准确性，从而提高车辆的实际定位的准确性。
下面结合附图，通过实施例来详细说明本申请实施例提供的一种图像识别方法及装置的具体实现方式。
参考图2所示为本申请实施例提供的一种图像识别方法的流程图，可以包括以下步骤。
S101，获取包括目标车辆的待识别图像。
在本申请实施例中，目标车辆为图像获取设备前方的车辆，待识别图像中是图像获取设备对目标车辆进行拍摄得到，图像获取设备可以是摄像头、相机等，也可以是其他图像获取设备。待识别图像中可以包括目标车辆的头部或尾部，也可以包括目标车辆的侧身，参考图1所示。
待识别图像中可以包括一个目标车辆，也可以包括多个目标车辆，在对待识别图像进行分析时，可以分别对待识别图像中的单个目标车辆进行分析。
可以理解的是，待识别图像也可以是对原始图像进行截取得到的，例如在原始图像中识别出包括目标车辆的候选区域作为待识别图像，这样待识别图像中就可以只包括一个目标车辆，以减少后续图像识别中的工序。对原始图像的截取可以通过机器学习模型实现。
S102，在待识别图像中识别目标车辆上的水平边和车底特征。
在待识别图像中识别目标车辆上的水平边可以有多种方式，例如通过边缘检测算法识别水平边，也可以通过计算水平方向上的像素梯度值来识别水平边，水平方向上的像素梯度值为左列的各个像素点的像素值与右列的各个像素点的像素值的差值，通常水平边的像素梯度值较小。
目标车辆上，可以有一条或多条水平边，参考图3所示，为本申请实施例提供的一种待识别图像的示意图，其中的白色线段为目标车辆上的水平边，目 标车辆上的水平边可以包括以下线段的至少一种：车顶所在的边、车牌的上边界、车牌的下边界、尾门边界、尾灯连线、车轮的下边界连线以及阴影处像素值相近的水平线等。
目标车辆的车底特征可以包括车底阴影，车底阴影是目标车辆车身之外的阴影区域，通常像素值较低，参考图3所示，其中较大的灰色椭圆为目标车辆的车底阴影区域。
在待识别图像中识别目标车辆上的车底阴影也可以有多种方式。具体的，可以根据待识别图像中像素点的像素梯度值确定目标车辆的车底阴影，其中像素梯度值可以是水平方向上的，也可以是竖直方向上的，竖直方向上的像素梯度值为上一行的各个像素点的像素值与下一行的各个像素点的像素值的差值。具体来说，车底阴影的像素值较低，通常高于待识别图像中的其他位置，以竖直方向上的像素梯度值来说，可以从下往上获取计算像素梯度值为负值的区域，到像素梯度值为正值截止，将像素梯度值为负值的区域作为车底阴影区域。事实上，这样确定出的车底阴影区域仅包括车底像素值最低的位置以上，在识别阴影区域时，还可以确定出完整的车底阴影区域，在此不再详细说明。
目标车辆的车底特征也可以包括车轮，车轮也是位于目标车辆的车底的特征，车轮的颜色从后侧放看，通常边缘较深，中间区域较浅，而从正后方看则边缘位置相较于其他区域较深，参考图3所示，其中较小的两个灰色椭圆为目标车辆的车轮区域。
在待识别图像中识别目标车辆上的车轮也可以有多种方式。具体的，可以根据待识别图像中像素点的像素梯度值以及像素值，确定目标车辆的车轮区域，像素梯度值可以是水平方向上的，也可以是竖直方向上的。具体来说，以水平方向上的像素梯度值来说，车轮区域的像素梯度值会在左右边缘位置出现最大值和最小值，同时像素值会在中间位置呈现较低值，从而确定车轮区域。
可以理解的是，目标车底特征可以包括车底阴影，也可以包括车轮，还可以同时包括车底阴影和车轮。以上示出的椭圆形车底阴影和椭圆形车轮均为一种示例，仅为了清楚说明，实际上，为了方便计算，也可以将车底阴影区域和车轮区域近似成为矩形。
S103，根据目标车辆的车底特征与水平边的相对位置，从水平边中确定目标车辆的车底边。
在本申请实施例中，在目标车辆上识别出了至少一个水平边，然而这些水平边不一定是有用的，因此需要对水平边进行筛选，具体的，可以根据目标车辆的车底特征从水平边中确定目标车辆的车底边，车底边通常是目标车辆上最为凸出的边，因此基于车底边计算得到的目标车辆与图像获取设备的相对位置是最接近实际距离的。
在实际操作中，可以确定车底特征的边界，根据车底特征的边界与水平边的距离，为水平边确定分值，将分值大于或等于预设值的水平边作为目标车辆的车底边。
具体的，车底特征的边界可以包括以下边界的至少一个：车底阴影的上边界、车底阴影的下边界、车轮的上边界和车轮的下边界。其中车底阴影的上边界或下边界可以是直线段，也可以是其他形状的曲线段，车轮的上边界或下边界可以是直线段，也可以是其他形状的曲线段，可以是一个车轮的上边界或下边界，也可以是两个车轮的上边界的连线或两个车轮的下边界的连线。
在确定车底特征的边界后，可以计算车底特征的边界与各个水平边的距离，从而计算水平边的分值。通常来说，车底阴影的下边界和车轮的下边界与车底边的距离较近，甚至于是重合的，因此可以计算水平边与车底阴影的下边界的距离，和/或，水平边与车轮的下边界的距离，设置水平边的分值与计算得到的距离成反比，这样，距离车底阴影的下边界或车辆的下边界越远的水平线的分值越小，即越不可能是车底边。
在确定水平边的分值后，可以将分值大于或等于预设值的水平边作为目标车辆的车底边，也可以将分值最高的水平边作为目标车辆的车底边，参考图3所示，可以将最靠下的一条白色线段作为车底边。
在确定目标车辆的车底边之后，可以根据目标车辆的车底边在待识别图像中的位置，计算目标车辆与获取待识别图像的图像获取设备的相对位置。参考图4所示，为本申请实施例提供的另一种待识别图像的示意图，其中白色线段为确定出的各个目标车辆的车底边，再结合图像获取设备的焦距等拍摄参数，即可计算得到各个目标车辆的车底边与图像获取设备的相对位置，实现了目标车辆的准确定位。
在本申请实施例中，还可以在待识别图像中识别目标车辆上的竖直线，竖直线的识别方式可以参考水平边的识别方式，在此不做赘述。参考图3所示， 左右两边的竖直黑色线段为竖直线。
在识别出确定竖直线后，还可以根据目标车辆的车底边端点特征以及目标车辆的对称性特征，从竖直线中确定目标车辆的竖直边界线，竖直边界线可以包括左边界线和/或右边界线。具体的，根据车底特征的左边界或右边界与竖直线的距离，为各个竖直线确定分值，将分值大于或等于预设值的竖直线作为目标车辆的竖直边界线。
通常来说，左边界线和右边界线关于中轴线对称，因此还可以根据目标车辆的对称性特征，在待识别图像中识别目标车辆上的中轴线，参考图3所示，位于中间位置的黑色线段为目标车辆的中轴线，目标车辆关于中轴线对称。在根据分值确定竖直边界线时，也可以计算关于中轴线对称的两条竖直线的分数和，将分数和最大的一对竖直线作为左边界线和右边界线。
在本申请实施例中，还可以在确定目标车辆的车底边之后，根据目标车辆的车底边的端点确定目标车辆的竖直边界线，根据车底边和竖直边界线，计算目标车辆与获取待识别图像的图像获取设备的相对位置。具体的，可以过车底边的两个端点作出垂直于车底边的线段，即为竖直边界线。
在确定目标车辆的竖直边界线之后，可以根据目标车辆的车底边和竖直边界线在待识别图像中的位置，计算目标车辆与获取待识别图像的图像获取设备的相对位置。参考图4所示，其中白色线段为确定出的各个目标车辆的车底边，黑色线段为确定出的各个目标车辆的竖直边界线，再结合图像获取设备的焦距等拍摄参数，即可计算得到各个目标车辆的车底边与图像获取设备的相对位置，实现了目标车辆的准确定位。
本申请实施例提供了一种图像识别方法，获取包括目标车辆的待识别图像，在待识别图像中识别目标车辆上的水平边和车底特征，车底特征可以包括车底阴影和/或车轮，根据目标车辆的车底特征和水平边的相对位置，从识别出的水平边中确定目标车辆的车底边。在本申请实施例中，由于车底边通常是目标车辆上最为凸出的边，可以用来表征目标车辆的位置，而车底边由于车底特征接近，因此根据车底特征和水平边的相对位置来识别车底边，可以提高车底边的识别的准确性，从而提高车辆的实际定位的准确性。
基于以上一种图像识别方法，本申请实施例还提供了一种图像识别装置， 参考图5所示，为本申请实施例提供的一种图像识别装置的结构框图，所述装置包括：
图像获取单元110，用于获取包括目标车辆的待识别图像；
水平边获取单元120，用于在所述待识别图像中识别所述目标车辆上的水平边；
特征获取单元130，用于在所述待识别图像中识别所述目标车辆上的车底特征，所述车底特征包括车底阴影和/或车轮；
车底边确定单元140，用于根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边。
可选的，所述车底边确定单元，包括：
上边界确定单元，用于确定所述车底特征的边界；
分值确定单元，用于根据所述车底特征的边界与各个所述水平边的距离，为各个所述水平边确定分值；
车底边确定子单元，用于将所述分值大于或等于预设值的水平边作为所述目标车辆的车底边。
可选的，所述特征获取单元具体用于：
根据所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车底阴影区域；和/或，
根据所述待识别图像中的像素点的像素值以及所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车轮区域。
可选的，所述装置还包括：
第一位置计算单元，用于根据所述目标车辆的车底边在所述待识别图像中的位置，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
可选的，所述装置还包括：
竖直线获取单元，用于在所述待识别图像中识别所述目标车辆上的竖直线；
竖直边界线确定单元，用于根据所述目标车辆的车底边左右端点特征以及 所述目标车辆的对称性特征，从所述竖直线中确定所述目标车辆的竖直边界线，所述竖直边界线包括左边界线和/或右边界线；
第二位置计算单元，用于根据所述车底边和所述竖直边界线，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
本申请实施例提供了一种图像识别装置，获取包括目标车辆的待识别图像，在待识别图像中识别目标车辆上的水平边和车底特征，车底特征可以包括车底阴影和/或车轮，根据目标车辆的车底特征和水平边的相对位置，从识别出的水平边中确定目标车辆的车底边。在本申请实施例中，由于车底边通常是目标车辆上最为凸出的边，可以用来表征目标车辆的位置，而车底边由于车底特征在位置上接近，因此根据车底特征和水平边的相对位置来识别车底边，可以提高车底边的识别的准确性，从而提高车辆的实际定位的准确性。
本申请实施例中提到的“第一……”、“第一……”等名称中的“第一”只是用来做名字标识，并不代表顺序上的第一。该规则同样适用于“第二”等。
通过以上的实施方式的描述可知，本领域的技术人员可以清楚地了解到上述实施例方法中的全部或部分步骤可借助软件加通用硬件平台的方式来实现。基于这样的理解，本申请的技术方案可以以软件产品的形式体现出来，该计算机软件产品可以存储在存储介质中，如只读存储器(英文：read-only memory，ROM)/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者诸如路由器等网络通信设备)执行本申请各个实施例或者实施例的某些部分所述的方法。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于装置实施例而言，由于其基本相似于方法实施例，所以描述得比较简单，相关之处参见方法实施例的部分说明即可。以上所描述的设备及系统实施例仅仅是示意性的，其中作为分离部件说明的模块可以是或者也可以不是物理上分开的，作为模块显示的部件可以是或者也可以不是物理模块，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性劳动的情况下，即可以理解并实施。
以上所述仅是本申请的优选实施方式，并非用于限定本申请的保护范围。应当指出，对于本技术领域的普通技术人员来说，在不脱离本申请的前提下，还可以作出若干改进和润饰，这些改进和润饰也应视为本申请的保护范围。
Claims (10)
- 一种图像识别方法，其特征在于，所述方法包括：获取包括目标车辆的待识别图像；在所述待识别图像中识别所述目标车辆上的水平边和车底特征，所述车底特征包括车底阴影和/或车轮；根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边。
- 根据权利要求1所述的方法，其特征在于，所述根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边，包括：确定所述车底特征的边界；根据所述车底特征的边界与各个所述水平边的距离，为各个所述水平边确定分值；将所述分值大于或等于预设值的水平边作为所述目标车辆的车底边。
- 根据权利要求1或2所述的方法，其特征在于，在所述待识别图像中识别所述目标车辆上的车底阴影，包括：根据所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车底阴影区域；在所述待识别图像中识别所述目标车辆上的车轮，包括：根据所述待识别图像中的像素点的像素值以及所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车轮区域。
- 根据权利要求1-3任意一项所述的方法，其特征在于，所述方法还包括：根据所述目标车辆的车底边在所述待识别图像中的位置，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
- 根据权利要求1-3任意一项所述的方法，其特征在于，所述方法还包括：在所述待识别图像中识别所述目标车辆上的竖直线；根据所述目标车辆的车底边端点特征以及所述目标车辆的对称性特征，从所述竖直线中确定所述目标车辆的竖直边界线，所述竖直边界线包括左边界线 和/或右边界线；根据所述车底边和所述竖直边界线，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
- 一种图像识别装置，其特征在于，所述装置包括：图像获取单元，用于获取包括目标车辆的待识别图像；水平边获取单元，用于在所述待识别图像中识别所述目标车辆上的水平边；特征获取单元，用于在所述待识别图像中识别所述目标车辆上的车底特征，所述车底特征包括车底阴影和/或车轮；车底边确定单元，用于根据所述目标车辆的车底特征与所述水平边的相对位置，从所述水平边中确定所述目标车辆的车底边。
- 根据权利要求6所述的装置，其特征在于，所述车底边确定单元，包括：上边界确定单元，用于确定所述车底特征的边界；分值确定单元，用于根据所述车底特征的边界与各个所述水平边的距离，为各个所述水平边确定分值；车底边确定子单元，用于将所述分值大于或等于预设值的水平边作为所述目标车辆的车底边。
- 根据权利要求6或7所述的装置，其特征在于，所述特征获取单元具体用于：根据所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车底阴影区域；和/或，根据所述待识别图像中的像素点的像素值以及所述待识别图像中的像素点的像素梯度值，确定所述目标车辆的车轮区域。
- 根据权利要求6-8任意一项所述的装置，其特征在于，所述装置还包括：第一位置计算单元，用于根据所述目标车辆的车底边在所述待识别图像中的位置，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
- 根据权利要求6-8任意一项所述的装置，其特征在于，所述装置还包括：竖直线获取单元，用于在所述待识别图像中识别所述目标车辆上的竖直线；竖直边界线确定单元，用于根据所述目标车辆的车底边端点特征以及所述目标车辆的对称性特征，从所述竖直线中确定所述目标车辆的竖直边界线，所述竖直边界线包括左边界线和/或右边界线；第二位置计算单元，用于根据所述车底边和所述竖直边界线，计算所述目标车辆与获取所述待识别图像的图像获取设备的相对位置。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910213120.2 | 2019-03-20 | ||
| CN201910213120.2A CN110059566A (zh) | 2019-03-20 | 2019-03-20 | 一种图像识别方法及装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020187311A1 true WO2020187311A1 (zh) | 2020-09-24 |
Family
ID=67317115
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2019/084536 WO2020186603A1 (zh) | 2019-03-20 | 2019-04-26 | 一种图像识别方法及装置 |
| PCT/CN2020/080321 WO2020187311A1 (zh) | 2019-03-20 | 2020-03-20 | 一种图像识别方法及装置 |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2019/084536 WO2020186603A1 (zh) | 2019-03-20 | 2019-04-26 | 一种图像识别方法及装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110059566A (zh) |
| WO (2) | WO2020186603A1 (zh) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110059566A (zh) * | 2019-03-20 | 2019-07-26 | 东软睿驰汽车技术（沈阳）有限公司 | 一种图像识别方法及装置 |
| CN111290386B (zh) * | 2020-02-20 | 2023-08-04 | 北京小马慧行科技有限公司 | 路径规划方法及装置、运载工具 |
| CN115817463B (zh) * | 2023-02-23 | 2023-05-26 | 禾多科技（北京）有限公司 | 车辆避障方法、装置、电子设备和计算机可读介质 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004355139A (ja) * | 2003-05-27 | 2004-12-16 | Nissan Motor Co Ltd | 車両認識装置 |
| CN101029824A (zh) * | 2006-02-28 | 2007-09-05 | 沈阳东软软件股份有限公司 | 基于车辆特征的车辆定位方法和装置 |
| KR20130000023A (ko) * | 2011-06-22 | 2013-01-02 | (주)새하소프트 | 영상의 장면 정보를 이용한 전방 차량 검출 방법 |
| JP2014034219A (ja) * | 2012-08-07 | 2014-02-24 | Nissan Motor Co Ltd | 車両走行支援装置及び車両走行支援方法 |
| CN105551046A (zh) * | 2015-12-17 | 2016-05-04 | 浙江宇视科技有限公司 | 车脸定位方法及装置 |
| CN105844222A (zh) * | 2016-03-18 | 2016-08-10 | 上海欧菲智能车联科技有限公司 | 基于视觉的前方车辆碰撞预警系统及方法 |
| CN109389064A (zh) * | 2018-09-27 | 2019-02-26 | 东软睿驰汽车技术（沈阳）有限公司 | 一种车辆特征获取方法及装置 |
| CN110059566A (zh) * | 2019-03-20 | 2019-07-26 | 东软睿驰汽车技术（沈阳）有限公司 | 一种图像识别方法及装置 |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7720580B2 (en) * | 2004-12-23 | 2010-05-18 | Donnelly Corporation | Object detection system for vehicle |
| CN104915642B (zh) * | 2015-05-26 | 2018-05-11 | 奇瑞汽车股份有限公司 | 前方车辆测距方法及装置 |
| CN106952478A (zh) * | 2017-05-11 | 2017-07-14 | 南宁市正祥科技有限公司 | 一种基于对称性的车辆检测方法 |
| CN108319910B (zh) * | 2018-01-30 | 2021-11-16 | 海信集团有限公司 | 一种车辆识别方法、装置及终端 |
-
2019
- 2019-03-20 CN CN201910213120.2A patent/CN110059566A/zh active Pending
- 2019-04-26 WO PCT/CN2019/084536 patent/WO2020186603A1/zh active Application Filing
-
2020
- 2020-03-20 WO PCT/CN2020/080321 patent/WO2020187311A1/zh active Application Filing
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004355139A (ja) * | 2003-05-27 | 2004-12-16 | Nissan Motor Co Ltd | 車両認識装置 |
| CN101029824A (zh) * | 2006-02-28 | 2007-09-05 | 沈阳东软软件股份有限公司 | 基于车辆特征的车辆定位方法和装置 |
| KR20130000023A (ko) * | 2011-06-22 | 2013-01-02 | (주)새하소프트 | 영상의 장면 정보를 이용한 전방 차량 검출 방법 |
| JP2014034219A (ja) * | 2012-08-07 | 2014-02-24 | Nissan Motor Co Ltd | 車両走行支援装置及び車両走行支援方法 |
| CN105551046A (zh) * | 2015-12-17 | 2016-05-04 | 浙江宇视科技有限公司 | 车脸定位方法及装置 |
| CN105844222A (zh) * | 2016-03-18 | 2016-08-10 | 上海欧菲智能车联科技有限公司 | 基于视觉的前方车辆碰撞预警系统及方法 |
| CN109389064A (zh) * | 2018-09-27 | 2019-02-26 | 东软睿驰汽车技术（沈阳）有限公司 | 一种车辆特征获取方法及装置 |
| CN110059566A (zh) * | 2019-03-20 | 2019-07-26 | 东软睿驰汽车技术（沈阳）有限公司 | 一种图像识别方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020186603A1 (zh) | 2020-09-24 |
| CN110059566A (zh) | 2019-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12002269B2 (en) | Vehicle feature acquisition method and device | |
| WO2020187311A1 (zh) | 一种图像识别方法及装置 | |
| CN103824066B (zh) | 一种基于视频流的车牌识别方法 | |
| US8098933B2 (en) | Method and apparatus for partitioning an object from an image | |
| US9245188B2 (en) | Lane detection system and method | |
| CN104899554A (zh) | 一种基于单目视觉的车辆测距方法 | |
| JP6769477B2 (ja) | 画像処理装置、撮像装置、移動体機器制御システム、画像処理方法、及びプログラム | |
| US11164012B2 (en) | Advanced driver assistance system and method | |
| CN110008891B (zh) | 一种行人检测定位方法、装置、车载计算设备及存储介质 | |
| WO2020133488A1 (zh) | 车辆检测方法和设备 | |
| CN110991264A (zh) | 前方车辆检测方法和装置 | |
| CN106022268A (zh) | 一种限速标识的识别方法和装置 | |
| WO2019149213A1 (zh) | 基于影像识别路锥的方法、装置、存储介质以及车辆 | |
| JP2014127027A (ja) | 境界線認識装置 | |
| US10970570B2 (en) | Device and a method for clustering light spots | |
| CN112400094B (zh) | 物体探测装置 | |
| CN111376904B (zh) | 一种自动跟车方法及装置 | |
| US20240203134A1 (en) | Lane line detection method, vehicle-mounted device and storage medium | |
| JPH10261065A (ja) | 車線認識装置 | |
| KR20190103503A (ko) | 차선 인식 장치 및 방법 | |
| Chingting et al. | Traffic Lane Line Classification System by Real-time Image Processing | |
| JP2001005979A (ja) | 識別関数学習装置及び車両検出装置 | |
| Rosebrock et al. | Using the shadow as a single feature for real-time monocular vehicle pose determination | |
| CN116977980A (zh) | 一种车辆交通路口通行检测方法及装置 | |
| KR20240002501A (ko) | 차량 및 그 제어 방법 |
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: 20774482 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: 20774482 Country of ref document: EP Kind code of ref document: A1 |