WO2020221012A1 - 图像特征点的运动信息确定方法、任务执行方法和设备 - Google Patents

图像特征点的运动信息确定方法、任务执行方法和设备 Download PDF

Info

Publication number
WO2020221012A1
WO2020221012A1 PCT/CN2020/085000 CN2020085000W WO2020221012A1 WO 2020221012 A1 WO2020221012 A1 WO 2020221012A1 CN 2020085000 W CN2020085000 W CN 2020085000W WO 2020221012 A1 WO2020221012 A1 WO 2020221012A1
Authority
WO
WIPO (PCT)
Prior art keywords
pixel
image
target
pixel area
feature point
Prior art date
Application number
PCT/CN2020/085000
Other languages
English (en)
French (fr)
Inventor
凌永根
张晟浩
Original Assignee
腾讯科技（深圳）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 腾讯科技（深圳）有限公司 filed Critical 腾讯科技（深圳）有限公司
Priority to JP2021529019A priority Critical patent/JP7305249B2/ja
Priority to EP20798875.9A priority patent/EP3965060A4/en
Publication of WO2020221012A1 publication Critical patent/WO2020221012A1/zh
Priority to US17/323,740 priority patent/US20210272294A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence

Definitions

  • This application relates to the field of Internet technology, and in particular to a method for determining motion information of image feature points, a task execution method and equipment.
  • Feature point tracking technology is the process of analyzing the position changes of the same feature point in consecutive multiple frames of images.
  • motion information is usually used to describe the change process of feature points in different images.
  • the process of determining the motion information of the image feature points may include: the computer device obtains the pixel area where the feature points are located in the current frame of image , In order to make the result accurate, the pixel area is usually selected larger. The computer device determines the motion information of image feature points based on multiple pixel points in multiple pixel areas and multiple pixel points in multiple corresponding pixel areas in a subsequent frame of image.
  • the traditional solution due to the large selected pixel area, the number of pixels to be calculated is large, resulting in low efficiency in the process of determining the motion information of the image feature points.
  • a method, device, computer device, and storage medium for determining motion information of image feature points are provided, and a task execution method, device, target device, and storage medium are also provided.
  • a method for determining motion information of image feature points which is executed by a computer device, and includes:
  • a second pixel area including the target feature point in the first image is determined, and the second pixel area is The pixel difference of the plurality of second pixels is greater than the pixel difference of the plurality of first pixels, the number of the plurality of second pixels is the same as the number of the plurality of first pixels, and the pixel difference is used to indicate The degree of change in pixel values of multiple pixels;
  • the motion information of the target feature point is acquired, and the motion information is used to indicate that the target feature point is in the first image and the second image The position changes in.
  • a task execution method executed by a target device, the method including:
  • the target task is executed.
  • a device for determining motion information of image feature points comprising:
  • a determining module configured to determine a first image and a second image, where the first image and the second image include the same object
  • the determining module is further configured to determine a first pixel area in the first image that includes the target feature point based on the target feature point on the object in the first image;
  • the determining module is further configured to determine the second pixel area in the first image that includes the target feature point according to the pixel difference of the multiple first pixel points in the first pixel area and the target feature point ,
  • the pixel difference of the plurality of second pixels in the second pixel area is greater than the pixel difference of the plurality of first pixels, and the number of the plurality of second pixels is the same as the number of the plurality of first pixels ,
  • the pixel difference is used to indicate the degree of change of the pixel values of multiple pixels;
  • the acquiring module is configured to acquire the motion information of the target feature point according to the plurality of second pixel points and the second image, and the motion information is used to indicate that the target feature point is in the first image and The position in the second image changes.
  • the determining module is further configured to, when the pixel difference of a plurality of first pixel points in the first pixel area is less than a target difference value, obtain a value greater than the first pixel difference according to the target feature point.
  • the pixel area includes the second pixel area of the target feature point.
  • the determining module is further configured to, when the pixel difference of the plurality of first pixel points in the first pixel area is less than the target difference value, obtain the difference from the first pixel point according to the target feature point. For a second pixel area with the same size of the pixel area, the pixel points included in the first pixel area and the second pixel area are different.
  • the determining module is further configured to expand the first pixel area into a second pixel area including the target feature point according to a target expansion coefficient and centering on the target feature point.
  • the determining module is further configured to move the first pixel area to a second pixel area including the target feature point according to the target feature point and according to the target movement track.
  • the determining module is further configured to obtain a third pixel that includes the target feature point according to the target feature point when the pixel difference of the plurality of first pixel points is less than the target difference value Area, the number of the plurality of third pixels in the third pixel area is the same as the number of the plurality of first pixels; the number of the plurality of pixels is determined according to the pixel values of the plurality of third pixels in the third pixel area Pixel differences of three third pixels; and when the pixel differences of the plurality of third pixels are not less than the target difference value, the third pixel area is determined as the second pixel area.
  • the determining module is further configured to increase the first sampling step of the first pixel area to the first pixel area according to the expansion coefficient from the first pixel area to the third pixel area.
  • a two-sampling step size according to the second sampling step size, obtain third pixel points with the same number as the plurality of first pixel points from the third pixel area.
  • the determining module is further configured to obtain, from the third pixel area, the same number of first pixels as the plurality of first pixels according to the first sampling step of the first pixel area. Three pixels.
  • the device further includes:
  • a detection module configured to detect whether the size of the third pixel area is greater than a target threshold when the pixel difference of the plurality of third pixels is smaller than the target difference value
  • the determining module is further configured to determine a fourth pixel area larger than the third pixel area when the size of the third pixel area is not greater than the target threshold;
  • the determining module is further configured to determine a second pixel area in the first image that includes the target feature point based on the pixel difference of a plurality of fourth pixel points in the fourth pixel area and the target feature point ,
  • the number of the plurality of fourth pixels is the same as the number of the plurality of first pixels.
  • the pixel difference is the pixel variance of the multiple pixels or the smallest feature value of the gradient matrix of the multiple pixels, and the pixel variance is used to represent the pixels of the multiple pixels.
  • the degree of change of the value relative to the average value of the pixel, and the gradient matrix is used to indicate the degree of change of the pixel values of the plurality of pixels in the horizontal gradient relative to the pixel average value and the change in the vertical gradient relative to the pixel average value. degree.
  • the shape of the first pixel area and the second pixel area is any shape of a square, a rectangle, a circle, a ring, an irregular polygon, or an irregular curve.
  • the determining module is further configured to obtain the multiple pixels among multiple pixel points on the area boundary of the first pixel area according to the first sampling step of the first pixel area. First pixel points; and determining the pixel difference of the plurality of first pixel points according to the pixel values of the plurality of first pixel points.
  • a task execution device includes:
  • An obtaining module configured to obtain a first image and a second image, the first image and the second image including the same object
  • the acquisition module is further configured to acquire the motion information of the target feature point on the object in the first image, and the motion information is used to indicate that the target feature point is in the first image and the second image. Position changes in the image;
  • the task processing module is used to execute the target task based on the motion information of the target feature point.
  • the target task includes a route planning task
  • the task processing module is further configured to determine a distance target based on the movement information of multiple target feature points when the number of target feature points is multiple. At least one scene object whose device does not exceed a first threshold, and the target device is a device that collects the first image and the second image; and according to the location of the destination whose distance to the target device does not exceed the second threshold and the At least one scene object determines a first target route for the target device to reach the destination, and the second threshold is greater than the first threshold.
  • the target task includes an object recognition task
  • the task processing module is further configured to determine the target feature point based on the movement information of the multiple target feature points when the number of the target feature points is multiple.
  • the multiple first feature points whose motion information meets the target condition among the multiple target feature points, and the multiple first feature points are used to indicate the motion in multiple objects included in the first image and the second image Object; and determining the object category to which the moving object belongs based on the positions of the plurality of first feature points in the first image or the second image.
  • a computer device including a processor and a memory, and computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the processor, the processor executes the image feature The steps of the method for determining the movement information of the point.
  • a target device includes a processor and a memory, and computer-readable instructions are stored in the memory.
  • the processor executes the task execution Method steps.
  • a non-volatile computer-readable storage medium that stores computer-readable instructions.
  • the one or more processors execute the image feature The steps of the method for determining the movement information of the point.
  • a non-volatile computer-readable storage medium that stores computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the task execution Method steps.
  • FIG. 1 is a schematic diagram of an implementation environment of a method for determining motion information of image feature points provided by an embodiment of the present application
  • FIG. 2 is a flowchart of a method for determining motion information of image feature points according to an embodiment of the present application
  • FIG. 3 is a schematic diagram of a pixel area provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of a pixel position provided by an embodiment of the present application.
  • FIG. 5 is a schematic diagram of motion information of a pixel provided by an embodiment of the present application.
  • Fig. 6 is a schematic diagram of a process for determining motion information of image feature points according to an embodiment of the present application
  • FIG. 7 is a flowchart of a task execution method provided by an embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of an apparatus for determining motion information of image feature points according to an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of a task execution device provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of a terminal provided by an embodiment of the present application.
  • FIG. 11 is a schematic structural diagram of a server provided by an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an implementation environment of a method for determining motion information of image feature points provided by an embodiment of the present application.
  • the implementation environment includes: a computer device 101 that can obtain information with the same object With two or more frames of images, the computer device 101 can determine the movement information of the target feature point on the object from one frame of image to another frame of image, so as to track the position change of the target feature point in different images,
  • Feature points refer to pixels with prominent pixel characteristics in an image.
  • the computer device 101 can acquire two or more consecutive frames of images in the video, for example, the first image and the second image, and the computer device 101 can acquire target feature points from the first image.
  • the movement information to the second image the movement information is used to indicate the position change of the target feature point in the first image and the second image.
  • the computer device 101 may perform position tracking based on the feature points in the pixel area of the image.
  • the pixel area of the feature point includes the target feature point pixel.
  • the computer device can determine that the target feature point is in the first pixel area of the first image, and based on the pixel difference of the plurality of first pixel points in the first pixel area, determine the second pixel with the larger pixel difference in the first image Area, and the second pixel area has the same number of pixels as the first pixel area, so that the target feature points from the first image to the second image can be determined based on the multiple second pixels and the second image in the second pixel area Sports information.
  • the implementation environment may further include: a computer device 102, the computer device 102 may collect more than one frame of image, and send the more than one frame of image to the computer device 101.
  • the computer device 101 can be provided as any device such as a server or a terminal, or any device such as an intelligent robot, an unmanned vehicle, a drone, and an unmanned vessel.
  • the computer device 102 may be provided as a mobile phone, a camera, a terminal, a monitoring device, etc.
  • the embodiment of the present application does not specifically limit the actual form of the computer device 101 and the computer device 102.
  • Fig. 2 is a flowchart of a method for determining motion information of image feature points according to an embodiment of the present application.
  • the execution subject of this embodiment of the invention is a computer device. Referring to FIG. 2, the method includes:
  • the computer device determines the first image and the second image.
  • the first image and the second image include the same object, and the object can be any object with an actual display form, for example, a house, a road sign, or a vehicle.
  • the first image and the second image may be two frames of images including the same object in the video, and the time stamp of the second image may be after the first image.
  • the computer device can capture the surrounding environment in real time to obtain a video, and the computer device can determine the first image in the video and the second image whose time stamp is located after the first image.
  • the computer device may be any device such as a movable unmanned vehicle, an intelligent robot, or a drone.
  • the computer device can also be a mobile device with a camera installed, and the computer device can also record a video during the movement.
  • the computer device may obtain the first image and the second image from the target device.
  • the target device can take a first image and a second image that include the same object during movement
  • the computer device receives a video sent by the target device, and obtains the first image and the second image in the video
  • the target device It can be a surveillance device, a mobile phone, or a camera.
  • the computer device may also determine the target feature point on the object in the first image.
  • the computer device can extract the target feature points on the object in the first image through a target algorithm.
  • the target algorithm can be SIFT (Scale-invariant feature transform) algorithm or SURF (Speeded Up Robust Features) , Accelerate robust features) algorithms, etc.
  • the target feature point refers to a point with a significant pixel feature in an image, for example, the difference between the pixel value and the pixel values of multiple surrounding pixels is greater than the target pixel difference. Based on the difference in image resolution, the number of sampled pixels in the same image is also different. Sampling pixels refer to the pixels where the image is sampled based on the target resolution.
  • the target feature point may be a sampling pixel point in the first image, or a pixel point located between two sampling pixel points in the first image.
  • the computer device may use integer coordinates to indicate the position of the sampling pixel in the first image, and integer coordinates mean that the values of the coordinates are all integers.
  • the coordinates of the target feature point can be integer coordinates, and when the target feature point is a pixel between two sampled pixels, the coordinates of the target feature point can also be non- Integer coordinates.
  • the computer device determines a first pixel area in the first image that includes the target feature point according to the target feature point on the object in the first image.
  • the computer device may determine, in the first image, a first pixel region that includes the target feature point and has a shape of the target shape according to the position of the target feature point.
  • the target shape of the first pixel area may be any shape of a square, a rectangle, a circle, a ring, an irregular polygon or a curved side.
  • the shape of the first pixel area may be a center-symmetric shape, and the computer device uses the target feature point as a center to determine the first pixel area of the target shape.
  • the first pixel area is an area including at least a first target number of pixels.
  • the computer device may also regard the target feature point as the center according to the first target quantity and the target shape, and use an area of the target shape that includes not less than the first target quantity as the first pixel area.
  • the computer device determines a square area with the target feature point as the center and the side length as the first side length as the first pixel area according to the first target number; The square of the first side length is not less than the first target quantity. For example, if the target number is 9, the first side length can be an integer of 3 or more, for example, the side length is 5 or 9, and so on.
  • the computer device determines a circular area centered on the target feature point and a radius of the target radius as the first pixel area according to the number of targets; when the target shape is a rectangle, a diamond, or a circle In the case of a ring shape or any other shape, the computer device has the same principle for determining the first pixel area, and will not be repeated here.
  • the pixel points in the first pixel area reflect the pixel features around the target feature point, and the computer device can also select a part of the pixels in the first pixel area to represent the pixel features around the target feature point.
  • the computer device may also adopt a sampling method to determine a plurality of first pixel points from the first pixel area, and the plurality of first pixel points reflect changes in pixel characteristics around the target characteristic point.
  • the computer device determines the first sampling step size according to the total number of pixels in the first pixel area and the second target number, and according to the first sampling step size, every first sampling from the plurality of pixels in the first pixel area For the pixel points of the step size, one first pixel point is obtained, and the first pixel point of the second target quantity is obtained.
  • the first sampling step is 1.
  • the target number of first pixel points may be evenly distributed on the boundary of the first pixel region with the target feature point as the center, and the computer device will determine the pixels located on the boundary of the first pixel region. The total number of pixels of a point is divided by the target number, and the obtained quotient is used as the first sampling step.
  • the computer device starts from a plurality of pixels located on the boundary of the first pixel area. , Obtain the first pixel of the second target quantity.
  • the computer device may also use a non-uniform sampling method to filter out the first pixel point.
  • the computer device may also filter the first pixel points according to the distance between the multiple pixels in the first pixel area and the target feature point, and the computer device obtains the multiple pixels and the The multiple distances between the target feature points, according to the multiple distances, the multiple pixels are divided into multiple pixel point sets, and the distance between the pixel point and the target feature point in each pixel point set is located at the pixel point Within the distance range corresponding to the collection.
  • the distance range corresponding to set A is 1.0 to 4.9
  • the distance range corresponding to set B is 5.0 to 9.9.
  • the computer device determines the first sampling step corresponding to the pixel point set according to the distance between each pixel point set and the target feature point, and according to the first sampling step corresponding to the pixel point set
  • the step size is to obtain multiple first pixel points among multiple pixels in the pixel point set, so as to obtain the second target number of first pixels.
  • the distance between the pixel point set and the target feature point may be the average value of the multiple distances between multiple pixels in the pixel point set and the target feature point.
  • the computer device can select a dense center with a sparse edge, that is, the larger the distance between the pixel point set and the target feature point, the smaller the first sampling step corresponding to the pixel point set.
  • the computer device may also adopt a random sampling method to filter out the second target number of first pixels from the first pixel area.
  • a square pixel area is taken as an example for description.
  • the black points are pixels and the white points are filtered out pixels.
  • the computer device selects from multiple pixels on the boundary of the square area. Filter out the pixels.
  • the first sampling step may be 1, and when the number of pixels on the boundary of the pixel area is greater than the second target number, the sampling step is Can be greater than 1.
  • the second target number may be 9, as shown in FIG. 3(a), the pixel area is a square area including 3 ⁇ 3 pixels, and the sampling step may be 1 pixel.
  • the pixel area may be a square area including 5 ⁇ 5 pixels, and the sampling step may be 2 pixels.
  • the pixel area may be a 9 ⁇ 9 square area, and the sampling step may be 4 pixels.
  • S203 The computer device determines the pixel difference of the plurality of first pixels according to the pixel values of the plurality of first pixels.
  • the pixel difference is used to indicate the degree of change of the pixel values of multiple pixels.
  • the pixel difference of the multiple pixels may be the pixel variance of the multiple pixels or the smallest value of the gradient matrix of the multiple pixels.
  • Characteristic value the pixel variance of the multiple pixels is used to indicate the degree of change of the pixel values of the multiple pixels relative to the pixel average value
  • the gradient matrix of the multiple pixels is used to represent the pixel values of the multiple pixels, respectively The degree of change from the average on the horizontal gradient and the degree of change from the average on the vertical gradient.
  • the computer device may use the pixel variance of the plurality of first pixels or the minimum characteristic value of the gradient matrix of the plurality of first pixels to represent the pixel difference of the plurality of first pixels.
  • the computer determines the average value of the pixel values of the plurality of first pixels, and determines the average value of the plurality of first pixels according to the difference between the pixel value of each first pixel and the average value. Pixel difference.
  • the computer device can use the following formula 1 according to the average value of the pixel values of the plurality of first pixels and the pixel value of each first pixel: Determine the pixel variance of the plurality of first pixels;
  • N is the number of the plurality of first pixels
  • l is 3
  • M is the average value of the pixel values of the plurality of first pixels
  • I u is the pixel value of the multiple first pixels
  • var represents the pixel square difference of the multiple first pixels.
  • u represents multiple first pixels.
  • the computer device may also use the gradient matrix of the plurality of first pixels to represent the pixel difference. Then the computer device can determine the horizontal gradient of the pixel value and the vertical gradient of the pixel value according to the pixel value of the plurality of first pixels, and the horizontal gradient of the pixel value and the vertical gradient of the pixel value.
  • Gradient Obtain the minimum characteristic value of the gradient matrix of the plurality of first pixels, and determine the minimum characteristic value as the pixel difference of the plurality of first pixels.
  • the gradient matrix may be a covariance matrix, and the variables are vertical gradients and horizontal gradients of pixel values of multiple pixels. Then the computer device can determine the minimum eigenvalue of the gradient matrix of the plurality of first pixels according to the horizontal gradient of the pixel values of the plurality of first pixels and the vertical gradient of the pixel values through the following formula two;
  • g x (u) is the horizontal gradient of the pixel values of the multiple first pixels u
  • g y (u) is the vertical gradient of the pixel values of the multiple first pixels u
  • u represents the multiple first pixels .
  • the matrix S represents the gradient matrix of a plurality of first pixels.
  • the computer device can determine the minimum eigenvalue of the matrix S to obtain the pixel difference of the plurality of first pixels.
  • the coordinates of the target feature point may be non-integer coordinates, so the coordinates of the first pixel point may also be non-integer coordinates.
  • the device may determine the pixel value of the first pixel according to the pixel values of the surrounding pixels of the first pixel.
  • the computer device may use a bidirectional linear interpolation algorithm to determine the pixel value of a pixel at a non-integer position, the computer device obtains the pixel value of the surrounding pixel of the first pixel, and determines according to the following formula 3. The pixel value of the first pixel;
  • I represents the pixel value of the first pixel
  • i0 represents the pixel value of the upper left corner of the first pixel
  • the position is represented as (u0, v0)
  • i1 represents the upper right corner of the first pixel.
  • the pixel value of the pixel, the position is expressed as (u1, v1)
  • i2 is the pixel value of the lower left corner of the first pixel
  • the position is expressed as (u2, v2)
  • i3 is the lower right pixel of the first pixel
  • the pixel value of the point, the position is expressed as (u3, v3)
  • (u, v) is the position of the first pixel.
  • the computer device determines a second pixel area in the first image that includes the target feature point according to the pixel difference of the plurality of first pixel points in the first pixel area and the target feature point.
  • the pixel difference of the plurality of second pixels in the second pixel area is greater than the pixel difference of the plurality of first pixels, and the number of the plurality of second pixels is the same as the number of the plurality of first pixels.
  • the computer device obtains a target difference value, and when the pixel difference of the plurality of first pixels is smaller than the target difference value, the computer device adjusts the first pixel area to obtain a second pixel whose pixel difference is greater than the target difference value area.
  • the computer device may obtain a second pixel area with greater pixel difference by expanding the first pixel area, or obtain a second pixel area by moving the first pixel area.
  • this step can be implemented in the following two ways.
  • the computer device obtains a second pixel area larger than the first pixel area and including the target characteristic point according to the target feature point.
  • the computer device may expand the first pixel area according to a certain expansion rule, or randomly expand the first pixel area.
  • the shape of the first pixel area may be a center-symmetrical shape, and the computer device expands the first pixel area to include the first pixel area of the target feature point according to the target expansion coefficient and centering on the target feature point.
  • the computer device may adjust the first pixel area multiple times to obtain a second pixel area with a pixel difference greater than the target difference value.
  • the process may be: when the pixel difference of the plurality of first pixel points is less than the target difference value, the computer device may obtain a third pixel area including the target characteristic point according to the target characteristic point, and the third pixel area is The number of the plurality of third pixels is the same as the number of the plurality of first pixels; the computer device determines the pixel difference of the plurality of third pixels according to the pixel values of the plurality of third pixels in the third pixel area; If the pixel difference of the plurality of third pixel points is not less than the target difference value, the computer device determines the third pixel area as the second pixel area.
  • the computer device may expand the first pixel area into a third pixel area including the target feature point according to the target expansion coefficient and centering on the target feature point.
  • the computer device may also acquire multiple third pixel points from the third pixel area according to a certain sampling step.
  • the process of the computer device acquiring a plurality of third pixel points in the third pixel area may include: the computer device may according to an expansion coefficient from the first pixel area to the third pixel area, The first sampling step size of the first pixel area is increased to the second sampling step size, and according to the second sampling step size, the third pixel area with the same quantity as the plurality of first pixels is obtained from the third pixel area.
  • pixel For example, for a 3 ⁇ 3 square, the computer device collects a first pixel every two pixels, and for a 5 ⁇ 5 square, the computer device collects a third pixel every three pixels.
  • the computer device continues to obtain and expand the third pixel area until the second pixel area whose pixel difference is not less than the target difference value is obtained.
  • the process of obtaining the pixel differences of the plurality of third pixels by the computer device is the same as the process of obtaining the pixel differences of the plurality of first pixels, and will not be repeated here.
  • the computer device may also determine whether to continue expanding the third pixel area based on the size of the pixel area.
  • the process may include: when the pixel difference of the plurality of third pixel points is less than the target difference value, the computer device detects whether the size of the third pixel area is greater than the target threshold; when the size of the third pixel area is not greater than the target threshold , The computer device determines a fourth pixel area that is larger than the third pixel area; the computer device determines that the first image includes the pixel difference of a plurality of fourth pixel points in the fourth pixel area and the target feature point In the second pixel area of the target feature point, the number of the fourth pixel points is the same as the number of the first pixel points.
  • the computer device obtains the pixel difference of the plurality of third pixels according to the pixel value of each third pixel. In some embodiments, the computer device obtains an average value of the pixel values of the plurality of third pixels according to the pixel values of the plurality of third pixels, and according to the average value, the computer device obtains the plurality of third pixels. The pixel variance of the pixel point is determined as the pixel difference of the plurality of third pixel points.
  • the computer device obtains the horizontal gradient of the pixel value and the vertical gradient of the pixel value of the plurality of third pixels according to the pixel values of the plurality of third pixels, and the computer device obtains the horizontal gradient of the pixel value and the vertical gradient of the pixel value according to the pixel values of the plurality of third pixels.
  • the horizontal gradient of the pixel value of the three pixels and the vertical gradient of the pixel value are obtained, the minimum characteristic value of the gradient matrix of the plurality of third pixels is obtained, and the minimum characteristic value is determined as the pixel difference of the plurality of third pixels.
  • the target difference value and the target threshold can be set based on needs, and the embodiment of the present application does not specifically limit this.
  • the target threshold may be 13 ⁇ 13, that is, the maximum pixel area of the target feature point may be An area including 13 ⁇ 13 pixels.
  • the computer device obtains a second pixel area having the same size as the first pixel area according to the target feature point.
  • the computer device may pre-store a target movement track, and the target movement track is used to indicate the movement process of the first pixel area. Then this step may include: the computer device moves the first pixel area to the second pixel area including the target feature point according to the target feature point and the target movement track.
  • the target movement trajectory can be one unit to the right. After moving the 3 ⁇ 3 square centered on the target feature point, in the second pixel area obtained, the target feature point is the middle left of the square. Point location.
  • the computer device may move the first pixel area multiple times to obtain a second pixel area with a pixel difference greater than the target difference value. That is, the computer device may obtain the third pixel area in the second way, and obtain the second pixel area based on the pixel difference of the plurality of third pixel points in the third pixel area.
  • This process is the same as the process of the first method above, and will not be detailed here. The difference is that in the second method, the first pixel area is not expanded. Therefore, the computer device does not need to increase the first sampling step to obtain the third pixel point.
  • the computer device can directly follow the first pixel area.
  • the first sampling step is to obtain the same number of third pixels as the plurality of first pixels from the third pixel area.
  • the number of the first pixel or the second pixel obtained by the computer device is the same. Therefore, the number of pixels actually involved in the calculation is guaranteed when the pixel difference is increased. No change, and based on the difference between the pixel points in the pixel area and the target difference value, the continuous tentative increase or movement of the first pixel area will result in a larger difference between the pixels in the second pixel area to ensure The pixel characteristics in the special pixel area change significantly, which avoids the problem of inaccurate motion information determination due to the insignificant change in pixel brightness in some smooth or poorly textured pixel areas. Moreover, as the pixel area is continuously enlarged or moved, the pixel difference of the pixel area becomes larger.
  • the area with larger pixel difference is used to represent the pixel change characteristics of the area around the target feature point, which ensures that the pixels between the pixels involved in the calculation
  • the large difference ensures that the number of pixels involved in the calculation is fixed and does not increase the amount of calculation, thereby balancing the number of pixels involved in the calculation and the pixel difference of the pixels, making the calculation complexity of tracking the target feature points In the case of unchanged, the pixel richness is increased, which improves the robustness of target feature point tracking, so that it can be executed stably in a smooth or lacking texture environment, and the wide range of applicable scenes is improved.
  • the computer device acquires the movement information of the target feature point according to the multiple second pixel points and the second image.
  • the motion information is used to indicate the position change of the target feature point in the first image and the second image.
  • the computer device can determine the starting pixel point at the corresponding position in the second image according to the target feature point at the first position of the first image, and obtain multiple fourth pixels in the pixel area of the starting pixel point, according to The plurality of fourth pixel points and the plurality of second pixel points determine the motion information of the target feature point based on a Gauss-Newton algorithm.
  • the computer device can also obtain the position of the target feature point in the second image.
  • the number of the fourth pixel points is the same as the number of the second pixel points.
  • the motion information may include the moving distance of the target feature point on the x-axis and the y-axis of the image coordinate axis respectively.
  • the computer device determines the movement information of the target feature point from the first image to the second image based on the following formula 4 based on the pixel values of the target number of second pixels and the pixel values of the plurality of fourth pixels;
  • T t i multiple second pixel points in the second pixel area of the target feature point i
  • d i t+1 represents the motion information of the target feature point i
  • d i t+1 (u x , u y ) respectively represent the two-dimensional movement distance on the x-axis and y-axis.
  • I t represents the pixel value of multiple second pixels in the first image
  • I t+1 represents the pixel value of multiple fourth pixels in the second image, assuming multiple second pixels
  • the pixel value of is the same in the first image and the second image, and the computer equipment can minimize the pixel difference between the pixel values of multiple second pixels in the first image and multiple fourth pixels in the second image to solve d i t+1 .
  • the motion information may be represented by a homography matrix, and the computer device determines according to the pixel value of the target number of second pixels and the pixel values of the plurality of fourth pixels based on the following formula 5 Motion information of the target feature point from the first image to the second image;
  • T t i represents multiple second pixel points in the second pixel area of the target feature point i
  • H i t+1 represents the motion information of the target feature point i
  • the computer device may express the motion information as: Among them, h 11 and h 22 respectively represent the scaling factor of the second pixel area from the first image to the second image in the x-axis direction and the y-axis direction of the image coordinate system.
  • the x-axis direction and the y-axis direction can be respectively the image In the horizontal and vertical directions.
  • h 11 and h 221 also indicate the process of rotating along the x axis and the normal vector of the second pixel area together with h 12 and h 21 .
  • h 12 and h 21 respectively represent the projection of the second pixel area from the first image to the second image in the x-axis direction and y-axis direction of the image coordinate system.
  • h 13 and h 23 represent the second pixel area from the first image.
  • the moving distance between the image and the second image in the x-axis direction and the y-axis direction of the image coordinate system, h 31 and h 32 respectively represent the shear parameters of the second pixel area in the x-axis direction and y-axis direction in the image coordinate system,
  • the shear parameter may be a deformation ratio of the second pixel region in the x-axis direction and the y-axis direction.
  • the first pixel area of the feature point in the first image can be a square
  • the second pixel area corresponding to the feature point in the second image can be a trapezoid. If the upper and lower sides of the square and the trapezoid are in the x-axis direction, then This h 31 represents the rate of change of the side lengths of the upper and lower sides of the trapezoid, and h 32 represents the rate of change of the side lengths of the left and right sides of the trapezoid.
  • I t represents the pixel value of multiple second pixels in the first image
  • I t+1 represents the pixel value of multiple fourth pixels in the second image, assuming multiple second pixels
  • the pixel value of is the same in the first image and the second image, and the computer device can minimize the pixel difference between the pixel values of multiple second pixels in the first image and multiple fourth pixels in the second image to solve the problem H i t+1 .
  • the target feature points in the embodiment of the present application are sparsely distributed in the image, and therefore may also become sparse target feature points.
  • the first image and the second image may be two consecutive images in the video shot by the camera.
  • the video stream captured by the camera in real time provides the camera's observation of the external environment at different moments.
  • the motion information of target feature points is often used in processes such as motion detection, motion estimation, real-time positioning, three-dimensional reconstruction, and object segmentation. For example, as shown in Figure 5, for the target feature point tracking process.
  • the movement trajectory of the target feature point that is, from the image position detected for the first time to the current position of the target feature point on the image, is represented by a white line after each target feature point.
  • the computer device can obtain the first pixel area with the size of l ⁇ l, based on the pixels in the first pixel area.
  • the pixel difference is less than the target difference value
  • the step size is increased to filter out multiple third pixels in the third pixel area.
  • the third pixel area is no longer increased, and the third pixel area is used as the second pixel area. If the pixel difference of the pixels in the third pixel area is less than the target difference value, and the third pixel If the size of the area is not greater than l m ⁇ l m , continue to increase the third pixel area until the second pixel area whose pixel difference is not less than the target difference value or greater than l m ⁇ l m is obtained.
  • the second pixel area whose pixel difference is larger than the first pixel area and the number of pixel points is not changed is obtained, so that multiple second pixel areas based on the number of original pixels can be obtained.
  • Two pixels are calculated to obtain the motion information of the target feature point. Because the pixel difference of the data involved in the calculation is increased under the premise of keeping the number of pixels unchanged, the calculation complexity and information richness are balanced to ensure the target feature point motion On the premise of the accuracy of the information, the efficiency of determining the movement information of the target feature point is improved.
  • FIG. 7 is a schematic flowchart of a task execution method provided by an embodiment of the present application. The method is applied on the target device, see Figure 7. The method includes:
  • the target device acquires a first image and a second image.
  • the first image and the second image include the same object;
  • the target device is a device that collects the first image and the second image.
  • the target device may take a first image and a second image that include the same object during the movement.
  • the target device may be a computer device, and the computer device may be used to determine the target feature point in the first image.
  • Motion information which is used to indicate the position change of the target feature point in the first image and the second image.
  • the computer device may be a mobile phone, and the computer device may move while capturing images in real time during the movement.
  • the target device may not be the computer device.
  • the target device captures the first image and the second image while moving, and sends the captured first image and second image to the computer in real time. equipment.
  • the target device obtains motion information of a target feature point on the object in the first image.
  • the motion information is used to indicate the position change of the target feature point in the first image and the second image.
  • the target device may obtain the movement information of the target feature point based on the process of steps S201-S205 in the foregoing invention embodiment.
  • the computer device can obtain the motion information of the target feature point based on the process of steps S201-S205 in the above-mentioned embodiment of the invention, and send the motion information of the target feature point to the target device in real time.
  • the target device receives the movement information of the target feature point sent by the computer device.
  • the target device executes the target task based on the movement information of the target feature point.
  • the target task can be a route planning task or an object recognition task.
  • the target task may be a route planning task, and the target device may construct a scene object surrounding the target device based on the movement information of the target feature point, and plan the route based on the scene object.
  • the process may include: when the number of the target feature points is multiple, the target device may determine at least one scene object whose distance from the target device does not exceed a first threshold based on the motion information of the multiple target feature points; and the target The device may determine a first target route for the target device to reach the destination according to the location of the destination that is not more than a second threshold from the target device and the at least one scene object, where the second threshold is greater than the first threshold.
  • the computer device can obtain the current location through positioning, and based on the current location, guide the user to the nearest restaurant, store, bathroom, etc.; for example, if the user is on a certain floor of a large shopping mall, the target device can be based on the surrounding The name of the store accurately locates the location of the target device on this floor. For example, in front of clothing store A, the target device can guide the user to watch store B.
  • the target task may be a route planning task.
  • the target device may also perform some operations on scene objects around the target device.
  • the target device may be a robot, and the target device may place a water cup on a table. on. Then, after the target device determines at least one scene object that is not more than the first threshold from the target device, the target device can determine the second target route of the target device according to the at least one scene object and the target task to be executed.
  • the task refers to performing a target operation on the target scene object in the at least one scene object.
  • the VR device and the AR device can combine virtual objects based on the scene objects in the scene. Placed in the real environment.
  • the target task includes an object recognition task.
  • the target device can identify moving objects and stationary objects in the surrounding environment based on the first image and the second image, and can also identify the object category of the moving object, for example Whether the moving object is a person or a vehicle, the process may include: when the number of target feature points is multiple, the target device may determine that the motion information of the multiple target feature points conforms to the motion information of the multiple target feature points. Multiple first feature points of the target condition, the multiple first feature points are used to indicate a moving object among multiple objects included in the first image and the second image; the target device may be based on the multiple first features Point on the position of the first image or the second image to determine the object category to which the moving object belongs.
  • the object category may include: vehicles, people, or animals.
  • the target condition may include: the motion information of the multiple feature points is different from the motion information of other feature points, or is different from the motion information of the feature points that exceed the target number among the multiple feature points. For example, it is different from the motion information of 80% of the 200 feature points.
  • the target device may further determine a third target route based on people walking around, a moving vehicle, or a stationary house.
  • the three-target route may be a route to avoid the moving object. In order to avoid obstacles in the process of moving, more accurate route planning.
  • the computer equipment can determine whether the certain object is moving, and if there is movement, in which direction to move; after detecting the moving object, it can also move the object Divide it out and judge whether it is a human or other animal, because the animal is often in the front and the scene is often in the back.
  • the computer equipment can determine whether the certain object is moving, and if there is movement, in which direction to move; after detecting the moving object, it can also move the object Divide it out and judge whether it is a human or other animal, because the animal is often in the front and the scene is often in the back.
  • the computer device can determine whether the target device is moving based on the position changes of all target feature points in the entire image. If the target device moves, it can also determine the moving direction and moving track of the target device, and further calculate the surroundings.
  • Scene object For example, surrounding seats, obstacles, etc., based on surrounding scene objects, reconstruct a scene of the target device, and the scene includes multiple scene objects located around the target device. For example, taking the target device as a robot as an example, when the robot performs a task, the robot can avoid obstacles due to the reconstruction of the scene; another example, the robot can further plan the next path based on the current position to reach its desired Where it reaches, perform the corresponding task.
  • the target device may further combine the scene object with a virtual object according to the at least one scene object, as The user shows the virtual and real environment.
  • the process may include: the target device obtains at least one virtual scene object, the target device constructs a virtual scene according to the position of the at least one virtual scene object and the position of the at least one scene object, and displays the virtual scene on the screen of the target device .
  • the real environment and the virtual environment can be rendered in a corresponding position relationship and displayed on the screen of the computer device.
  • the target device can use the motion information of the target feature point on the object in the first image to perform target tasks, such as planning a route, motion detection, and identifying the object category to which the object belongs, and the target feature point is
  • target tasks such as planning a route, motion detection, and identifying the object category to which the object belongs
  • target feature point is The determination process of motion information is accurate and efficient, which improves the accuracy and efficiency of the target task.
  • FIG. 8 is a schematic structural diagram of an apparatus for determining motion information of image feature points provided by an embodiment of the present application.
  • the device includes:
  • the determining module 801 is configured to determine a first image and a second image, where the first image and the second image include the same object.
  • the determining module 801 is further configured to determine a first pixel area in the first image that includes the target feature point based on the target feature point on the object in the first image.
  • the determining module 801 is further configured to determine a second pixel area including the target feature point in the first image according to the pixel difference of the plurality of first pixel points in the first pixel area and the target feature point.
  • the pixel difference of the plurality of second pixels in the pixel area is greater than the pixel difference of the plurality of first pixels, the number of the plurality of second pixels is the same as the number of the plurality of first pixels, and the pixel difference is used to indicate more The degree of change in the pixel value of each pixel.
  • the acquiring module 802 is configured to acquire motion information of the target feature point according to the plurality of second pixel points and the second image, and the motion information is used to indicate that the target feature point is in the first image and the second image The location changes.
  • the determining module 801 is configured to obtain, according to the target feature point, when the pixel difference of the first pixel points in the first pixel area is smaller than the target difference value The second pixel area of the target feature point.
  • the determining module 801 is configured to, when the pixel difference of a plurality of first pixel points in the first pixel area is less than the target difference value, obtain the size of the first pixel area according to the target feature point. For the same second pixel area, the first pixel area and the second pixel area include different pixel points.
  • the determining module 801 is further configured to expand the first pixel area into a second pixel area including the target feature point according to the target expansion coefficient and centering on the target feature point.
  • the determining module 801 is further configured to move the first pixel area to a second pixel area including the target feature point according to the target feature point and according to the target movement track.
  • the determining module 801 is further configured to obtain a third pixel region including the target feature point according to the target feature point when the pixel difference of the plurality of first pixel points is less than the target difference value, and the The number of the plurality of third pixels in the third pixel area is the same as the number of the plurality of first pixels; according to the pixel values of the plurality of third pixels in the third pixel area, the pixels of the plurality of third pixels are determined Difference; and when the pixel difference of the plurality of third pixel points is not less than the target difference value, the third pixel area is determined as the second pixel area.
  • the determining module 801 is further configured to increase the first sampling step of the first pixel area to the second sampling step according to the expansion coefficient from the first pixel area to the third pixel area. Long, according to the second sampling step size, third pixel points with the same number as the plurality of first pixels are obtained from the third pixel area.
  • the determining module 801 is further configured to obtain, from the third pixel area, the same number of third pixels as the plurality of first pixels according to the first sampling step of the first pixel area .
  • the device further includes:
  • the detection module is configured to detect whether the size of the third pixel area is greater than the target threshold when the pixel difference of the plurality of third pixel points is less than the target difference value.
  • the determining module 801 is further configured to determine a fourth pixel area larger than the third pixel area when the size of the third pixel area is not greater than the target threshold.
  • the determining module 801 is further configured to determine a second pixel area including the target feature point in the first image based on the pixel difference of a plurality of fourth pixel points in the fourth pixel area and the target feature point.
  • the number of the fourth pixel is the same as the number of the first pixel.
  • the pixel difference is the pixel variance of the multiple pixels or the smallest characteristic value of the gradient matrix of the multiple pixels
  • the pixel variance is used to indicate that the pixel values of the multiple pixels are relative to the pixel average
  • the degree of change of the value, the gradient matrix is used to indicate the degree of change of the plurality of pixels on the horizontal gradient relative to the pixel average value and the degree of change on the vertical gradient relative to the pixel average value.
  • the shape of the first pixel area and the second pixel area is any shape of a square, a rectangle, a circle, a ring, an irregular polygon, or an irregular curve.
  • the determining module 801 is further configured to obtain the plurality of first pixel points among the plurality of pixels on the area boundary of the first pixel area according to the first sampling step of the first pixel area. Pixel points; and determining the pixel difference of the plurality of first pixel points according to the pixel values of the plurality of first pixel points.
  • the second pixel area whose pixel difference is larger than the first pixel area and the number of pixels is unchanged is obtained, so that calculation can be performed based on multiple second pixel points of the original pixel number,
  • the pixel difference of the data involved in the calculation is increased while keeping the number of pixels unchanged, which balances the computational complexity and information richness, and guarantees the premise of the accuracy of the target feature point motion information This improves the efficiency of determining the movement information of the target feature point.
  • Fig. 9 is a schematic structural diagram of a task execution device provided by an embodiment of the present application. Referring to Figure 9, the device includes:
  • the obtaining module 901 is configured to obtain a first image and a second image of a target device, where the first image and the second image include the same object.
  • the acquiring module 901 is also configured to acquire motion information of a target feature point on the object in the first image, and the motion information is used to indicate a position change of the target feature point in the first image and the second image.
  • the task processing module 902 is configured to execute the target task based on the motion information of the target feature point.
  • the target task includes a route planning task
  • the task processing module 902 is further configured to determine the distance to the target device based on the movement information of multiple target feature points when the number of target feature points is multiple. At least one scene object that does not exceed the first threshold; and the first target route for the target device to reach the destination is determined based on the location of the target device that does not exceed the second threshold and the at least one scene object.
  • the second threshold is greater than the first threshold.
  • the target task includes an object recognition task
  • the task processing module 902 is further configured to determine the multiple target feature points based on the motion information of the multiple target feature points when the number of target feature points is multiple.
  • a plurality of first feature points in the target feature point whose motion information meets the target condition, and the plurality of first feature points are used to indicate a moving object among a plurality of objects included in the first image and the second image; and based on the multiple The position of the first feature point in the first image or the second image determines the object category to which the moving object belongs.
  • the target device can use the motion information of the target feature point on the object in the first image to perform the target task, such as planning a route, motion detection, or identifying the object category to which the object belongs, and the target feature point
  • the target task such as planning a route, motion detection, or identifying the object category to which the object belongs
  • the determination process of the movement information is accurate and efficient, which improves the accuracy and efficiency of the target task.
  • the device for determining the motion information of image feature points provided in the above embodiment determines the motion information of the feature points, as well as the task execution device and the task execution, only the division of the above functional modules is used as an example.
  • the above-mentioned function allocation can be completed by different function modules according to needs, that is, the internal structure of the computer device is divided into different function modules to complete all or part of the functions described above.
  • the device for determining motion information of image feature points and the method for determining motion information of image feature points, as well as the embodiments of the task execution device and the task execution method provided by the foregoing embodiments belong to the same concept.
  • the specific implementation process please refer to the method embodiment. No longer.
  • a computer device including a memory and a processor, and computer-readable instructions are stored in the memory.
  • the processor executes the computer-readable instructions to implement the method for determining motion information of image feature points. Steps in.
  • a target device including a memory and a processor, where computer-readable instructions are stored in the memory, and the processor implements the steps in the above-mentioned task execution method embodiments when executing the computer-readable instructions.
  • a computer-readable storage medium which stores computer-readable instructions that, when executed by a processor, implement the steps in the above-mentioned method for determining motion information of image feature points, or , To realize the steps in the above-mentioned task execution method embodiment.
  • FIG. 10 is a schematic structural diagram of a terminal provided by an embodiment of the present application.
  • the terminal 1000 can be: a smart phone, a tablet computer, an MP3 player (Moving Picture Experts Group Audio Layer III, moving picture expert compression standard audio layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, moving picture expert compressing standard audio Level 4) Player, laptop or desktop computer.
  • the terminal 1000 may also be called user equipment, portable terminal, laptop terminal, desktop terminal and other names.
  • the terminal 1000 includes a processor 1001 and a memory 1002.
  • the processor 1001 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so on.
  • the processor 1001 may adopt at least one hardware form among DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array, Programmable Logic Array). achieve.
  • the processor 1001 may also include a main processor and a coprocessor.
  • the main processor is a processor used to process data in the wake state, also called a CPU (Central Processing Unit, central processing unit); the coprocessor is A low-power processor used to process data in the standby state.
  • the processor 1001 may be integrated with a GPU (Graphics Processing Unit, image processor), and the GPU is used for rendering and drawing content that needs to be displayed on the display screen.
  • the processor 1001 may further include an AI (Artificial Intelligence, artificial intelligence) processor, which is used to process computing operations related to machine learning.
  • AI Artificial Intelligence, artificial intelligence
  • the memory 1002 may include one or more computer-readable storage media, which may be non-transitory.
  • the memory 1002 may also include high-speed random access memory and non-volatile memory, such as one or more magnetic disk storage devices and flash memory storage devices.
  • the non-transitory computer-readable storage medium in the memory 1002 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 1001 to implement the image features provided in the method embodiments of this application. Point’s motion information determination method or task execution method.
  • the terminal 1000 optionally further includes: a peripheral device interface 1003 and at least one peripheral device.
  • the processor 1001, the memory 1002, and the peripheral device interface 1003 may be connected by a bus or a signal line.
  • Each peripheral device can be connected to the peripheral device interface 1003 through a bus, a signal line or a circuit board.
  • the peripheral device includes: at least one of a radio frequency circuit 1004, a touch display screen 1005, a camera 1006, an audio circuit 1007, a positioning component 1008, and a power supply 1009.
  • the peripheral device interface 1003 may be used to connect at least one peripheral device related to I/O (Input/Output) to the processor 1001 and the memory 1002.
  • the processor 1001, the memory 1002, and the peripheral device interface 1003 are integrated on the same chip or circuit board; in some other embodiments, any one of the processor 1001, the memory 1002, and the peripheral device interface 1003 or The two can be implemented on separate chips or circuit boards, which are not limited in this embodiment.
  • the radio frequency circuit 1004 is used for receiving and transmitting RF (Radio Frequency, radio frequency) signals, also called electromagnetic signals.
  • the radio frequency circuit 1004 communicates with a communication network and other communication devices through electromagnetic signals.
  • the radio frequency circuit 1004 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals into electrical signals.
  • the radio frequency circuit 1004 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, and so on.
  • the radio frequency circuit 1004 can communicate with other terminals through at least one wireless communication protocol.
  • the wireless communication protocol includes but is not limited to: metropolitan area network, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area network and/or WiFi (Wireless Fidelity, wireless fidelity) network.
  • the radio frequency circuit 1004 may also include NFC (Near Field Communication) related circuits, which is not limited in this application.
  • the display screen 1005 is used to display a UI (User Interface, user interface).
  • the UI can include graphics, text, icons, videos, and any combination thereof.
  • the display screen 1005 also has the ability to filter out touch signals on or above the surface of the display screen 1005.
  • the touch signal may be input to the processor 1001 as a control signal for processing.
  • the display screen 1005 may also be used to provide virtual buttons and/or virtual keyboards, also called soft buttons and/or soft keyboards.
  • the display screen 1005 there may be one display screen 1005, which is provided with the front panel of the terminal 1000; in other embodiments, there may be at least two display screens 1005, which are respectively arranged on different surfaces of the terminal 1000 or in a folded design; In still other embodiments, the display screen 1005 may be a flexible display screen, which is arranged on a curved surface or a folding surface of the terminal 1000. Furthermore, the display screen 1005 can also be set as a non-rectangular irregular pattern, that is, a special-shaped screen.
  • the display screen 1005 may be made of materials such as LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode, organic light emitting diode).
  • the camera assembly 1006 is used to filter out images or videos.
  • the camera assembly 1006 includes a front camera and a rear camera.
  • the front camera is set on the front panel of the terminal, and the rear camera is set on the back of the terminal.
  • the camera assembly 1006 may also include a flash.
  • the flash can be a single-color flash or a dual-color flash. Dual color temperature flash refers to a combination of warm light flash and cold light flash, which can be used for light compensation under different color temperatures.
  • the audio circuit 1007 may include a microphone and a speaker.
  • the microphone is used to filter out the sound waves of the user and the environment, and convert the sound waves into electrical signals and input to the processor 1001 for processing, or input to the radio frequency circuit 1004 to implement voice communication.
  • the microphone can also be an array microphone or an omnidirectional screening microphone.
  • the speaker is used to convert the electrical signal from the processor 1001 or the radio frequency circuit 1004 into sound waves.
  • the speaker can be a traditional membrane speaker or a piezoelectric ceramic speaker.
  • the speaker When the speaker is a piezoelectric ceramic speaker, it can not only convert the electrical signal into human audible sound waves, but also convert the electrical signal into human inaudible sound waves for purposes such as distance measurement.
  • the audio circuit 1007 may also include a headphone jack.
  • the positioning component 1008 is used to locate the current geographic position of the terminal 1000 to implement navigation or LBS (Location Based Service, location-based service).
  • the positioning component 1008 may be a positioning component based on the GPS (Global Positioning System, Global Positioning System) of the United States, the Beidou system of China, the Granus system of Russia, or the Galileo system of the European Union.
  • the power supply 1009 is used to supply power to various components in the terminal 1000.
  • the power source 1009 may be alternating current, direct current, disposable batteries or rechargeable batteries.
  • the rechargeable battery may support wired charging or wireless charging.
  • the rechargeable battery can also be used to support fast charging technology.
  • the terminal 1000 further includes one or more sensors 1010.
  • the one or more sensors 1010 include, but are not limited to: an acceleration sensor 1011, a gyroscope sensor 1012, a pressure sensor 1013, a fingerprint sensor 1014, an optical sensor 1015, and a proximity sensor 1016.
  • the acceleration sensor 1011 can detect the magnitude of acceleration on the three coordinate axes of the coordinate system established by the terminal 1000.
  • the acceleration sensor 1011 can be used to detect the components of the gravitational acceleration on three coordinate axes.
  • the processor 1001 may control the touch screen 1005 to display the user interface in a horizontal view or a vertical view according to the gravity acceleration signal filtered by the acceleration sensor 1011.
  • the acceleration sensor 1011 can also be used to filter out the game or user's motion data.
  • the gyroscope sensor 1012 can detect the body direction and rotation angle of the terminal 1000, and the gyroscope sensor 1012 can cooperate with the acceleration sensor 1011 to screen out the user's 3D actions on the terminal 1000.
  • the processor 1001 can implement the following functions according to the data filtered by the gyroscope sensor 1012: motion sensing (such as changing the UI according to the user's tilt operation), image stabilization during shooting, game control, and inertial navigation.
  • the pressure sensor 1013 may be disposed on the side frame of the terminal 1000 and/or the lower layer of the touch screen 1005.
  • the processor 1001 performs left and right hand recognition or quick operation according to the holding signal filtered by the pressure sensor 1013.
  • the processor 1001 controls the operability controls on the UI interface according to the user's pressure operation on the touch display screen 1005.
  • the operability control includes at least one of a button control, a scroll bar control, an icon control, and a menu control.
  • the fingerprint sensor 1014 is used to screen out the user's fingerprint.
  • the processor 1001 identifies the user's identity according to the fingerprint screened out by the fingerprint sensor 1014, or the fingerprint sensor 1014 identifies the user's identity according to the screened fingerprint.
  • the processor 1001 authorizes the user to perform related sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, paying, and changing settings.
  • the fingerprint sensor 1014 may be provided on the front, back or side of the terminal 1000. When a physical button or a manufacturer logo is provided on the terminal 1000, the fingerprint sensor 1014 can be integrated with the physical button or the manufacturer logo.
  • the optical sensor 1015 is used to filter out the ambient light intensity.
  • the processor 1001 may control the display brightness of the touch screen 1005 according to the intensity of the ambient light filtered by the optical sensor 1015. Specifically, when the ambient light intensity is high, the display brightness of the touch display screen 1005 is increased; when the ambient light intensity is low, the display brightness of the touch display screen 1005 is decreased.
  • the processor 1001 may also dynamically adjust the shooting parameters of the camera assembly 1006 according to the ambient light intensity filtered by the optical sensor 1015.
  • the proximity sensor 1016 also called a distance sensor, is usually arranged on the front panel of the terminal 1000.
  • the proximity sensor 1016 is used to filter the distance between the user and the front of the terminal 1000.
  • the processor 1001 controls the touch screen 1005 to switch from the on-screen state to the off-screen state; when the proximity sensor 1016 detects When the distance between the user and the front of the terminal 1000 gradually increases, the processor 1001 controls the touch display screen 1005 to switch from the rest screen state to the bright screen state.
  • FIG. 10 does not constitute a limitation on the terminal 1000, and may include more or fewer components than shown in the figure, or combine certain components, or adopt different component arrangements.
  • FIG. 11 is a schematic structural diagram of a server provided by an embodiment of the present application.
  • the server 1100 may have relatively large differences due to different configurations or performance, and may include one or more processors (central processing units, CPU) 1101 and one Or more than one memory 1102, wherein the memory 1102 stores at least one computer-readable instruction, and the at least one computer-readable instruction is loaded and executed by the processor 1101 to implement the image feature points provided by the foregoing method embodiments. Movement information determination method or task execution method.
  • the server may also have components such as a wired or wireless network interface, a keyboard, an input and output interface for input and output, and the server may also include other components for implementing device functions, which are not described here.
  • a computer-readable storage medium such as a memory including instructions, which can be executed by a processor in a terminal to complete the method or task of determining motion information of image feature points in the foregoing embodiment.
  • the computer-readable storage medium may be ROM (Read-Only Memory), RAM (random access memory), CD-ROM (Compact Disc Read-Only Memory, CD-ROM), Tapes, floppy disks and optical data storage devices, etc.
  • the program can be stored in a computer-readable storage medium.
  • the storage medium can be read-only memory, magnetic disk or optical disk, etc.

Abstract

一种图像特征点的运动信息确定方法，包括：确定第一图像和第二图像，该第一图像和该第二图像包括相同的对象；基于该第一图像中该对象上的目标特征点，确定该第一图像中包括该目标特征点的第一像素区域；根据该第一像素区域中多个第一像素点的像素差异以及该目标特征点，确定该第一图像中包括该目标特征点的第二像素区域；该第二像素区域中多个第二像素点的像素差异大于该多个第一像素点的像素差异，该多个第二像素点与该多个第一像素点的数量相同，该像素差异用于指示多个像素点的像素值的变化程度；根据该多个第二像素点和该第二图像，获取该目标特征点的运动信息；该运动信息用于指示该目标特征点在该第一图像和该第二图像中的位置变化。

Description

图像特征点的运动信息确定方法、任务执行方法和设备
本申请要求于2019年04月29日提交中国专利局，申请号为201910356752.4、发明名称为“图像特征点的运动信息确定方法、任务执行方法和设备”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
本申请涉及互联网技术领域，特别涉及一种图像特征点的运动信息确定方法、任务执行方法和设备。
背景技术
随着互联网技术的发展，特征点跟踪技术在移动机器人、或虚拟现实等多种场景中广泛应用。特征点跟踪技术是分析同一特征点在连续多帧图像中位置变化情况的过程。本领域中，通常采用运动信息来描述特征点在不同图像的位置变化过程。
在传统方案中，以确定特征点从视频中当前帧图像到后一帧图像的运动信息为例，图像特征点的运动信息确定过程可以包括：计算机设备获取当前帧图像中特征点所在的像素区域，为了使得结果准确，该像素区域通常会选取的较大。该计算机设备基于多个像素区域内多个像素点和后一帧图像中多个对应像素区域内多个像素点，来确定图像特征点的运动信息。然而，传统方案中，由于所选取的像素区域较大，需要计算的像素点数量较大，导致图像特征点的运动信息确定过程的效率较低。
发明内容
根据本申请的各种实施例提供了一种图像特征点的运动信息确定方法、装置、计算机设备及存储介质，还提供了一种任务执行方法、装置、目标设备及存储介质。
一种图像特征点的运动信息确定方法，由计算机设备执行，所述方法包 括：
确定第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
基于所述第一图像中所述对象上的目标特征点，确定所述第一图像中包括所述目标特征点的第一像素区域；
根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述第二像素区域中多个第二像素点的像素差异大于所述多个第一像素点的像素差异，所述多个第二像素点与所述多个第一像素点的数量相同，所述像素差异用于指示多个像素点的像素值的变化程度；及
根据所述多个第二像素点和所述第二图像，获取所述目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化。
一种任务执行方法，由目标设备执行，所述方法包括：
获取第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
获取所述第一图像中所述对象上的目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化；及
基于所述目标特征点的运动信息，执行目标任务。
一种图像特征点的运动信息确定装置，所述装置包括：
确定模块，用于确定第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
所述确定模块，还用于基于所述第一图像中所述对象上的目标特征点，确定所述第一图像中包括所述目标特征点的第一像素区域；
所述确定模块，还用于根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述第二像素区域中多个第二像素点的像素差异大于所述多个第 一像素点的像素差异，所述多个第二像素点与所述多个第一像素点的数量相同，所述像素差异用于指示多个像素点的像素值的变化程度；及
获取模块，用于根据所述多个第二像素点和所述第二图像，获取所述目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化。
在一些实施例中，所述确定模块，还用于当所述第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取大于所述第一像素区域且包括所述目标特征点的第二像素区域。
在一些实施例中，所述确定模块，还用于当所述第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取与所述第一像素区域的大小相同的第二像素区域，所述第一像素区域和所述第二像素区域所包括的像素点不同。
在一些实施例中，所述确定模块，还用于按照目标扩大系数，以所述目标特征点为中心，将所述第一像素区域扩大为包括所述目标特征点的第二像素区域。
在一些实施例中，所述确定模块，还用于根据所述目标特征点，按照目标移动轨迹，将所述第一像素区域移动至包括所述目标特征点的第二像素区域。
在一些实施例中，所述确定模块，还用于当所述多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取包括所述目标特征点的第三像素区域，所述第三像素区域中多个第三像素点与所述多个第一像素点的数量相同；根据所述第三像素区域中多个第三像素点的像素值，确定所述多个第三像素点的像素差异；及当所述多个第三像素点的像素差异不小于目标差异值时，将所述第三像素区域确定为所述第二像素区域。
在一些实施例中，所述确定模块，还用于根据从所述第一像素区域到所述第三像素区域的扩大系数，将所述第一像素区域的第一采样步长增大为第二采样步长，按照所述第二采样步长，从所述第三像素区域中获取与所述多个第一像素点数量相同的第三像素点。
在一些实施例中，所述确定模块，还用于按照所述第一像素区域的第一采样步长，从所述第三像素区域中获取与所述多个第一像素点数量相同的第 三像素点。
在一些实施例中，所述装置还包括：
检测模块，用于当所述多个第三像素点的像素差异小于所述目标差异值时，检测所述第三像素区域的大小是否大于目标阈值；
所述确定模块，还用于当所述第三像素区域的大小不大于所述目标阈值时，确定大于所述第三像素区域的第四像素区域；及
所述确定模块，还用于基于所述第四像素区域中多个第四像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述多个第四像素点与所述多个第一像素点的数量相同。
在一些实施例中，所述像素差异为所述多个像素点的像素方差或者所述多个像素点的梯度矩阵的最小特征值，所述像素方差用于表示所述多个像素点的像素值相对于像素平均值的变化程度，所述梯度矩阵用于表示所述多个像素点的像素值分别在水平梯度上相对于像素平均值的变化程度和垂直梯度上相对于像素平均值的变化程度。
在一些实施例中，所述第一像素区域和所述第二像素区域的形状为正方形、长方形、圆形、环形、不规则多边形或不规则曲边形中的任一形状。
在一些实施例中，所述确定模块，还用于根据所述第一像素区域的第一采样步长，在所述第一像素区域的区域边界上的多个像素点中，获取所述多个第一像素点；及根据所述多个第一像素点的像素值，确定所述多个第一像素点的像素差异。
一种任务执行装置，所述装置包括：
获取模块，用于获取第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
所述获取模块，还用于获取所述第一图像中所述对象上的目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化；及
任务处理模块，用于基于所述目标特征点的运动信息，执行目标任务。
在一些实施例中，所述目标任务包括路线规划任务，所述任务处理模块，还用于当所述目标特征点的数目为多个时，基于多个目标特征点的运动信息， 确定距离目标设备不超过第一阈值的至少一个场景对象，所述目标设备为采集所述第一图像和第二图像的设备；及根据距离所述目标设备不超过第二阈值的目的地的位置和所述至少一个场景对象，确定所述目标设备到达所述目的地的第一目标路线，所述第二阈值大于所述第一阈值。
在一些实施例中，所述目标任务包括对象识别任务，所述任务处理模块，还用于当所述目标特征点的数目为多个时，基于多个目标特征点的运动信息，确定出所述多个目标特征点中运动信息符合目标条件的多个第一特征点，所述多个第一特征点用于指示所述第一图像和所述第二图像包括的多个对象中的运动对象；及基于所述多个第一特征点在所述第一图像或第二图像的位置，确定所述运动对象所属的对象类别。
一种计算机设备，所述计算机设备包括处理器和存储器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行所述图像特征点的运动信息确定方法的步骤。
一种目标设备，所述目标设备包括处理器和存储器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行所述任务执行方法的步骤。
一种非易失性的计算机可读存储介质，存储有计算机可读指令，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行所述图像特征点的运动信息确定方法的步骤。
一种非易失性的计算机可读存储介质，存储有计算机可读指令，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行所述任务执行方法的步骤。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征、目的和优点将从说明书、附图以及权利要求书变得明显。
附图说明
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的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的 前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例提供的一种图像特征点的运动信息确定方法的实施环境的示意图；
图2是本申请实施例提供的一种图像特征点的运动信息确定方法的流程图；
图3是本申请实施例提供的一种像素区域的示意图；
图4是本申请实施例提供的一种像素点位置示意图；
图5是本申请实施例提供的一种像素点的运动信息的示意图；
图6是本申请实施例提供的一种图像特征点的运动信息确定流程示意图；
图7是本申请实施例提供的一种任务执行方法的流程图；
图8是本申请实施例提供的一种图像特征点的运动信息确定装置的结构示意图；
图9是本申请实施例提供的一种任务执行装置的结构示意图；
图10是本申请实施例提供的一种终端的结构示意图；
图11是本申请实施例提供的一种服务器的结构示意图。
具体实施方式
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。
图1是本申请实施例提供的一种图像特征点的运动信息确定方法的实施环境的示意图，参见图1，该实施环境包括：计算机设备101，该计算机设备101可以获取到具备相同的对象的两帧或多于两帧图像，该计算机设备101可以确定该对象上的目标特征点从一帧图像到另一帧图像的运动信息，以对目标特征点在不同图像的位置变化情况进行追踪，特征点是指图像中像素特征显著的像素点。
在一种可能的场景中，该计算机设备101可以获取视频中连续的两帧或多于两帧图像，例如，第一图像和第二图像，该计算机设备101可以获取目 标特征点从第一图像到第二图像的运动信息，该运动信息用于指示该目标特征点在该第一图像和该第二图像的位置变化。例如，该计算机设备101可以基于特征点在图像的像素区域，进行位置追踪。特征点的像素区域包括目标特征点像素点。该计算机设备可以确定目标特征点在第一图像的第一像素区域，基于该第一像素区域内多个第一像素点的像素差异，在第一图像中确定出像素差异更大的第二像素区域，且第二像素区域与第一像素区域的像素点数量相同，从而可以基于第二像素区域中多个第二像素点和第二图像，确定出目标特征点从第一图像到第二图像的运动信息。
在一种可能的实施场景中，该实施环境还可以包括：计算机设备102，该计算机设备102可以采集多于一帧图像，将该多于一帧图像发送至计算机设备101。
需要说明的是，该计算机设备101可以被提供为服务器、终端等任一设备或智能机器人、无人驾驶车辆、无人机、无人船只等任一设备。该计算机设备102可以被提供为手机、摄像机、终端、监控设备等，本申请实施例对计算机设备101和计算机设备102的实际形态不做具体限定。
图2是本申请实施例提供的一种图像特征点的运动信息确定方法的流程图。该发明实施例的执行主体为计算机设备，参见图2，该方法包括：
S201、计算机设备确定第一图像和第二图像。
该第一图像和第二图像中包括相同的对象，该对象可以为任一具备实际显示形态的物体，例如，房屋、路标、或车辆等。在一种可能的场景中，该第一图像和第二图像可以为视频中包括相同对象的两帧图像，该第二图像的时间戳可以在该第一图像之后。
在一些实施例中，该计算机设备可以实时对周围环境进行拍摄，得到视频，该计算机设备可以确定该视频中第一图像和时间戳位于该第一图像之后的第二图像。在一种可能的实施场景中，该计算机设备可以为可移动的无人驾驶车辆、智能机器人、或无人机等任一设备。该计算机设备还可以是安装有摄像头的可移动设备中，则该计算机设备还可以在移动过程中录制视频。在另一种可能的实施场景中，该计算机设备可以从目标设备上获取该第一图像和第二图像。例如，该目标设备可以在移动过程中拍摄包括相同的对象的 第一图像和第二图像，该计算机设备接收目标设备发送的视频，获取该视频中的第一图像和第二图像，该目标设备可以为监控设备、手机、或摄像头等。
在一些实施例中，该计算机设备还可以确定第一图像中该对象上的目标特征点。该计算机设备可以通过目标算法，提取该第一图像中该对象上的目标特征点，该目标算法可以为SIFT(Scale-invariant feature transform，尺度不变特征转换)算法、或SURF(Speeded Up Robust Features，加速稳健特征)算法等。
需要说明的是，目标特征点是指图像中像素特征显著的点，例如，像素值与周围多个像素点的像素值之间的差值大于目标像素差。基于图像分辨率大小的不同，同一图像中采样像素点的数量也不同。采样像素点是指基于目标分辨率对图像进行采样的像素点。该目标特征点可以为第一图像中的采样像素点，也可以为位于第一图像中两个采样像素点之间的像素点。在一些实施例中，该计算机设备可以采用整数坐标表示第一图像中采样像素点的位置，整数坐标是指坐标中数值的取值均为整数。则目标特征点为第一图像中的采样像素点时，目标特征点的坐标可以为整数坐标，目标特征点为两个采样像素点之间的像素点时，目标特征点的坐标也可以为非整数坐标。
S202、计算机设备根据该第一图像中该对象上的目标特征点，确定该第一图像中包括该目标特征点的第一像素区域。
该计算机设备可以根据该目标特征点的位置，在该第一图像中确定包括该目标特征点且形状为目标形状的第一像素区域。该第一像素区域的目标形状可以为正方形、长方形、圆形、环形或者不规则多边形或曲边形中的任一形状。在一些实施例中，该第一像素区域的形状可以为中心对称形状，则该计算机设备以该目标特征点为中心，确定目标形状的第一像素区域。
在一些实施例中，该第一像素区域为至少包括有第一目标数量的像素点的区域。该计算机设备还可以根据该第一目标数量和该目标形状，以该目标特征点为中心，将目标形状的、且包括的像素点数量不小于该第一目标数量的区域作为第一像素区域。在一些实施例中，当该目标形状为正方形时，该计算机设备根据该第一目标数量，将以目标特征点为中心、边长为第一边长的正方形区域确定为该第一像素区域；该第一边长的平方不小于该第一目标数量。例如，如果该目标数量为9，则该第一边长可以为3或3以上的整数， 例如，边长为5或9等。如果该目标形状为圆形，该计算机设备根据该目标数量，将以目标特征点为中心、半径为目标半径的圆形区域确定为该第一像素区域；当该目标形状为长方形、菱形、圆环形或其他任一形状时，该计算机设备确定第一像素区域的原理相同，此处不再赘述。
该第一像素区域内的像素点反映了该目标特征点周围的像素特征，该计算机设备还可以选取该第一像素区域内部分像素点，表示该目标特征点周围的像素特征。在一些实施例中，该计算机设备还可以采用采样的方式，从该第一像素区域中确定多个第一像素点，该多个第一像素点反映了该目标特征点周围像素特征变化情况。计算机设备根据该第一像素区域内像素点总数量和第二目标数量，确定第一采样步长，根据该第一采样步长，从该第一像素区域内多个像素点中每第一采样步长的像素点，获取一个第一像素点，得到第二目标数量的第一像素点。当该第一目标数量与第二目标数量相同时，该第一采样步长为1。在一些实施例中，该目标数量的第一像素点可以以目标特征点为中心均匀分布在该第一像素区域的边界上，则该计算机设备将该第一像素区域内位于区域边界上的像素点的像素点总数除以该目标数量，将得到的商值作为该第一采样步长，该计算机设备按照该第一采样步长，从位于该第一像素区域的边界上的多个像素点中，获取该第二目标数量的第一像素点。
需要说明的是，上述是均匀采样的方式来筛选出第一像素点，该计算机设备还可以采用非均匀采样的方式筛选出第一像素点。在一种可能示例中，该计算机设备还可以根据第一像素区域内多个像素点距离该目标特征点的远近，进行第一像素点的筛选，该计算机设备获取该多个像素点分别与该目标特征点之间的多个距离，根据该多个距离，将该多个像素点划分为多个像素点集合，每个像素点集合中像素点与目标特征点之间的距离位于该像素点集合对应的距离范围内。例如，集合A对应的距离范围为1.0到4.9，集合B对应的距离范围为5.0到9.9等。对于每个像素点集合，该计算机设备根据每个像素点集合与该目标特征点之间的距离，确定该像素点集合所对应的第一采样步长，根据该像素点集合对应的第一采样步长，获取该像素点集合内的多个像素点中多个第一像素点，从而得到第二目标数量的第一像素点。在一种可能示例中，像素点集合与目标特征点之间的距离可以为该像素点集合内多 个像素点与该目标特征点之间多个距离的平均值。例如，该计算机设备可以采取中心稠密边缘稀疏的方式进行筛选，也即是，该像素点集合与该目标特征点之间的距离越大，该像素点集合对应的第一采样步长越小，像素点集合与该目标特征点之间的距离越小，该像素点集合对应的第一采样步长越大；该计算机设备可以采取中心稀疏边缘稠密的方式进行筛选，则该像素点集合与该目标特征点之间的距离，与第一采样步长的对应关系则正好与中间稠密边缘稀疏的方式相反，此处不再赘述。在另一种示例中，该计算机设备还可以采取随机采样的方式，从该第一像素区域内筛选出第二目标数量的第一像素点。
还需要说明的是，本申请实施例中所提及的“多个”、“多种”及“多次”等具体指“多于一个”、“多于一种”、及“多于一次”。
如图3所示，以正方形的像素区域为例进行说明，黑色点为像素点，白色点为筛选出的像素点，对于正方形的像素区域，该计算机设备从正方形区域边界的多个像素点中筛选出像素点。当该像素区域边界上的像素点的数量等于第二目标数量时，该第一采样步长可以为1，当该像素区域边界上的像素点的数量大于第二目标数量时，该采样步长可以大于1。例如，该第二目标数量可以为9，如图3中(a)所示，像素区域为一个包括3×3个像素点的正方形区域，该采样步长可以为1个像素点。如图3中(b)所示，像素区域可以为一个包括5×5个像素点的正方形区域，该采样步长可以为2个像素点。如图3中(c)所示，该像素区域可以为一个9×9的正方形区域，该采样步长可以为4个像素点。
S203、该计算机设备根据该多个第一像素点的像素值，确定该多个第一像素点的像素差异。
像素差异用于指示多个像素点的像素值的变化程度，在一些实施例中，多个像素点的像素差异可以为该多个像素点的像素方差或者该多个像素点的梯度矩阵的最小特征值；该多个像素点的像素方差用于表示多个像素点的像素值相对于像素平均值的变化程度，该多个像素点的梯度矩阵用于表示该多个像素点的像素值分别在水平梯度上相对于平均值的变化程度和垂直梯度上相对于平均值的变化程度。则该计算机设备可以采用该多个第一像素点的像素方差或者多个第一像素点的梯度矩阵的最小特征值，表示该多个第一像素 点的像素差异。
在一些实施例中，该计算机确定多个第一像素点的像素值的平均值，根据该每个第一像素点的像素值与该平均值的差值，确定该多个第一像素点的像素差异。在一种可能示例中，如果采用像素方差表示像素差异，则该计算机设备可以根据该多个第一像素点的像素值的平均值、每个第一像素点的像素值，通过以下公式一，确定该多个第一像素点的像素方差；
Figure PCTCN2020085000-appb-000001
其中，N为该多个第一像素点的数量，该N可以为第二目标数量，例如，N＝l×l，l为3，M为多个第一像素点的像素值的平均值，例如，
Figure PCTCN2020085000-appb-000002
I u为多个第一像素点的像素值，var表示该多个第一像素点的像素平方差。u表示多个第一像素点。
在一些实施例中，该计算机设备还可以采用该多个第一像素点的梯度矩阵，表示该像素差异。则该计算机设备可以根据该多个第一像素点的像素值，确定该多个第一像素点的像素值的水平梯度和像素值的垂直梯度，根据该像素值的水平梯度和像素值的垂直梯度，获取该多个第一像素点的梯度矩阵的最小特征值，将该最小特征值确定为该多个第一像素点的像素差异。
在一种可能示例中，该梯度矩阵可以为协方差矩阵，变量为多个像素点的像素值的垂直梯度和水平梯度。则该计算机设备可以根据多个第一像素点的像素值的水平梯度和像素值的垂直梯度，通过以下公式二，确定该多个第一像素点的梯度矩阵的最小特征值；
Figure PCTCN2020085000-appb-000003
其中，g x(u)为多个第一像素点u的像素值的水平梯度，g y(u)为多个第一像素点u的像素值的垂直梯度，u表示多个第一像素点。矩阵S表示多 个第一像素点的梯度矩阵。该计算机设备可以确定矩阵S的最小特征值，得到该多个第一像素点的像素差异。
需要说明的是，当目标特征点为位于两个采样像素点之间的像素点时，该目标特征点的坐标可能为非整数坐标，从而第一像素点的坐标也可能为非整数坐标，计算机设备可以根据该第一像素点的周围像素点的像素值，确定该第一像素点的像素值。在一种可能示例中，该计算机设备可以采用双向线性插值算法确定非整数位置的像素点的像素值，该计算机设备获取该第一像素点的周围像素点的像素值，根据以下公式三，确定该第一像素点的像素值；
公式三：
I＝i0*(u-u3)*(v-v3)+i1*(u-u2)*(v-v2)+i2*(u-u1)*(v-v1)+i3*(u-u0)*(v-v0)；
如图4所示，I表示第一像素点的像素值，i0表示该第一像素点的左上角像素点的像素值，位置表示为(u0,v0)，i1表示第一像素点的右上角像素点的像素值，位置表示为(u1,v1)，i2表示表示第一像素点的左下角像素点的像素值，位置表示为(u2,v2)，i3表示第一像素点的右下角像素点的像素值，位置表示为(u3,v3)，(u，v)表示第一像素点的位置。
S204、计算机设备根据该第一像素区域中多个第一像素点的像素差异以及该目标特征点，确定该第一图像中包括该目标特征点的第二像素区域。
其中，该第二像素区域中多个第二像素点的像素差异大于该多个第一像素点的像素差异，该多个第二像素点与该多个第一像素点的数量相同。该计算机设备获取目标差异值，当该多个第一像素点的像素差异小于该目标差异值时，该计算机设备对该第一像素区域进行调整，得到像素差异大于该目标差异值的第二像素区域。本申请实施例中，该计算机设备可以通过扩大该第一像素区域的方式，得到像素差异更大的第二像素区域，或者，通过移动该第一像素区域的方式，得到第二像素区域。相应的，本步骤可以通过以下两种方式实现。
第一种方式、当该多个第一像素点的像素差异小于目标差异值时，计算机设备根据该目标特征点，获取大于该第一像素区域且包括该目标特征点的 第二像素区域。
本申请实施例中，该计算机设备可以按照一定的扩大规则，扩大该第一像素区域，或者，随机扩大该第一像素区域。在一些实施例中，该第一像素区域的形状可以为中心对称形状，该计算机设备按照目标扩大系数，以该目标特征点为中心，将该第一像素区域扩大为包括该目标特征点的第二像素区域。例如，该第一像素区域可以为一个3×3的正方形，正方形的边长l＝3，该计算机设备按照(2l-1)的扩大系数，将3×3的正方形扩大为5×5的正方形。
本申请实施例中，该计算机设备可以通过多次调整第一像素区域，以得到像素差异大于目标差异值的第二像素区域。该过程可以为：当该多个第一像素点的像素差异小于目标差异值时，该计算机设备可以根据该目标特征点，获取包括该目标特征点的第三像素区域，该第三像素区域中多个第三像素点与该多个第一像素点的数量相同；该计算机设备根据该第三像素区域中多个第三像素点的像素值，确定该多个第三像素点的像素差异；如果该多个第三像素点的像素差异不小于目标差异值，该计算机设备将该第三像素区域确定为该第二像素区域。在一种可能实施方式中，该计算机设备可以按照目标扩大系数，以该目标特征点为中心，将该第一像素区域扩大为包括该目标特征点的第三像素区域。与上述获取多个第一像素点的方式同理，该计算机设备也可以按照一定的采样步长，从第三像素区域中获取多个第三像素点。在一种可能实施方式中，该计算机设备获取该第三像素区域内多个第三像素点的过程可以包括：该计算机设备可以根据从该第一像素区域到该第三像素区域的扩大系数，将该第一像素区域的第一采样步长增大为第二采样步长，按照该第二采样步长，从该第三像素区域中获取与该多个第一像素点数量相同的第三像素点。例如，对于3×3的正方形，该计算机设备每两个像素点采集一个第一像素点，则对于5×5的正方形，该计算机设备每三个像素点采集一个第三像素点。
如果该第三像素区域内多个第三像素点的像素差异小于目标差异值，则该计算机设备继续获取扩大第三像素区域，直至获取到像素差异不小于目标差异值的第二像素区域。该计算机设备获取该多个第三像素点的像素差异的过程，与上述获取多个第一像素点的像素差异的过程同理，此处不再赘述。
在一些实施例中，该计算机设备还可以基于像素区域的大小，判断是否继续扩大第三像素区域。该过程可以包括：当多个第三像素点的像素差异小于该目标差异值时，该计算机设备检测该第三像素区域的大小是否大于目标阈值；当第三像素区域的大小不大于该目标阈值时，该计算机设备确定大于该第三像素区域的第四像素区域；该计算机设备基于该第四像素区域中多个第四像素点的像素差异以及该目标特征点，确定该第一图像中包括该目标特征点的第二像素区域，其中，该多个第四像素点与该多个第一像素点的数量相同。
该计算机设备根据每个第三像素点的像素值，获取该多个第三像素点的像素差异。在一些实施例中，该计算机设备根据该多个第三像素点的像素值，获取该多个第三像素点的像素值的平均值，根据该平均值，该计算机设备获取该多个第三像素点的像素方差，将该像素方差确定为该多个第三像素点的像素差异。在一些实施例中，该计算机设备根据该多个第三像素点的像素值，获取该多个第三像素点的像素值的水平梯度和像素值的垂直梯度，该计算机设备根据该多个第三像素点的像素值的水平梯度和像素值的垂直梯度，获取该多个第三像素点的梯度矩阵的最小特征值，将该最小特征值确定为该多个第三像素点的像素差异。
该过程与上述对第一像素点的处理过程同理，此处不再赘述。该目标差异值、该目标阈值均可以基于需要进行设置，本申请实施例对此不做具体限定，例如，该目标阈值可以为13×13，也即是该目标特征点的像素区域最大可以为一个包括13×13个像素点的区域。
第二种方式、当该多个第一像素点的像素差异小于目标差异值时，计算机设备根据该目标特征点，获取与该第一像素区域的大小相同的第二像素区域。
在一些实施例中，该计算机设备可以预先存储有目标移动轨迹，该目标移动轨迹用于指示第一像素区域的移动过程。则本步骤可以包括：该计算机设备根据该目标特征点，按照目标移动轨迹，将该第一像素区域移动至包括该目标特征点的第二像素区域。例如，该目标移动轨迹可以为向右移动一个单位，则该以目标特征点为中心的3×3的正方形进行移动后，得到的第二像素区域中，该目标特征点为正方形中左边的中点位置。
本申请实施例中，该计算机设备可以通过多次移动第一像素区域，以得到像素差异大于目标差异值的第二像素区域。也即是，该计算机设备可以通过第二种方式，获取第三像素区域，基于第三像素区域内多个第三像素点的像素差异，获取第二像素区域。该过程与上述第一种方式的过程同理，此处不再具体赘述。不同的是，由于第二种方式中，未对第一像素区域进行扩大，因此，该计算机设备无需增大第一采样步长获取第三像素点，该计算机设备可以直接按照该第一像素区域的第一采样步长，从该第三像素区域中获取与该多个第一像素点数量相同的第三像素点。
需要说明的是，无论像素区域如何变化，由于计算机设备所获取的第一像素点或者第二像素点的数量均相同，因此，保证了在增大像素差异时，实际参与计算的像素点的数量不变，并且，基于像素区域内像素点的差异与目标差异值进行判断，持续的试探性增大或移动第一像素区域，从而得到的第二像素区域内像素点之间差异较大，保证特像素区域内像素特征变化显著，避免了在一些平滑或者纹理不丰富的像素区域，由于像素区域内像素亮度变化不显著，而导致的确定运动信息不准确的问题。并且，由于不断增大或移动像素区域时，像素区域的像素差异变大，采用像素差异较大的区域来表示目标特征点周围区域的像素变化特征，既保证了参与计算的像素点之间像素差异较大，又保证了参与计算的像素点数量固定不变，不会增大计算量，从而平衡了参与计算的像素点的数量和像素点的像素差异，使得在目标特征点跟踪计算复杂度不变的情况下，其像素丰富程度增大，提高了目标特征点跟踪的鲁棒性，使其可以在平滑或者缺乏纹理的环境下稳定执行，提高了所适用的场景的广泛性。
S205、计算机设备根据该多个第二像素点和第二图像，获取目标特征点的运动信息。
该运动信息用于指示该目标特征点在该第一图像和该第二图像中的位置变化。该计算机设备可以根据该目标特征点在第一图像的第一位置，确定第二图像中对应位置上的起始像素点，获取该起始像素点的像素区域中多个第四像素点，根据该多个第四像素点和该多个第二像素点，基于高斯牛顿算法，确定出该目标特征点的运动信息。该计算机设备还可以得到该目标特征点在该第二图像的位置。该多个第四像素点的数量与多个第二像素点的数量相同。
在一些实施例中，该运动信息可以包括在该目标特征点分别在图像坐标轴的x轴、y轴的移动距离。该计算机设备根据该目标数量的第二像素点的像素值和该多个第四像素点的像素值，基于以下公式四，确定该目标特征点从第一图像到第二图像的运动信息；
Figure PCTCN2020085000-appb-000004
其中，T t i表示目标特征点i的第二像素区域内多个第二像素点，d i t+1表示目标特征点i的运动信息。d i t+1＝(u x，u y)分别表示x轴、y轴上两维的移动距离。u表示第二像素点，I t表示第一图像中多个第二像素点的像素值，I t+1表示第二图像中多个第四像素点的像素值，假设多个第二像素点的像素值在第一图像和第二图像中相同，计算机设备可以最小化第一图像中多个第二像素点和第二图像中多个第四像素点的像素值之间的像素差求解d i t+1。
在一些实施例中，该运动信息可以采用单应矩阵表示，则该计算机设备根据该目标数量的第二像素点的像素值和该多个第四像素点的像素值，基于以下公式五，确定该目标特征点从第一图像到第二图像的运动信息；
Figure PCTCN2020085000-appb-000005
其中，T t i表示目标特征点i的第二像素区域内多个第二像素点，H i t+1表示目标特征点i的运动信息。在一些实施例中，该计算机设备可以将该运动信息表示为：
Figure PCTCN2020085000-appb-000006
其中，h 11和h 22分别表示第二像素区域从第一图像到第二图时像在图像坐标系的x轴方向和y轴方向的缩放系数，x轴方向和y轴方向可以分别为图像中的水平方向和垂直方向。h 11、 h 221同时也和h 12和h 21共同指示沿着x轴和第二像素区域法向量旋转的过程。h 12和h 21分别表示第二像素区域从第一图像到第二图像时在图像坐标系的x轴方向和y轴方向上的投影，h 13和h 23分别表示第二像素区域从第一图像到第二图像时在图像坐标系的x轴方向和y轴方向的移动距离，h 31和h 32分别表示第二像素区域在图像坐标系中x轴方向和y轴方向的切变参数，在一些实施例中，该切变参数可以为第二像素区域在x轴方向和y轴方向上的形变比率。例如，特征点在第一图像中第一像素区域可以为正方形，特征点在第二图像中对应的第二像素区域可以梯形，如果该正方形和梯形的上下边的方向均为x轴方向，则该h 31表示梯形上边和下边的边长的变化率，h 32表示梯形的左边和右边的边长的变化率。u表示第二像素点，I t表示第一图像中多个第二像素点的像素值，I t+1表示第二图像中多个第四像素点的像素值，假设多个第二像素点的像素值在第一图像和第二图像中相同，计算机设备可以最小化第一图像中多个第二像素点和第二图像中多个第四像素点的像素值之间的像素差异来求解H i t+1。
需要说明的是，本申请实施例中的目标特征点在图像中分布稀疏，因此也可以成为稀疏目标特征点。该第一图像和第二图像可以为摄像机所拍摄的视频内两帧连续图像。摄像机实时拍摄的视频流提供了摄像机对外部环境在不同时刻的观察。目标特征点的运动信息常常用于运动检测，运动估计，实时定位，三维重建，物体分割等过程。例如，如图5所示，对于目标特征点跟踪过程。目标特征点的运动轨迹，即从第一次检测到的图像位置到目标特征点目前在图像上的位置，用每个目标特征点后的白色线表示。
下面以图6所示的流程图，对上述步骤S201-S205进行介绍，如图6所示，该计算机设备可以获取l×l大小的第一像素区域，基于第一像素区域内像素点的像素差异，当像素差异小于目标差异值时，判断该第一像素区域的大小是否大于目标阈值l m×l m，如果不大于l m×l m，则继续增大第一像素区域， 基于增大后的第三像素区域，增大采用步长，以筛选出第三像素区域内多个第三像素点，当多个像素点的像素差异不小于目标差异值时，或者，第三像素区域的大小大于l m×l m，不再增大第三像素区域，将第三像素区域作为第二像素区域，如果该第三像素区域内像素点的像素差异小于目标差异值时，且第三像素区域的大小不大于l m×l m，则继续增大第三像素区域，直到获取到像素差异不小于目标差异值或者大于l m×l m的第二像素区域。
本申请实施例中，通过基于包括目标特征点的第一像素区域，获取了像素差异大于第一像素区域且像素点数量不变的第二像素区域，从而可以基于原像素点数量的多个第二像素点进行计算，以得到目标特征点的运动信息，由于保持像素点数量不变的前提下增大参与计算的数据的像素差异，平衡了计算复杂度和信息丰富程度，保证目标特征点运动信息的准确性的前提下，提高了确定目标特征点运动信息的效率。
图7是本申请实施例提供的一种任务执行方法的流程示意图。该方法应用在目标设备上，参见图7，该方法包括：
S701、目标设备获取第一图像和第二图像。
该第一图像和该第二图像包括相同的对象；该目标设备为采集该第一图像和第二图像的设备。该目标设备可以在移动过程中拍摄包括相同的对象的第一图像和第二图像，在一些实施例中，该目标设备可以为计算机设备，该计算机设备用于确定第一图像中目标特征点的运动信息，该运动信息用于指示该目标特征点在该第一图像和该第二图像的位置变化。例如，该计算机设备可以为手机，该则该计算机设备可以一边移动，一边在移动过程中实时拍摄图像。在一些实施例中，该目标设备可以不为该计算机设备，例如，该目标设备在移动过程中拍摄第一图像和第二图像，实时将拍摄得到的第一图像和第二图像发送至该计算机设备。
S702、目标设备获取该第一图像中该对象上的目标特征点的运动信息。
本申请实施例中，该运动信息用于指示该目标特征点在该第一图像和该第二图像中的位置变化。如果该目标设备为该计算机设备，该目标设备可以基于上述发明实施例中，步骤S201-S205的过程，获取该目标特征点的运动信息。如果目标设备不为该计算机设备，该计算机设备可以基于上述发明实 施例中，步骤S201-S205的过程，获取该目标特征点的运动信息，实时将该目标特征点的运动信息发送至目标设备。该目标设备接收该计算机设备发送的目标特征点的运动信息。
S703、目标设备基于该目标特征点的运动信息，执行目标任务。
该目标任务可以为路线规划任务，也可以为对象识别任务。
在一些实施例中，该目标任务可以为路线规划任务，该目标设备可以基于目标特征点的运动信息，构建该目标设备周围环境的场景对象，基于该场景对象，来规划路线。该过程可以包括：当该目标特征点的数目为多个时，该目标设备可以基于多个目标特征点的运动信息，确定距离该目标设备不超过第一阈值的至少一个场景对象；及该目标设备可以根据距离该目标设备不超过第二阈值的目的地的位置和该至少一个场景对象，确定该目标设备到达该目的地的第一目标路线，该第二阈值大于该第一阈值。
在一个具体示例中，计算机设备可以通过定位获取当前位置，基于该当前位置，指引用户到附近最近的餐厅，商店，洗手间等；例如，用户在一个大型商场的某一层，目标设备可以基于周围店铺的名称，精确定位该目标设备在该层的位置，例如，位于服装店A门前，该目标设备可以指引用户到达手表店B处。
该目标任务可以为路线规划任务，在一些实施例中，该目标设备还可以对该目标设备周围的场景对象进行一些操作，例如，该目标设备可以为机器人，该目标设备可以将水杯放置在桌子上。则该目标设备确定距离该目标设备不超过第一阈值的至少一个场景对象后，该目标设备可以根据该至少一个场景对象和待执行的目标任务，确定该目标设备的第二目标路线，该目标任务是指在该至少一个场景对象中目标场景对象上执行目标操作。
在一个具体示例中，以目标设备为VR(Virtual Reality，虚拟现实)设备或者AR(Augmented Reality，增强现实技术)设备为例，VR设备和AR设备可以基于场景中的场景对象，将虚拟的物品放置在真实环境中。
在一些实施例中，该目标任务包括对象识别任务，该目标设备可以基于该第一图像和第二图像，识别出周围环境中运动对象和静止对象，还可以识别出运动对象的对象类别，例如运动对象是人还是车辆，该过程可以包括：当该目标特征点的数目为多个时，该目标设备可以基于多个目标特征点的运 动信息，确定出该多个目标特征点中运动信息符合目标条件的多个第一特征点，该多个第一特征点用于指示该第一图像和该第二图像包括的多个对象中的运动对象；该目标设备可以基于该多个第一特征点在该第一图像或第二图像的位置，确定该运动对象所属的对象类别。其中，该对象类别可以包括：车辆、人、或动物等。该目标条件可以包括：多个特征点中运动信息与其他特征点的运动信息不同，或者，与多个特征点中超过目标数量的特征点的运动信息不同。例如，与200个特征点中80％的特征点的运动信息不同。
在一些实施例中，该目标设备识别出周围环境中运动对象所属的类别后，该目标设备还可以进一步基于周围走动的人、行驶的车辆，或者静止的房屋，确定第三目标路线，该第三目标路线可以是避开该运动对象的路线。以在移动过程中避开障碍物，更准确的进行路线规划。
在一个具体示例中，如果选取目标特征点的在某一个物体内，计算机设备可以判断某一个物体是否有移动，如果有移动的话，往哪个方向移动；检测出运动的物体后，还可以把物体分割出来，判断是否为人或其他动物，因为动物常常在前面，场景常常在后面。有了这些信息，手机、机器人、或自动驾驶的车辆等具备了基本的对象识别功能。
在一个具体示例中，计算机设备可以基于整个图像所有目标特征点的位置变化，判断目标设备是否移动，如果该目标设备移动，还可以确定该目标设备的移动方向、移动轨迹等，进一步反算出周围的场景对象。例如，周围的座椅、障碍物等，基于周围的场景对象，重建出该目标设备的场景，该场景包括位于该目标设备周围的多个场景对象。例如，以目标设备为机器人为例，在机器人执行任务时，由于重建了场景，机器人可以躲开障碍物；又如，机器人可以基于当前所处的位置，进一步规划下一步的路径，到达其要达到的地方，执行相应的任务。
在一些实施例中，该目标设备确定距离该目标设备不超过第一阈值的至少一个场景对象后，该目标设备还可以根据该至少一个场景对象，将该场景对象和虚拟的对象进行结合，为用户展示虚实结合环境。该过程可以包括：该目标设备获取至少一个虚拟场景对象，该目标设备根据该至少一个虚拟场景对象的位置和该至少一个场景对象的位置，构建虚拟场景，在目标设备的屏幕上显示该虚拟场景。在一个具体示例中，由于已知当前所处位置，可以 将真实的环境和虚拟的环境通过相对应的位置关系进行渲染，显示在计算机设备的屏幕上。
本申请实施例中，目标设备可以利用第一图像中对象上的目标特征点的运动信息，执行目标任务，例如，规划路线、运动检测、识别对象所属的对象类别等，且该目标特征点的运动信息的确定过程准确、高效，提高了执行目标任务的准确性和效率。
图8是本申请实施例提供的一种图像特征点的运动信息确定装置的结构示意图。参见图8，该装置包括：
确定模块801，用于确定第一图像和第二图像，该第一图像和该第二图像包括相同的对象。
该确定模块801，还用于基于该第一图像中该对象上的目标特征点，确定该第一图像中包括该目标特征点的第一像素区域。
该确定模块801，还用于根据该第一像素区域中多个第一像素点的像素差异以及该目标特征点，确定该第一图像中包括该目标特征点的第二像素区域，该第二像素区域中多个第二像素点的像素差异大于该多个第一像素点的像素差异，该多个第二像素点与该多个第一像素点的数量相同，该像素差异用于指示多个像素点的像素值的变化程度。
获取模块802，用于根据该多个第二像素点和该第二图像，获取该目标特征点的运动信息，该运动信息用于指示该目标特征点在该第一图像和该第二图像中的位置变化。
在一些实施例中，该确定模块801，用于当该第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据该目标特征点，获取大于该第一像素区域且包括该目标特征点的第二像素区域。
在一些实施例中，该确定模块801，用于当该第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据该目标特征点，获取与该第一像素区域的大小相同的第二像素区域，该第一像素区域和该第二像素区域所包括的像素点不同。
在一些实施例中，该确定模块801，还用于按照目标扩大系数，以该目标特征点为中心，将该第一像素区域扩大为包括该目标特征点的第二像素区 域。
在一些实施例中，该确定模块801，还用于根据该目标特征点，按照目标移动轨迹，将该第一像素区域移动至包括该目标特征点的第二像素区域。
在一些实施例中，该确定模块801，还用于当该多个第一像素点的像素差异小于目标差异值时，根据该目标特征点，获取包括该目标特征点的第三像素区域，该第三像素区域中多个第三像素点与该多个第一像素点的数量相同；根据该第三像素区域中多个第三像素点的像素值，确定该多个第三像素点的像素差异；及当该多个第三像素点的像素差异不小于目标差异值时，将该第三像素区域确定为该第二像素区域。
在一些实施例中，该确定模块801，还用于根据从该第一像素区域到该第三像素区域的扩大系数，将该第一像素区域的第一采样步长增大为第二采样步长，按照该第二采样步长，从该第三像素区域中获取与该多个第一像素点数量相同的第三像素点。
在一些实施例中，该确定模块801，还用于按照该第一像素区域的第一采样步长，从该第三像素区域中获取与该多个第一像素点数量相同的第三像素点。
在一些实施例中，该装置还包括：
检测模块，用于当该多个第三像素点的像素差异小于该目标差异值时，检测该第三像素区域的大小是否大于目标阈值。
该确定模块801，还用于当第三像素区域的大小不大于该目标阈值时，确定大于该第三像素区域的第四像素区域。
该确定模块801，还用于基于该第四像素区域中多个第四像素点的像素差异以及该目标特征点，确定该第一图像中包括该目标特征点的第二像素区域，该多个第四像素点与该多个第一像素点的数量相同。
在一些实施例中，该像素差异为该多个像素点的像素方差或者该多个像素点的梯度矩阵的最小特征值，该像素方差用于表示该多个像素点的像素值相对于像素平均值的变化程度，该梯度矩阵用于表示该多个像素点分别在水平梯度上相对于像素平均值的变化程度和垂直梯度上相对于像素平均值的变化程度。
在一些实施例中，该第一像素区域和该第二像素区域的形状为正方形、 长方形、圆形、环形、不规则多边形或不规则曲边形中的任一形状。
在一些实施例中，该确定模块801，还用于根据该第一像素区域的第一采样步长，在该第一像素区域的区域边界上的多个像素点中，获取该多个第一像素点；及根据该多个第一像素点的像素值，确定该多个第一像素点的像素差异。
通过基于包括目标特征点的第一像素区域，获取了像素差异大于第一像素区域且像素点数量不变的第二像素区域，从而可以基于原像素点数量的多个第二像素点进行计算，以得到目标特征点的运动信息，由于保持像素点数量不变的前提下增大参与计算的数据的像素差异，平衡了计算复杂度和信息丰富程度，保证目标特征点运动信息的准确性的前提下，提高了确定目标特征点运动信息的效率。
图9是本申请实施例提供的一种任务执行装置的结构示意图。参见图9，该装置包括：
获取模块901，用于获取目标设备的第一图像和第二图像，该第一图像和该第二图像包括相同的对象。
该获取模块901，还用于获取该第一图像中该对象上的目标特征点的运动信息，该运动信息用于指示该目标特征点在该第一图像和该第二图像中的位置变化。
任务处理模块902，用于基于该目标特征点的运动信息，执行目标任务。
在一些实施例中，该目标任务包括路线规划任务，该任务处理模块902，还用于当该目标特征点的数目为多个时，基于多个目标特征点的运动信息，确定距离该目标设备不超过第一阈值的至少一个场景对象；及根据距离该目标设备不超过第二阈值的目的地的位置和该至少一个场景对象，确定该目标设备到达该目的地的第一目标路线，该第二阈值大于该第一阈值。
在一些实施例中，该目标任务包括对象识别任务，该任务处理模块902，还用于当该目标特征点的数目为多个时，基于多个目标特征点的运动信息，确定出该多个目标特征点中运动信息符合目标条件的多个第一特征点，该多个第一特征点用于指示该第一图像和该第二图像包括的多个对象中的运动对象；及基于该多个第一特征点在该第一图像或第二图像的位置，确定该运动 对象所属的对象类别。
本申请实施例中，目标设备可以利用第一图像中对象上的目标特征点的运动信息，执行目标任务，例如，规划路线、运动检测、或识别对象所属的对象类别等，且该目标特征点的运动信息的确定过程准确、高效，提高了执行目标任务的准确性和效率。
上述所有可选技术方案，可以采用任意结合形成本公开的可选实施例，在此不再一一赘述。
需要说明的是：上述实施例提供的图像特征点的运动信息确定装置在确定特征点的运动信息时，以及任务执行装置与执行任务时，仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将计算机设备的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。另外，上述实施例提供的图像特征点的运动信息确定装置与图像特征点的运动信息确定方法，以及任务执行装置与任务执行方法实施例属于同一构思，其具体实现过程详见方法实施例，这里不再赘述。
在一些实施例中，还提供了一种计算机设备，包括存储器和处理器，存储器中存储有计算机可读指令，该处理器执行计算机可读指令时实现上述图像特征点的运动信息确定方法实施例中的步骤。
在一些实施例中，还提供了一种目标设备，包括存储器和处理器，存储器中存储有计算机可读指令，该处理器执行计算机可读指令时实现上述任务执行方法实施例中的步骤。
在一些实施例中，提供了一种计算机可读存储介质，存储有计算机可读指令，该计算机可读指令被处理器执行时实现上述图像特征点的运动信息确定方法实施例中的步骤，或者，实现上述任务执行方法实施例中的步骤。
图10是本申请实施例提供的一种终端的结构示意图。该终端1000可以是：智能手机、平板电脑、MP3播放器(Moving Picture Experts Group Audio Layer III，动态影像专家压缩标准音频层面3)、MP4(Moving Picture Experts  Group Audio Layer IV，动态影像专家压缩标准音频层面4)播放器、笔记本电脑或台式电脑。终端1000还可能被称为用户设备、便携式终端、膝上型终端、台式终端等其他名称。
通常，终端1000包括有：处理器1001和存储器1002。
处理器1001可以包括一个或多个处理核心，比如4核心处理器、8核心处理器等。处理器1001可以采用DSP(Digital Signal Processing，数字信号处理)、FPGA(Field－Programmable Gate Array，现场可编程门阵列)、PLA(Programmable Logic Array，可编程逻辑阵列)中的至少一种硬件形式来实现。处理器1001也可以包括主处理器和协处理器，主处理器是用于对在唤醒状态下的数据进行处理的处理器，也称CPU(Central Processing Unit，中央处理器)；协处理器是用于对在待机状态下的数据进行处理的低功耗处理器。在一些实施例中，处理器1001可以在集成有GPU(Graphics Processing Unit，图像处理器)，GPU用于负责显示屏所需要显示的内容的渲染和绘制。一些实施例中，处理器1001还可以包括AI(Artificial Intelligence，人工智能)处理器，该AI处理器用于处理有关机器学习的计算操作。
存储器1002可以包括一个或多个计算机可读存储介质，该计算机可读存储介质可以是非暂态的。存储器1002还可包括高速随机存取存储器，以及非易失性存储器，比如一个或多个磁盘存储设备、闪存存储设备。在一些实施例中，存储器1002中的非暂态的计算机可读存储介质用于存储至少一个指令，该至少一个指令用于被处理器1001所执行以实现本申请中方法实施例提供的图像特征点的运动信息确定方法或任务执行方法。
在一些实施例中，终端1000还可选包括有：外围设备接口1003和至少一个外围设备。处理器1001、存储器1002和外围设备接口1003之间可以通过总线或信号线相连。各个外围设备可以通过总线、信号线或电路板与外围设备接口1003相连。具体地，外围设备包括：射频电路1004、触摸显示屏1005、摄像头1006、音频电路1007、定位组件1008和电源1009中的至少一种。
外围设备接口1003可被用于将I/O(Input/Output，输入/输出)相关的至少一个外围设备连接到处理器1001和存储器1002。在一些实施例中，处理器1001、存储器1002和外围设备接口1003被集成在同一芯片或电路板上； 在一些其他实施例中，处理器1001、存储器1002和外围设备接口1003中的任意一个或两个可以在单独的芯片或电路板上实现，本实施例对此不加以限定。
射频电路1004用于接收和发射RF(Radio Frequency，射频)信号，也称电磁信号。射频电路1004通过电磁信号与通信网络以及其他通信设备进行通信。射频电路1004将电信号转换为电磁信号进行发送，或者，将接收到的电磁信号转换为电信号。可选地，射频电路1004包括：天线系统、RF收发器、一个或多个放大器、调谐器、振荡器、数字信号处理器、编解码芯片组、用户身份模块卡等等。射频电路1004可以通过至少一种无线通信协议来与其它终端进行通信。该无线通信协议包括但不限于：城域网、各代移动通信网络(2G、3G、4G及5G)、无线局域网和/或WiFi(Wireless Fidelity，无线保真)网络。在一些实施例中，射频电路1004还可以包括NFC(Near Field Communication，近距离无线通信)有关的电路，本申请对此不加以限定。
显示屏1005用于显示UI(User Interface，用户界面)。该UI可以包括图形、文本、图标、视频及其它们的任意组合。当显示屏1005是触摸显示屏时，显示屏1005还具有筛选出在显示屏1005的表面或表面上方的触摸信号的能力。该触摸信号可以作为控制信号输入至处理器1001进行处理。此时，显示屏1005还可以用于提供虚拟按钮和/或虚拟键盘，也称软按钮和/或软键盘。在一些实施例中，显示屏1005可以为一个，设置终端1000的前面板；在另一些实施例中，显示屏1005可以为至少两个，分别设置在终端1000的不同表面或呈折叠设计；在再一些实施例中，显示屏1005可以是柔性显示屏，设置在终端1000的弯曲表面上或折叠面上。甚至，显示屏1005还可以设置成非矩形的不规则图形，也即异形屏。显示屏1005可以采用LCD(Liquid Crystal Display，液晶显示屏)、OLED(Organic Light-Emitting Diode,有机发光二极管)等材质制备。
摄像头组件1006用于筛选出图像或视频。可选地，摄像头组件1006包括前置摄像头和后置摄像头。通常，前置摄像头设置在终端的前面板，后置摄像头设置在终端的背面。在一些实施例中，后置摄像头为至少两个，分别为主摄像头、景深摄像头、广角摄像头、长焦摄像头中的任意一种，以实现主摄像头和景深摄像头融合实现背景虚化功能、主摄像头和广角摄像头融合 实现全景拍摄以及VR(Virtual Reality，虚拟现实)拍摄功能或者其它融合拍摄功能。在一些实施例中，摄像头组件1006还可以包括闪光灯。闪光灯可以是单色温闪光灯，也可以是双色温闪光灯。双色温闪光灯是指暖光闪光灯和冷光闪光灯的组合，可以用于不同色温下的光线补偿。
音频电路1007可以包括麦克风和扬声器。麦克风用于筛选出用户及环境的声波，并将声波转换为电信号输入至处理器1001进行处理，或者输入至射频电路1004以实现语音通信。出于立体声筛选出或降噪的目的，麦克风可以为多个，分别设置在终端1000的不同部位。麦克风还可以是阵列麦克风或全向筛选出型麦克风。扬声器则用于将来自处理器1001或射频电路1004的电信号转换为声波。扬声器可以是传统的薄膜扬声器，也可以是压电陶瓷扬声器。当扬声器是压电陶瓷扬声器时，不仅可以将电信号转换为人类可听见的声波，也可以将电信号转换为人类听不见的声波以进行测距等用途。在一些实施例中，音频电路1007还可以包括耳机插孔。
定位组件1008用于定位终端1000的当前地理位置，以实现导航或LBS(Location Based Service，基于位置的服务)。定位组件1008可以是基于美国的GPS(Global Positioning System，全球定位系统)、中国的北斗系统、俄罗斯的格雷纳斯系统或欧盟的伽利略系统的定位组件。
电源1009用于为终端1000中的各个组件进行供电。电源1009可以是交流电、直流电、一次性电池或可充电电池。当电源1009包括可充电电池时，该可充电电池可以支持有线充电或无线充电。该可充电电池还可以用于支持快充技术。
在一些实施例中，终端1000还包括有一个或多个传感器1010。该一个或多个传感器1010包括但不限于：加速度传感器1011、陀螺仪传感器1012、压力传感器1013、指纹传感器1014、光学传感器1015以及接近传感器1016。
加速度传感器1011可以检测以终端1000建立的坐标系的三个坐标轴上的加速度大小。比如，加速度传感器1011可以用于检测重力加速度在三个坐标轴上的分量。处理器1001可以根据加速度传感器1011筛选出的重力加速度信号，控制触摸显示屏1005以横向视图或纵向视图进行用户界面的显示。加速度传感器1011还可以用于游戏或者用户的运动数据的筛选出。
陀螺仪传感器1012可以检测终端1000的机体方向及转动角度，陀螺仪 传感器1012可以与加速度传感器1011协同筛选出用户对终端1000的3D动作。处理器1001根据陀螺仪传感器1012筛选出的数据，可以实现如下功能：动作感应(比如根据用户的倾斜操作来改变UI)、拍摄时的图像稳定、游戏控制以及惯性导航。
压力传感器1013可以设置在终端1000的侧边框和/或触摸显示屏1005的下层。当压力传感器1013设置在终端1000的侧边框时，可以检测用户对终端1000的握持信号，由处理器1001根据压力传感器1013筛选出的握持信号进行左右手识别或快捷操作。当压力传感器1013设置在触摸显示屏1005的下层时，由处理器1001根据用户对触摸显示屏1005的压力操作，实现对UI界面上的可操作性控件进行控制。可操作性控件包括按钮控件、滚动条控件、图标控件、菜单控件中的至少一种。
指纹传感器1014用于筛选出用户的指纹，由处理器1001根据指纹传感器1014筛选出到的指纹识别用户的身份，或者，由指纹传感器1014根据筛选出到的指纹识别用户的身份。在识别出用户的身份为可信身份时，由处理器1001授权该用户执行相关的敏感操作，该敏感操作包括解锁屏幕、查看加密信息、下载软件、支付及更改设置等。指纹传感器1014可以被设置终端1000的正面、背面或侧面。当终端1000上设置有物理按键或厂商Logo时，指纹传感器1014可以与物理按键或厂商Logo集成在一起。
光学传感器1015用于筛选出环境光强度。在一个实施例中，处理器1001可以根据光学传感器1015筛选出的环境光强度，控制触摸显示屏1005的显示亮度。具体地，当环境光强度较高时，调高触摸显示屏1005的显示亮度；当环境光强度较低时，调低触摸显示屏1005的显示亮度。在另一个实施例中，处理器1001还可以根据光学传感器1015筛选出的环境光强度，动态调整摄像头组件1006的拍摄参数。
接近传感器1016，也称距离传感器，通常设置在终端1000的前面板。接近传感器1016用于筛选出用户与终端1000的正面之间的距离。在一个实施例中，当接近传感器1016检测到用户与终端1000的正面之间的距离逐渐变小时，由处理器1001控制触摸显示屏1005从亮屏状态切换为息屏状态；当接近传感器1016检测到用户与终端1000的正面之间的距离逐渐变大时，由处理器1001控制触摸显示屏1005从息屏状态切换为亮屏状态。
本领域技术人员可以理解，图10中示出的结构并不构成对终端1000的限定，可以包括比图示更多或更少的组件，或者组合某些组件，或者采用不同的组件布置。
图11是本申请实施例提供的一种服务器的结构示意图，该服务器1100可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上处理器(central processing units，CPU)1101和一个或一个以上的存储器1102，其中，该存储器1102中存储有至少一条计算机可读指令，该至少一条计算机可读指令由该处理器1101加载并执行以实现上述各个方法实施例提供的图像特征点的运动信息确定方法或任务执行方法。当然，该服务器还可以具有有线或无线网络接口、键盘以及输入输出接口等部件，以便进行输入输出，该服务器还可以包括其他用于实现设备功能的部件，在此不做赘述。
在示例性实施例中，还提供了一种计算机可读存储介质，例如包括指令的存储器，上述指令可由终端中的处理器执行以完成上述实施例中的图像特征点的运动信息确定方法或任务执行方法。例如，该计算机可读存储介质可以是ROM(Read-Only Memory，只读存储器)、RAM(random access memory，随机存取存储器)、CD-ROM(Compact Disc Read-Only Memory，只读光盘)、磁带、软盘和光数据存储设备等。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，该的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
以上所述仅为本申请的较佳实施例，并不用以限制本申请，凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (20)

  1. 一种图像特征点的运动信息确定方法，由计算机设备执行，所述方法包括：
    确定第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
    基于所述第一图像中所述对象上的目标特征点，确定所述第一图像中包括所述目标特征点的第一像素区域；
    根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述第二像素区域中多个第二像素点的像素差异大于所述多个第一像素点的像素差异，所述多个第二像素点与所述多个第一像素点的数量相同，所述像素差异用于指示多个像素点的像素值的变化程度；及
    根据所述多个第二像素点和所述第二图像，获取所述目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，包括：
    当所述第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取大于所述第一像素区域且包括所述目标特征点的第二像素区域。
  3. 根据权利要求1所述的方法，其特征在于，所述根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，包括：
    当所述第一像素区域中多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取与所述第一像素区域的大小相同的第二像素区域，所述第一像素区域和所述第二像素区域所包括的像素点不同。
  4. 根据权利要求2所述的方法，其特征在于，所述根据所述目标特征点，获取大于所述第一像素区域且包括所述目标特征点的第二像素区域包括：
    按照目标扩大系数，以所述目标特征点为中心，将所述第一像素区域扩 大为包括所述目标特征点的第二像素区域。
  5. 根据权利要求3所述的方法，其特征在于，所述根据所述目标特征点，获取与所述第一像素区域的大小相同的第二像素区域包括：
    根据所述目标特征点，按照目标移动轨迹，将所述第一像素区域移动至包括所述目标特征点的第二像素区域。
  6. 根据权利要求1所述的方法，其特征在于，所述根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域包括：
    当所述多个第一像素点的像素差异小于目标差异值时，根据所述目标特征点，获取包括所述目标特征点的第三像素区域，所述第三像素区域中多个第三像素点与所述多个第一像素点的数量相同；
    根据所述第三像素区域中多个第三像素点的像素值，确定所述多个第三像素点的像素差异；及
    当所述多个第三像素点的像素差异不小于目标差异值时，将所述第三像素区域确定为所述第二像素区域。
  7. 根据权利要求6所述的方法，其特征在于，所述根据所述第三像素区域中多个第三像素点的像素值，确定所述多个第三像素点的像素差异之前，所述方法还包括：
    根据从所述第一像素区域到所述第三像素区域的扩大系数，将所述第一像素区域的第一采样步长增大为第二采样步长，按照所述第二采样步长，从所述第三像素区域中获取与所述多个第一像素点数量相同的第三像素点。
  8. 根据权利要求6所述的方法，其特征在于，所述根据所述第三像素区域中多个第三像素点的像素值，确定所述多个第三像素点的像素差异之前，所述方法还包括：
    按照所述第一像素区域的第一采样步长，从所述第三像素区域中获取与所述多个第一像素点数量相同的第三像素点。
  9. 根据权利要求6所述的方法，其特征在于，所述根据所述第三像素区域中多个第三像素点的像素值，确定所述多个第三像素点的像素差异之后，所述方法还包括：
    当所述多个第三像素点的像素差异小于所述目标差异值时，检测所述第 三像素区域的大小是否大于目标阈值；
    当所述第三像素区域的大小不大于所述目标阈值时，确定大于所述第三像素区域的第四像素区域；及
    基于所述第四像素区域中多个第四像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述多个第四像素点与所述多个第一像素点的数量相同。
  10. 根据权利要求1所述的方法，其特征在于，所述像素差异为所述多个像素点的像素方差或者所述多个像素点的梯度矩阵的最小特征值，所述像素方差用于表示所述多个像素点的像素值相对于像素平均值的变化程度，所述梯度矩阵用于表示所述多个像素点的像素值分别在水平梯度上相对于像素平均值的变化程度和垂直梯度上相对于像素平均值的变化程度。
  11. 根据权利要求1所述的方法，其特征在于，所述第一像素区域和所述第二像素区域的形状为正方形、长方形、圆形、环形、不规则多边形或不规则曲边形中的任一形状。
  12. 根据权利要求1所述的方法，其特征在于，所述基于所述第一图像中所述对象上的目标特征点，确定所述第一图像中包括所述目标特征点的第一像素区域之后，所述方法还包括：
    根据所述第一像素区域的第一采样步长，在所述第一像素区域的区域边界上的多个像素点中，获取所述多个第一像素点；及
    根据所述多个第一像素点的像素值，确定所述多个第一像素点的像素差异。
  13. 一种任务执行方法，由目标设备执行，所述方法包括：
    获取第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
    获取所述第一图像中所述对象上的目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化；及
    基于所述目标特征点的运动信息，执行目标任务。
  14. 根据权利要求13所述的方法，其特征在于，所述目标任务包括路线 规划任务，所述基于所述目标特征点的运动信息，执行目标任务包括：
    当所述目标特征点的数目为多个时，基于多个目标特征点的运动信息，确定距离目标设备不超过第一阈值的至少一个场景对象，所述目标设备为采集所述第一图像和所述第二图像的设备；及
    根据距离所述目标设备不超过第二阈值的目的地的位置和所述至少一个场景对象，确定所述目标设备到达所述目的地的第一目标路线，所述第二阈值大于所述第一阈值。
  15. 根据权利要求13所述的方法，其特征在于，所述目标任务包括对象识别任务，所述基于所述目标特征点的运动信息，执行目标任务包括：
    当所述目标特征点的数目为多个时，基于多个目标特征点的运动信息，确定出所述多个目标特征点中运动信息符合目标条件的多个第一特征点，所述多个第一特征点用于指示所述第一图像和所述第二图像包括的多个对象中的运动对象；及
    基于所述多个第一特征点在所述第一图像或第二图像的位置，确定所述运动对象所属的对象类别。
  16. 一种图像特征点的运动信息确定装置，所述装置包括：
    确定模块，用于确定第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
    所述确定模块，还用于基于所述第一图像中所述对象上的目标特征点，确定所述第一图像中包括所述目标特征点的第一像素区域；
    所述确定模块，还用于根据所述第一像素区域中多个第一像素点的像素差异以及所述目标特征点，确定所述第一图像中包括所述目标特征点的第二像素区域，所述第二像素区域中多个第二像素点的像素差异大于所述多个第一像素点的像素差异，所述多个第二像素点与所述多个第一像素点的数量相同，所述像素差异用于指示多个像素点的像素值的变化程度；及
    获取模块，用于根据所述多个第二像素点和所述第二图像，获取所述目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化。
  17. 一种任务执行装置，所述装置包括：
    获取模块，用于获取第一图像和第二图像，所述第一图像和所述第二图像包括相同的对象；
    所述获取模块，还用于获取所述第一图像中所述对象上的目标特征点的运动信息，所述运动信息用于指示所述目标特征点在所述第一图像和所述第二图像中的位置变化；及
    任务处理模块，用于基于所述目标特征点的运动信息，执行目标任务。
  18. 一种计算机设备，所述计算机设备包括处理器和存储器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行如权利要求1至12中任一项所述的图像特征点的运动信息确定方法的步骤。
  19. 一种目标设备，所述目标设备包括处理器和存储器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行如权利要求13至15中任一项所述的任务执行方法的步骤。
  20. 一种非易失性的计算机可读存储介质，存储有计算机可读指令，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如权利要求1至15中任一项所述方法的步骤。
PCT/CN2020/085000 2019-04-29 2020-04-16 图像特征点的运动信息确定方法、任务执行方法和设备 WO2020221012A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2021529019A JP7305249B2 (ja) 2019-04-29 2020-04-16 画像特徴点の動き情報の決定方法、タスク実行方法およびデバイス
EP20798875.9A EP3965060A4 (en) 2019-04-29 2020-04-16 METHOD FOR DETERMINING MOTION INFORMATION OF AN IMAGE FEATURE POINT, TASK PERFORMANCE METHOD AND APPARATUS
US17/323,740 US20210272294A1 (en) 2019-04-29 2021-05-18 Method and device for determining motion information of image feature point, and task performing method and device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910356752.4A CN110097576B (zh) 2019-04-29 2019-04-29 图像特征点的运动信息确定方法、任务执行方法和设备
CN201910356752.4 2019-04-29

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/323,740 Continuation US20210272294A1 (en) 2019-04-29 2021-05-18 Method and device for determining motion information of image feature point, and task performing method and device

Publications (1)

Publication Number Publication Date
WO2020221012A1 true WO2020221012A1 (zh) 2020-11-05

Family

ID=67446382

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/085000 WO2020221012A1 (zh) 2019-04-29 2020-04-16 图像特征点的运动信息确定方法、任务执行方法和设备

Country Status (5)

Country Link
US (1) US20210272294A1 (zh)
EP (1) EP3965060A4 (zh)
JP (1) JP7305249B2 (zh)
CN (1) CN110097576B (zh)
WO (1) WO2020221012A1 (zh)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110097576B (zh) * 2019-04-29 2022-11-18 腾讯科技（深圳）有限公司 图像特征点的运动信息确定方法、任务执行方法和设备
CN112492249B (zh) * 2019-09-11 2024-04-09 瑞昱半导体股份有限公司 图像处理方法及电路
CN111068323B (zh) * 2019-12-20 2023-08-22 腾讯科技（深圳）有限公司 智能速度检测方法、装置、计算机设备及存储介质
US11756210B2 (en) 2020-03-12 2023-09-12 Adobe Inc. Video inpainting via machine-learning models with motion constraints
US11823357B2 (en) * 2021-03-09 2023-11-21 Adobe Inc. Corrective lighting for video inpainting
CN114119675B (zh) * 2021-11-10 2023-07-18 爱芯元智半导体(上海)有限公司 像素点的偏移获取方法、装置、电子设备及存储介质
CN114396911B (zh) * 2021-12-21 2023-10-31 中汽创智科技有限公司 一种障碍物测距方法、装置、设备及存储介质
CN114748872B (zh) * 2022-06-13 2022-09-02 深圳市乐易网络股份有限公司 一种基于信息融合的游戏渲染更新方法
CN115424353B (zh) * 2022-09-07 2023-05-05 杭银消费金融股份有限公司 基于ai模型的业务用户特征识别方法及系统
CN115330657B (zh) * 2022-10-14 2023-01-31 威海凯思信息科技有限公司 一种海洋探测图像的处理方法、装置和服务器
CN116542970B (zh) * 2023-06-30 2023-09-15 山东金有粮脱皮制粉设备有限公司 基于图像处理的面粉熟化控制方法及相关装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070183667A1 (en) * 2006-01-13 2007-08-09 Paul Wyatt Image processing apparatus
CN102378002A (zh) * 2010-08-25 2012-03-14 无锡中星微电子有限公司 动态调整搜索窗的方法及装置、块匹配方法及装置
CN104700415A (zh) * 2015-03-23 2015-06-10 华中科技大学 一种图像匹配跟踪中匹配模板的选取方法
US20180205926A1 (en) * 2017-01-17 2018-07-19 Seiko Epson Corporation Cleaning of Depth Data by Elimination of Artifacts Caused by Shadows and Parallax
CN109344742A (zh) * 2018-09-14 2019-02-15 腾讯科技（深圳）有限公司 特征点定位方法、装置、存储介质和计算机设备
CN110097576A (zh) * 2019-04-29 2019-08-06 腾讯科技（深圳）有限公司 图像特征点的运动信息确定方法、任务执行方法和设备

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR0171120B1 (ko) * 1995-04-29 1999-03-20 배순훈 특징점 기반 움직임 보상을 이용한 비디오 신호 부호화에서의 움직임 영역 설정방법 및 장치
JPH11218499A (ja) * 1998-02-03 1999-08-10 Hitachi Denshi Ltd 外観検査装置およびその画像処理方法
EP1978432B1 (en) * 2007-04-06 2012-03-21 Honda Motor Co., Ltd. Routing apparatus for autonomous mobile unit
JP4774401B2 (ja) 2007-04-06 2011-09-14 本田技研工業株式会社 自律移動体の経路設定装置
US9307251B2 (en) * 2009-08-19 2016-04-05 Sharp Laboratories Of America, Inc. Methods and systems for determining data-adaptive weights for motion estimation in a video sequence
JP5824936B2 (ja) 2011-07-25 2015-12-02 富士通株式会社 携帯型電子機器、危険報知方法及びプログラム
PL2791865T3 (pl) * 2011-12-12 2016-08-31 Aselsan Elektronik Sanayi Ve Ticaret Anonim Sirketi System i sposób do oszacowania rozmiaru celu
EP3020024B1 (en) * 2013-07-09 2017-04-26 Aselsan Elektronik Sanayi ve Ticaret Anonim Sirketi Method for updating target tracking window size
CN104881881B (zh) * 2014-02-27 2018-04-10 株式会社理光 运动对象表示方法及其装置
US10445862B1 (en) * 2016-01-25 2019-10-15 National Technology & Engineering Solutions Of Sandia, Llc Efficient track-before detect algorithm with minimal prior knowledge
US10002435B2 (en) * 2016-01-29 2018-06-19 Google Llc Detecting motion in images
CN108960012B (zh) * 2017-05-22 2022-04-15 中科创达软件股份有限公司 特征点检测方法、装置及电子设备
EP3435330B1 (en) * 2017-07-24 2021-09-29 Aptiv Technologies Limited Vehicule based method of object tracking
CN107590453B (zh) * 2017-09-04 2019-01-11 腾讯科技（深圳）有限公司 增强现实场景的处理方法、装置及设备、计算机存储介质
CN109523570B (zh) * 2017-09-20 2021-01-22 杭州海康威视数字技术股份有限公司 运动参数计算方法及装置
US11044404B1 (en) * 2018-11-28 2021-06-22 Vulcan Inc. High-precision detection of homogeneous object activity in a sequence of images
US10923084B2 (en) * 2019-03-29 2021-02-16 Intel Corporation Method and system of de-interlacing for image processing

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070183667A1 (en) * 2006-01-13 2007-08-09 Paul Wyatt Image processing apparatus
CN102378002A (zh) * 2010-08-25 2012-03-14 无锡中星微电子有限公司 动态调整搜索窗的方法及装置、块匹配方法及装置
CN104700415A (zh) * 2015-03-23 2015-06-10 华中科技大学 一种图像匹配跟踪中匹配模板的选取方法
US20180205926A1 (en) * 2017-01-17 2018-07-19 Seiko Epson Corporation Cleaning of Depth Data by Elimination of Artifacts Caused by Shadows and Parallax
CN109344742A (zh) * 2018-09-14 2019-02-15 腾讯科技（深圳）有限公司 特征点定位方法、装置、存储介质和计算机设备
CN110097576A (zh) * 2019-04-29 2019-08-06 腾讯科技（深圳）有限公司 图像特征点的运动信息确定方法、任务执行方法和设备

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3965060A4 *

Also Published As

Publication number Publication date
CN110097576B (zh) 2022-11-18
JP2022511427A (ja) 2022-01-31
EP3965060A1 (en) 2022-03-09
EP3965060A4 (en) 2022-10-05
US20210272294A1 (en) 2021-09-02
JP7305249B2 (ja) 2023-07-10
CN110097576A (zh) 2019-08-06

Similar Documents

Publication Publication Date Title
WO2020221012A1 (zh) 图像特征点的运动信息确定方法、任务执行方法和设备
US11205282B2 (en) Relocalization method and apparatus in camera pose tracking process and storage medium
CN110502954B (zh) 视频分析的方法和装置
WO2021008456A1 (zh) 图像处理方法、装置、电子设备及存储介质
KR102635373B1 (ko) 이미지 처리 방법 및 장치, 단말 및 컴퓨터 판독 가능 저장 매체
JP7026819B2 (ja) カメラの位置決め方法および装置、端末並びにコンピュータプログラム
CN111768454B (zh) 位姿确定方法、装置、设备及存储介质
CN109886208B (zh) 物体检测的方法、装置、计算机设备及存储介质
CN110290426B (zh) 展示资源的方法、装置、设备及存储介质
WO2020108041A1 (zh) 耳部关键点检测方法、装置及存储介质
US11386586B2 (en) Method and electronic device for adding virtual item
CN111754386B (zh) 图像区域屏蔽方法、装置、设备及存储介质
CN110570460A (zh) 目标跟踪方法、装置、计算机设备及计算机可读存储介质
CN113763228A (zh) 图像处理方法、装置、电子设备及存储介质
CN112308103B (zh) 生成训练样本的方法和装置
WO2021218926A1 (zh) 图像显示方法、装置和计算机设备
CN111860064A (zh) 基于视频的目标检测方法、装置、设备及存储介质
CN112990424A (zh) 神经网络模型训练的方法和装置
WO2021243955A1 (zh) 主色调提取方法及装置
CN112804481B (zh) 监控点位置的确定方法、装置及计算机存储介质
CN113052408B (zh) 一种社区聚合的方法及装置
CN116502382A (zh) 传感器数据的处理方法、装置、设备及存储介质
CN113486811A (zh) 悬崖检测方法、装置、电子设备及计算机可读存储介质
CN114897915A (zh) 图像分割方法、装置、电子设备和存储介质
CN110929675A (zh) 图像处理方法、装置、计算机设备及计算机可读存储介质

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: 20798875

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021529019

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2020798875

Country of ref document: EP

Effective date: 20211129