WO2018010371A1 - 全景图像映射方法 - Google Patents

全景图像映射方法 Download PDF

Info

Publication number
WO2018010371A1
WO2018010371A1 PCT/CN2016/109603 CN2016109603W WO2018010371A1 WO 2018010371 A1 WO2018010371 A1 WO 2018010371A1 CN 2016109603 W CN2016109603 W CN 2016109603W WO 2018010371 A1 WO2018010371 A1 WO 2018010371A1
Authority
WO
WIPO (PCT)
Prior art keywords
point
panoramic image
latitude
square
mapping
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2016/109603
Other languages
English (en)
French (fr)
Inventor
王荣刚
王悦名
王振宇
高文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Peking University Shenzhen Graduate School
Original Assignee
Peking University Shenzhen Graduate School
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 Peking University Shenzhen Graduate School filed Critical Peking University Shenzhen Graduate School
Priority to US16/313,731 priority Critical patent/US10796410B2/en
Publication of WO2018010371A1 publication Critical patent/WO2018010371A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T9/00—Image coding

Definitions

  • the present invention relates to the field of virtual reality (VR), and more particularly to a new panoramic image mapping technique that can be used for a rectangular image of a panoramic image, reducing the code rate required to encode a panoramic image.
  • VR virtual reality
  • mapping The most common mapping method today is rectangular mapping, but this mapping method has more serious oversampling in high latitudes.
  • Academia and industry have improved the way of mapping panoramic images, mapping rectangular mapped panoramic images to some new planes to reduce oversampling in high latitudes, although the new mapping method has a good bit rate. Savings, but due to interpolation during the mapping process, there are some distortion problems.
  • the present invention provides a new panoramic image mapping method, which can be used for a rectangular image (equirectangular), effectively reducing the distortion problem in the mapping process and reducing the code rate required for encoding the panoramic image. Can improve the problem of oversampling in rectangular mapping.
  • the mapping method provided by the invention can reduce the pixel number by 25% in a better setting and reduce the code rate by 11% in the test sequence (if there are more details in the high latitude area, the panorama for the high latitude area The image will save more code rate) and there is basically no distortion.
  • a panoramic image mapping method for a rectangular-mapped panoramic image with a resolution of 2M ⁇ M, the mapping area and the non-mapping area are set according to the latitude, and only the set mapping area is mapped to a square area, thereby effectively reducing oversampling, thereby achieving The purpose of effectively reducing the number of pixels of the panoramic image and the code rate required for encoding in the case of less distortion; including the following steps:
  • the first step for the point on the square area with coordinates (x, y), calculate the vertical distance and the horizontal distance from the point to the center of the square area, and take the larger one as m;
  • Step 2 Calculate the distance from the point to the zeroth point on the concentric square for the point on the square area with the coordinates (x, y), denoted as n; the zeroth point is an arbitrary position Point
  • the third step calculating the longitude latitude and the latitude longitude corresponding to the point of the (x, y) coordinate on the square area according to the n and m;
  • the fourth step according to the longitude latitude and the latitude longitude obtained in the third step, the points of the square region (x, y) on the square region are inversely mapped to the rectangular map corresponding to the panoramic image obtained by Equations 8 and 9.
  • Step 5 When (X, Y) is an integer pixel, assign the pixel value at (X, Y) in the rectangular mapped panoramic image to the (x, y) point on the square region; when (X, Y) When it is not an integer pixel, the pixel is interpolated near (X, Y) to obtain the pixel value of the (x, y) point on the square area.
  • a 2M ⁇ H region of the panoramic image at a high latitude is set as a mapping region, and the mapping region is two; and H is a height at which the mapping region is set, H ⁇ M/2.
  • the H is 0.25M.
  • the panoramic image high latitude range set as the mapping area is preferably between ⁇ 45° and ⁇ 90°.
  • the distance from the point to the zeroth point on the concentric square is calculated in the second step, and the pixel in the upper right corner is selected as the zeroth pixel, which is obtained by the following calculation process:
  • the longitude latitude and the latitude longitude corresponding to the point of the (x, y) coordinate on the square region are calculated by the following formula:
  • the interpolation method in the fifth step is a bicubic interpolation method, a bilinear interpolation method or a nearest neighbor interpolation method.
  • the interpolation method is preferably a bicubic interpolation method.
  • the invention provides a new panoramic image mapping method, which can be used for a rectangular image (equirectangular), effectively reducing the distortion problem in the mapping process and reducing the code rate required for encoding the panoramic image, and improving the existence of the rectangular mapping. Sampling problem.
  • high latitude areas have more detail, and panoramic images for high latitude areas save more code rate and are substantially free of distortion.
  • FIG. 1 is a schematic diagram of a mapping relationship in an embodiment of the present invention.
  • FIG. 2 is a flow chart of a panoramic image mapping method provided by the present invention.
  • FIG. 3 is a schematic diagram of a specific calculation process of a mapping method in an embodiment of the present invention.
  • (a) is a schematic diagram for calculating the vertical distance and horizontal distance from the point of the coordinate (x, y) to the center of the square area, in the figure The vertical distance and horizontal distance to the center respectively;
  • (b) is a schematic diagram for calculating the distance from the point of coordinates (x, y) to the zeroth point on the concentric square (the position of the zero point can be selected at will, we Select the point in the upper left 45° direction as the zeroth point).
  • FIG. 4 is a diagram of a mapping effect according to an embodiment of the present invention.
  • (a) is a panoramic image obtained by a rectangular mapping method; (b) is obtained by using the method provided by the present invention. Panoramic image.
  • the invention provides a new panoramic image mapping method, which can be used for a rectangular image (equirectangular), effectively reducing the distortion problem in the mapping process and reducing the code rate required for encoding the panoramic image, and improving the existence of the rectangular mapping. Sampling problem.
  • the mapping method provided by the invention does not map the region with low latitude in the rectangular mapping panoramic image, and only maps the region with higher latitude in the image to two square regions, thereby effectively reducing oversampling in the high latitude region.
  • the number of pixels of the panoramic image and the code rate required for encoding are effectively reduced.
  • the method of the invention can reduce the number of pixels by 25% and reduce the code rate by 11% in the tested sequence (if there are more details in the high latitude region, the panoramic image for the high latitude region will save more code rate), And basically no distortion.
  • the region with lower latitude is not mapped, and only the region with higher latitude with severe oversampling is mapped, and the latitude of the panoramic image with rectangular mapping is higher.
  • the area is mapped into two N ⁇ N square areas (the area with lower latitude and higher latitude can be set by itself.
  • the area of 2M ⁇ 0.5M between ⁇ 45° is not For mapping, only two 2M ⁇ 0.25M regions between ⁇ 45° and ⁇ 90° are mapped into two 0.5M ⁇ 0.5M square regions, which is better.
  • the area of two 2M ⁇ H (where H is the height of the region with a higher latitude set by H, H ⁇ M/2) at a high latitude in a rectangular-mapped panoramic image with a resolution of 2M ⁇ M can be mapped to two N ⁇ N square area.
  • the relationship of the mapping process of the present invention is as shown in FIG. 1.
  • the low-latitude regions of the rectangular-mapped panoramic image are not mapped, but are mapped to squares at high latitudes.
  • the low latitude area of the rectangular map panoramic image with a resolution of 4096 ⁇ 2048 can be reserved (the 4096 ⁇ 1024 area with low latitude is preserved in this embodiment), and the 4096 ⁇ 512 areas of the upper and lower high latitudes are mapped into two.
  • a square of 1024 ⁇ 1024, the number of pixels after mapping is reduced by 25%.
  • FIG. 2 is a flow chart of a panoramic image mapping method provided by the present invention.
  • the specific calculation of the high latitude area mapping is as follows (the starting value of the subscript in the computer is usually 0, so some of the following formulas are decremented by 1):
  • the first step is to calculate the vertical distance and horizontal distance from the point of the coordinate (x, y) in Fig. 3(a) to the center of the square area, and the larger one is recorded as m, which is calculated as follows:
  • Step 2 Calculate the distance from the point of (x, y) in Figure 3(b) to the zeroth point on the concentric square (the position of the zeroth point can be chosen at random, we select the point in the upper left 45° direction as The zeroth point), the result is recorded as n, and its calculation formula is as follows:
  • Step 3 Calculate the longitude and latitude corresponding to the point of the (x, y) coordinate on the square area according to n and m, and calculate the formula as follows:
  • Step 4 According to latitude and longitude, determine that the coordinates of the square region (x, y) are inversely mapped to the corresponding positions (X, Y) in the rectangular map panoramic image, where:
  • Step 5 If (X, Y) is an integer pixel, assign the pixel value at (X, Y) in the rectangular mapped panoramic image to the (x, y) point on the square region, if (X, Y) is not For a whole pixel, the pixel near (X, Y) is interpolated to obtain the pixel value of the (x, y) point on the square region.
  • the interpolation method used in this embodiment is bicubic interpolation; other difference methods, such as bilinear interpolation, nearest neighbor interpolation, etc., may also be used.
  • the embodiment shows the effect as shown in FIG. 4, wherein (a) is a panoramic image obtained by using a rectangular mapping method; and (b) is a panoramic image obtained by mapping using the method provided by the present invention.
  • this map preserves the low latitude area of the rectangular mapped panoramic image and maps the high latitude area to two square areas. Effectively reduces the area of the image (ie, reduces the number of pixels).

Landscapes

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

Abstract

本发明公布了一种全景图像映射方法，针对分辨率为2M×M的矩形映射全景图像，根据纬度设置映射区域和非映射区域，只将设置的映射区域映射为方形区域包括：对于所述方形区域上的点，计算得到该点到方形区域中心的垂直距离和水平距离，取其中较大者记为m；计算得到该点到同心方形上第零个点的距离n；计算得到该点所对应的经度和纬度；计算得到该点反映射到矩形映射全景图像中所对应的位置(X,Y)；再给该点赋值。本发明方法可有效降低过采样，由此达到在失真较小的情况下有效降低全景图像的像素数和编码所需的码率的目的。

Description

全景图像映射方法 技术领域
本发明涉及虚拟现实(VR)领域，尤其涉及一种新的全景图像映射技术，可用于矩形映射(equirectangular)的全景图像，减少编码全景图像所需的码率。
背景技术
随着虚拟现实技术的日益发展，对虚拟现实视频的需求日益增加。而360°的全景图像由于要求更广的视角，因此相比于传统的平面图像需要更高的分辨率，其编码所需的码率也相较传统的平面图像高出许多。
如今最常见的映射方式是矩形映射，但是这种映射方式在高纬度区域存在较严重的过采样现象。学术界和工业界都对全景图像的映射方式进行了改进，将矩形映射的全景图像映射到一些新的平面上来降低高纬度区域的过采样，虽然新的映射方法在码率上有比较不错的节省，但由于映射过程中的插值，均存在一些失真的问题。
发明内容
为了克服上述现有技术的不足，本发明提供一种新的全景图像映射方法，可用于矩形映射(equirectangular)的全景图像，有效减少映射过程中的失真问题并减少编码全景图像所需的码率，可改善矩形映射中存在的过采样的问题。
本发明提供的映射方法，在较优的设置下可以减少25％的像素数，在测试的序列中减少了11％的码率(高纬度区域如果有较多的细节，针对高纬度区域的全景图像会节省更多的码率)，并且基本没有失真。
本发明提供的技术方案是：
一种全景图像映射方法，针对分辨率为2M×M的矩形映射全景图像，根据纬度设置映射区域和非映射区域，只将设置的映射区域映射为方形区域，从而有效降低过采样，由此达到在失真较小的情况下有效降低全景图像的像素数和编码所需的码率的目的；包括如下步骤：
第一步：对于所述方形区域上的坐标为(x,y)的点，计算得到该点到方形区域中心的垂直距离和水平距离，取其中较大的一个记为m；
第二步：对于所述方形区域上的坐标为(x,y)的点，计算得到该点到同心方形上第零个点的距离，记为n；所述第零个点为任意位置的点；
第三步：根据所述n和m计算得到所述方形区域上的坐标为(x,y)的点所对应的经度latitude和纬度longitude；
第四步：根据第三步得到的经度latitude和纬度longitude，通过式8和式9计算得到所述方形区域上的坐标为(x,y)的点反映射到矩形映射全景图像中所对应的位置(X,Y):
Figure PCTCN2016109603-appb-000001
Figure PCTCN2016109603-appb-000002
第五步：当(X,Y)是整像素时，将矩形映射全景图像中(X,Y)处的像素值赋给所述方形区域上的(x,y)点；当(X,Y)不是整像素时，取(X,Y)附近的而像素进行插值得到所述方形区域上的(x,y)点的像素值。
针对上述全景图像映射方法，进一步地，将全景图像高纬度的2M×H的区域设置为映射区域，所述映射区域为两个；所述H为设置映射区域的高度，H≤M/2。优选地，所述H为0.25M。
针对上述全景图像映射方法，进一步地，所述设置为映射区域的全景图像高纬度范围优选为±45°到±90°之间。
针对上述全景图像映射方法，进一步地，第二步所述计算得到该点到同心方形上第零个点的距离，具体选取右上角的像素为第零个像素，通过如下计算过程得到：
nup＝y-m′              (式2)
nright＝s+x-m′              (式3)
ndown＝2×s+(1024-1-y-m′)              (式4)
nleft＝3×s+(1024-1-x-m′)              (式5)
其中，nup,nright,ndown,nleft分别表示同心方框的上侧、右侧、下侧和左侧的点到第零个点的距离；s＝2×m,为方框上像素数的
Figure PCTCN2016109603-appb-000003
m为步骤一所述方形区域上的坐标为(x,y)的点到方形区域中心的垂直距离和水平距离取其中较大者。
针对上述全景图像映射方法，进一步地，第三步根据所述n和m，具体通过以下公式计算得到所述方形区域上的坐标为(x,y)的点所对应的经度latitude和纬度longitude：
Figure PCTCN2016109603-appb-000004
Figure PCTCN2016109603-appb-000005
其中，4s是点所在的同心方形的像素个数；longitude的取值从-180°到180°，表示西经180°到东经180°；latitude的取指从
Figure PCTCN2016109603-appb-000006
到±90°，正值代表北纬，负值代表南纬。
针对上述全景图像映射方法，进一步地，第五步所述插值方法为双三次插值方法、双线性插值方法或最近邻插值方法。插值方法优选为双三次插值方法。
与现有技术相比，本发明的有益效果是：
本发明提供一种新的全景图像映射方法，可用于矩形映射(equirectangular)的全景图像，有效减少映射过程中的失真问题并减少编码全景图像所需的码率，可改善矩形映射中存在的过采样的问题。一般地，高纬度区域有较多的细节，针对高纬度区域的全景图像会节省更多的码率，并且基本没有失真。
本发明具有以下优点：
(一)降低了纬度较高的区域存在的过采样现象，降低图像的像素数从而降低编码所需要的码率，比如将分辨率为2M×M的矩形映射的纬度较高的两个2M×0.5M区域映射成两个
Figure PCTCN2016109603-appb-000007
的方形区域可节省25％的像素数，在测试序列上平均节省11％的码率。
(二)映射过程中造成的失真很小。
附图说明
图1是本发明实施例中的映射关系示意图。
图2是本发明提供的全景图像映射方法的流程框图。
图3是本发明实施例中映射方法具体计算过程示意图；
其中，(a)为计算坐标为(x,y)的点到方形区域中心的垂直距离和水平距离的示意图，图中
Figure PCTCN2016109603-appb-000008
分别为其到中心的垂直距离和水平距离；(b)为计算坐标为(x,y)的点到同心方形上第零个点的距离的示意图(第零个点的位置可随意选取，我们选择左上45°方向的点作为第零个点)。
图4是本发明实施例的映射效果图；
其中，(a)是采用矩形映射方法得到的全景图像；(b)是利用本发明提供方法映射得到 的全景图像。
具体实施方式
下面结合附图，通过实施例进一步描述本发明，但不以任何方式限制本发明的范围。
本发明提供一种新的全景图像映射方法，可用于矩形映射(equirectangular)的全景图像，有效减少映射过程中的失真问题并减少编码全景图像所需的码率，可改善矩形映射中存在的过采样的问题。
本发明提供的映射方法，对于矩形映射全景图像中纬度较低的区域不进行映射，只将图像中纬度较高的区域映射到两个方形区域上，有效降低了高纬度区域的过采样，在失真较小的情况下有效的降低了全景图像的像素数和编码所需的码率。本发明方法可以减少25％的像素数，在测试的序列中减少了11％的码率(高纬度区域如果有较多的细节，针对高纬度区域的全景图像会节省更多的码率)，并且基本没有失真。
对于分辨率为2M×M的矩形映射全景图像，对其纬度较低的区域不进行映射，只对存在严重过采样现象的纬度较高的区域进行映射，将矩形映射的全景图像中纬度较高的区域的映射成两个N×N的方形区域(纬度较低的区域和纬度较高的区域的范围可自己设置。实验中，我们发现对±45°之间的2M×0.5M的区域不进行映射，只将±45°到±90°之间的两个2M×0.25M的区域进行映射成两个0.5M×0.5M的方形区域，其效果较好)。可将分辨率为2M×M的矩形映射全景图像中的高纬度的两个2M×H(其中H为自己设置的纬度较高的区域的高度，H≤M/2)的区域映射到两个N×N的方形区域。
以下实施例描述了本发明的具体实现过程。本发明映射过程的关系如图1所示，矩形映射的全景图像的低纬度区域不进行映射，而对于高纬度映射成方形。可将分辨率为4096×2048的矩形映射全景图像的低纬度区域保留(本实施例中保留了低纬度的4096×1024的区域)，将上、下高纬度的4096×512的区域映射成两个1024×1024的方形，映射后像素数减少了25％。
图2是本发明提供的全景图像映射方法的流程框图。高纬度区域映射的具体计算如下(在计算机中下标的起始值通常为0，因此下面公式有的进行了减1的操作)：
第一步：计算图3(a)中坐标为(x,y)的点到方形区域中心的垂直距离和水平距离，取其中较大的一个记为m，其计算如下：
Figure PCTCN2016109603-appb-000009
第二步：计算图3(b)中坐标为(x,y)的点到同心方形上第零个点的距离(第零个点的位置可随意选取，我们选择左上45°方向的点作为第零个点)，其结果记为n，其计算公式如下：
nup＝y-m′               (式2)
nright＝s+x-m′               (式3)
ndown＝2×s+(1024-1-y-m′)               (式4)
nleft＝3×s+(1024-1-x-m′)               (式5)
其中，nup,nright,ndown,nleft分别表示同心方框的上侧、右侧、下侧和左侧的点到第零个点的距离；s＝2×m,为方框上像素数的
Figure PCTCN2016109603-appb-000010
第三步：根据n和m计算出方形区域上的坐标为(x,y)的点所对应的经度和纬度，计算公式如下：
Figure PCTCN2016109603-appb-000011
Figure PCTCN2016109603-appb-000012
第四步：根据latitude和longitude确定方形区域上的坐标为(x,y)的点反映射到矩形映射全景图像中所对应的位置(X,Y)，其中：
Figure PCTCN2016109603-appb-000013
Figure PCTCN2016109603-appb-000014
第五步：如果(X,Y)是整像素，则将矩形映射全景图像中(X,Y)处的像素值赋给方形区域上的(x,y)点，如果(X,Y)不是整像素，则取(X,Y)附近的像素进行插值得到方形区域上的(x,y)点的像素值。优选地，本实施例采用的插值方法是双三次插值；也可使用其他的差值方式，如双线性插值，最近邻插值等。
至此完成本映射实施例的所有步骤，实施例展示效果如图4所示，其中，(a)是采用矩形映射方法得到的全景图像；(b)是利用本发明提供方法映射得到的全景图像。从图中可看出，本映射将矩形映射的全景图像的低纬度区域保留，将高纬度区域映射到两个方形的区域， 有效的降低了图像的面积(即降低了像素个数)。
需要注意的是，公布实施例的目的在于帮助进一步理解本发明，但是本领域的技术人员可以理解：在不脱离本发明及所附权利要求的精神和范围内，各种替换和修改都是可能的。因此，本发明不应局限于实施例所公开的内容，本发明要求保护的范围以权利要求书界定的范围为准。

Claims (9)

  1. 一种全景图像映射方法，针对分辨率为2M×M的矩形映射全景图像，根据纬度设置映射区域和非映射区域，只将设置的映射区域映射为方形区域，从而有效降低过采样，由此达到在失真较小的情况下有效降低全景图像的像素数和编码所需的码率的目的；包括如下步骤：
    第一步：对于所述方形区域上的坐标为(x,y)的点，计算得到该点到方形区域中心的垂直距离和水平距离，取其中较大的一个记为m；
    第二步：对于所述方形区域上的坐标为(x,y)的点，计算得到该点到同心方形上第零个点的距离，记为n；所述第零个点为任意位置的点；
    第三步：根据所述n和m计算得到所述方形区域上的坐标为(x,y)的点所对应的经度latitude和纬度longitude；
    第四步：根据第三步得到的经度latitude和纬度longitude，计算所述方形区域上的坐标为(x,y)的点映射到矩形映射全景图像中所对应的位置(X,Y)；
    第五步：当(X,Y)是整像素时，将矩形映射全景图像中(X,Y)处的像素值赋给所述方形区域上的(x,y)点；当(X,Y)不是整像素时，取(X,Y)附近的而像素进行插值得到所述方形区域上的(x,y)点的像素值。
  2. 如权利要求1所述全景图像映射方法，其特征是，将全景图像高纬度的2M×H的区域设置为映射区域，所述映射区域为两个；所述H为设置映射区域的高度，H≤M/2。
  3. 如权利要求2所述全景图像映射方法，其特征是，所述设置为映射区域的全景图像高纬度范围优选为±45°到±90°之间。
  4. 如权利要求2所述全景图像映射方法，其特征是，所述H为0.25M。
  5. 如权利要求1所述全景图像映射方法，其特征是，第二步所述计算得到该点到同心方形上第零个点的距离，具体选取右上角的像素为第零个像素，通过如下计算过程得到：
    nup＝y-m′(式1)nright＝s+x-m′(式2)ndown＝2×s+(1024-1-y-m′)(式3)nleft＝3×s+(1024-1-x-m′)(式4)
    其中，nup,nright,ndown,nleft分别表示同心方框的上侧、右侧、下侧和左侧的点到第零个点的距离；s＝2×m,为方框上像素数的
    Figure PCTCN2016109603-appb-100001
    Figure PCTCN2016109603-appb-100002
    m为步骤一所述方形区域上的坐标为(x,y)的点到方形区域中心的垂直距离和水平距离取其中较大者。
  6. 如权利要求1所述全景图像映射方法，其特征是，第三步根据所述n和m，具体通过以下公式计算得到所述方形区域上的坐标为(x,y)的点所对应的经度latitude和纬度longitude：
    Figure PCTCN2016109603-appb-100003
    Figure PCTCN2016109603-appb-100004
    其中，4s是点所在的同心方形的像素个数；longitude的取值从-180°到180°，表示西经180°到东经180°；latitude的取指从
    Figure PCTCN2016109603-appb-100005
    到±90°，正值代表北纬，负值代表南纬。
  7. 如权利要求1所述全景图像映射方法，其特征是，第四步根据经度latitude和纬度longitude具体通过式7和式8计算方形区域上的坐标为(x,y)的点反映射到矩形映射全景图像中所对应的位置(X,Y)：
    Figure PCTCN2016109603-appb-100006
    Figure PCTCN2016109603-appb-100007
    其中，方形区域上点的坐标为(x,y)；该点反映射到矩形映射全景图像中所对应的位置为(X,Y)；矩形映射全景图像的分辨率为2M×M。
  8. 如权利要求1所述全景图像映射方法，其特征是，第五步所述插值方法包括但不限于双三次插值方法、双线性插值方法或最近邻插值方法。
  9. 如权利要求1所述全景图像映射方法，其特征是，第五步所述插值方法优选为双三次插值方法。
PCT/CN2016/109603 2016-07-12 2016-12-13 全景图像映射方法 Ceased WO2018010371A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/313,731 US10796410B2 (en) 2016-07-12 2016-12-13 Panoramic image mapping method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610543056.0A CN106127691B (zh) 2016-07-12 2016-07-12 全景图像映射方法
CN201610543056.0 2016-07-12

Publications (1)

Publication Number Publication Date
WO2018010371A1 true WO2018010371A1 (zh) 2018-01-18

Family

ID=57283036

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/109603 Ceased WO2018010371A1 (zh) 2016-07-12 2016-12-13 全景图像映射方法

Country Status (3)

Country Link
US (1) US10796410B2 (zh)
CN (1) CN106127691B (zh)
WO (1) WO2018010371A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106127691B (zh) 2016-07-12 2019-04-12 北京大学深圳研究生院 全景图像映射方法
CN108269229B (zh) * 2016-12-30 2021-08-06 央视国际网络无锡有限公司 一种全景图像或视频的像素位置分布优化方法
CN106846245B (zh) * 2017-01-17 2019-08-02 北京大学深圳研究生院 基于主视点的全景视频映射方法
CN106875331B (zh) * 2017-01-19 2019-04-12 北京大学深圳研究生院 一种全景图像的非对称映射方法
CN108566554A (zh) * 2018-05-11 2018-09-21 北京奇艺世纪科技有限公司 一种vr全景视频处理方法、系统及电子设备
CN110956583B (zh) 2018-09-26 2022-05-10 华为技术有限公司 球面图像处理方法、装置及服务器
CN111062875B (zh) * 2019-12-19 2021-11-12 广州启量信息科技有限公司 一种用于空中全景漫游数据的坐标转换方法及装置
CN113411615B (zh) * 2021-06-22 2023-01-10 深圳市大数据研究院 一种面向虚拟现实的纬度自适应的全景图像编码方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6239808B1 (en) * 1998-04-03 2001-05-29 Nvidia Corporation Method and apparatus for determining texture values of graphical images
CN101777193A (zh) * 2010-02-03 2010-07-14 中国人民解放军国防科学技术大学 基于双向映射精确插值的折反射全向图柱面全景展开方法
CN102346923A (zh) * 2010-07-30 2012-02-08 中国科学院遥感应用研究所 一种基于经纬网格的数据分级组织方法
CN104835117A (zh) * 2015-05-11 2015-08-12 合肥工业大学 基于重叠方式的球面全景图生成方法
CN106127691A (zh) * 2016-07-12 2016-11-16 北京大学深圳研究生院 全景图像映射方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH1152845A (ja) * 1997-08-08 1999-02-26 Aisin Aw Co Ltd 地図表示装置及び記録媒体
CN103942754B (zh) * 2013-01-18 2017-07-04 深圳市腾讯计算机系统有限公司 全景图像补全方法及装置
GB2520822B (en) * 2014-10-10 2016-01-13 Aveva Solutions Ltd Image rendering of laser scan data
CN104333675B (zh) * 2014-10-20 2017-09-05 长春理工大学 一种基于球面投影的全景电子稳像方法
CN107409231B (zh) * 2015-03-05 2021-05-11 索尼公司 图像处理装置和图像处理方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6239808B1 (en) * 1998-04-03 2001-05-29 Nvidia Corporation Method and apparatus for determining texture values of graphical images
CN101777193A (zh) * 2010-02-03 2010-07-14 中国人民解放军国防科学技术大学 基于双向映射精确插值的折反射全向图柱面全景展开方法
CN102346923A (zh) * 2010-07-30 2012-02-08 中国科学院遥感应用研究所 一种基于经纬网格的数据分级组织方法
CN104835117A (zh) * 2015-05-11 2015-08-12 合肥工业大学 基于重叠方式的球面全景图生成方法
CN106127691A (zh) * 2016-07-12 2016-11-16 北京大学深圳研究生院 全景图像映射方法

Also Published As

Publication number Publication date
US10796410B2 (en) 2020-10-06
US20200175648A1 (en) 2020-06-04
CN106127691A (zh) 2016-11-16
CN106127691B (zh) 2019-04-12

Similar Documents

Publication Publication Date Title
WO2018010371A1 (zh) 全景图像映射方法
US10798301B2 (en) Panoramic image mapping method
CN102881000B (zh) 一种视频图像的超分辨率方法、装置和设备
CN103426184B (zh) 一种光流跟踪方法和装置
CN106875331B (zh) 一种全景图像的非对称映射方法
CN110458755B (zh) 图像处理方法、相关设备及计算机存储介质
WO2018133380A1 (zh) 基于主视点的全景视频映射方法
CN105931210B (zh) 一种高分辨率图像重建方法
CN109783177B (zh) 一种基于HTML5 Canvas的GIS栅格数据渲染方法
CN106778347B (zh) 一种基于三角函数的矢栅地理数据可逆脱密方法
WO2021073443A1 (zh) 感兴趣区域的检测方法、装置、电子设备及可读存储介质
CN106530313A (zh) 一种基于区域分割的海天线实时检测方法
CN107135397A (zh) 一种全景视频编码方法和装置
CN106780383A (zh) Tof相机的深度图像增强方法
CN108961182A (zh) 针对视频图像的竖直方向灭点检测方法及视频扭正方法
CN101325724B (zh) 一种立体图像对的极线校正方法
CN104103057B (zh) 一种快速的鱼眼图像生成全景图的方法
CN108648205A (zh) 一种亚像素边缘检测方法
CN106920213A (zh) 一种高分辨率图像的获取方法及系统
CN108470363A (zh) 一种基于地理参考点和三角剖分的手绘地图移动定位方法
KR100986607B1 (ko) 영상 보간 방법 및 그 방법이 기록된 컴퓨터로 읽을 수 있는 기록매체
CN109856657B (zh) 一种基于图像变换的路径记录方法
CN108596835B (zh) 一种圆形区域的图像放大方法
CN104182986A (zh) 由正交隐消点的性质求解平面折反射摄像机内参数的方法
CN102930549A (zh) 利用点阵模板和正交性求解摄像机内参数

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: 16908693

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16908693

Country of ref document: EP

Kind code of ref document: A1