WO2018054173A1 - 裸眼3d显示系统中可硬件实现的浮点型视点处理方法 - Google Patents

裸眼3d显示系统中可硬件实现的浮点型视点处理方法 Download PDF

Info

Publication number
WO2018054173A1
WO2018054173A1 PCT/CN2017/095502 CN2017095502W WO2018054173A1 WO 2018054173 A1 WO2018054173 A1 WO 2018054173A1 CN 2017095502 W CN2017095502 W CN 2017095502W WO 2018054173 A1 WO2018054173 A1 WO 2018054173A1
Authority
WO
WIPO (PCT)
Prior art keywords
viewpoint
floating point
sub
pixel
row
Prior art date
Application number
PCT/CN2017/095502
Other languages
English (en)
French (fr)
Inventor
秦剑飞
姜志强
韩大强
Original Assignee
四川长虹电器股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 四川长虹电器股份有限公司 filed Critical 四川长虹电器股份有限公司
Publication of WO2018054173A1 publication Critical patent/WO2018054173A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N13/00Stereoscopic video systems; Multi-view video systems; Details thereof
    • H04N13/10Processing, recording or transmission of stereoscopic or multi-view image signals
    • H04N13/106Processing image signals
    • H04N13/122Improving the 3D impression of stereoscopic images by modifying image signal contents, e.g. by filtering or adding monoscopic depth cues

Definitions

  • the present invention relates to a naked eye 3D display technology, and in particular to a hardware-implementable floating point view processing method in a naked eye 3D display system.
  • the naked eye processing schemes in the existing market can be divided into integer type and floating point type, the integer type scheme loses most of the signals, and the screen display effect is poor; the floating point type is complicated due to the complexity of the algorithm implementation and the calculation amount is too large. Only a small part of the market is used for hardware implementation, and there is no way to implement the entire hardware.
  • the technical problem to be solved by the present invention is to provide a floating point type viewpoint processing method which can be implemented by hardware in a naked eye 3D display system, aiming at efficient and high precision floating in the case of hardware achievable for the defects of the prior art. Point processing and error compensation to meet the display effect of different naked-eye 3D screens when the grating tilt angle and tilt distance change.
  • a hardware-implementable floating-point viewpoint processing method in a naked-eye 3D display system includes the following steps:
  • step A the method of calculating the viewpoint value of the first sub-pixel of each row is:
  • N M out /P*V view , and reserve the eight-bit floating point for the obtained viewpoint value; wherein N is the viewpoint value, and M out is the physical resolution of the display screen and the tilt of the raster The calculation constant of the moving relationship, P is the grating physical parameter, and V view is the number of viewpoints.
  • step C the integer parts of N of the sub-pixel positions of each line are sequentially arranged and estimated according to the law of cyclic reciprocation.
  • step D the method for estimating the floating point weight is:
  • the floating point is increased by the sub-pixel position, and the step is decremented to zero, and then incremented from zero to the value of the first sub-pixel of the row, and so on, until the last sub-item of the row.
  • the floating point weight estimation of the pixel location is completed.
  • step E the error caused by the viewpoint estimation is compensated by the reverse order and the same distribution law, and specifically includes:
  • the floating point portion to be discarded in addition to the reserved eight-bit floating point number in step A is defined as F, and the F of the first sub-pixel is distributed in the same order as the average distribution in step D. , the sub-pixel floating point weight of the same position is compensated, and the final floating point part is obtained as N F .
  • step F the method for calculating the final viewpoint value according to the estimated viewpoint integer part and the compensated floating point weight part is:
  • N final N Z * N F + (N Z +1) * (1 - N F ).
  • the invention has the beneficial effects that the efficiency of the algorithm is improved, and the floating-point naked-eye 3D display system can be realized in low-cost hardware, thereby saving cost.
  • the naked-eye 3D algorithm in the present invention starts from the input of picture data of the HDMI or LVDS path, and each step is implemented in hardware.
  • the DDR3 is used to store data to realize inter-frame switching, the picture recombination and the image quality post-processing algorithm are all implemented by hardware.
  • the most difficult of these is the hardware implementation of the floating-point naked-eye 3D screen and the algorithm for intercepting certain precision floating-point data processing, and the optimization of the algorithm for the error compensation hardware after the corresponding floating-point processing.
  • the specific steps are as follows: :
  • x represents the horizontal coordinate position of the input picture
  • y represents the vertical coordinate position of the input picture.
  • N is the viewpoint value
  • M out is the calculation constant of the physical resolution of the display screen and the upward and downward movement relationship of the raster tilt
  • P is the physical parameter of the grating
  • V view is the viewpoint. Number.
  • the viewpoint value N calculated by the first sub-pixel of each row is rounded off to obtain an integer part thereof, that is, an integer value N Z obtained by subtracting N from the floating-point portion;
  • N final N Z * N F + (N Z +1) * (1-N F ).
  • the method of comparing the software to calculate the viewpoint of the invention has a slight error, but combined with the visual characteristics of the human eye, the error is controlled within the range acceptable to the human eye, and the system architecture which can be implemented by hardware is used to improve the efficiency of the algorithm. , the cost of expensive naked-eye 3D display solutions is greatly reduced.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
  • Image Processing (AREA)

Abstract

本发明涉及裸眼3D显示技术,其公开了一种裸眼3D显示系统中可硬件实现的浮点型视点处理方法,旨在硬件可实现的情况下高效并且高精度的进行浮点处理及误差补偿,以满足不同裸眼3D屏在光栅倾斜角度和倾斜距离变化时的显示效果。在本发明中先精确计算每一行第一个子像素的视点值,再对其余子像素的视点值的整数部分和浮点部分分别进行排列和估算,并对估算的结果进行补偿,在估算和补偿时采用循环平均分布规律,最后以子像素视点值的整数部分和补偿后的浮点部分计算最终视点值。

Description

裸眼3D显示系统中可硬件实现的浮点型视点处理方法 技术领域
本发明涉及裸眼3D显示技术,具体涉及裸眼3D显示系统中可硬件实现的浮点型视点处理方法。
背景技术
近年来,普通消费者对于裸眼3D电视需求逐年提高,因此裸眼3D电视成本在不断地减小的同时,也开始出现各种实现方案不同版本的裸眼3D电视系统。从全SOC系统到全GPU系统,现在比较多的是全GPU系统实现。为了减少成本,应用A8芯片加简单FPGA控制的方法,但得到的画面显示效果不是很理想,成本减少也远远小于期望值。
现有市场上的裸眼处理方案可分为整型和浮点型,整型方案丢掉大部分信号,画面显示效果很差;浮点型的由于算法实现的复杂性和计算量太大的原因,市面上只有把其中很小一部分用于硬件实现,还没有整套硬件实现的方法。
如果把整套裸眼3D系统从RGB信号进入后的全部流程全部用硬件实现,裸眼3D电视机的成本将大大降低,普通家庭购买裸眼3D电视将成为可能。
发明内容
本发明所要解决的技术问题是:针对现有技术的缺陷,提出一种裸眼3D显示系统中可硬件实现的浮点型视点处理方法,旨在硬件可实现的情况下高效并且高精度的进行浮点处理及误差补偿,以满足不同裸眼3D屏在光栅倾斜角度和倾斜距离变化时的显示效果。
本发明解决其技术问题所采用的技术方案是:
裸眼3D显示系统中可硬件实现的浮点型视点处理方法,包括以下步骤:
A、根据裸眼3D屏的参数,对图片每一行的第一个子像素的视点值进行计算;
B、对每一行的第一个子像素的视点值N进行舍去处理得到其整数部分;
C、对每一行的子像素位置的N的整数部分进行估算;
D、对每一行的子像素位置的N的浮点部分进行平均分布排列并进行浮点权重估算;
E、对视点估算带来的误差进行逆序同分布规律的补偿;
F、根据估算的视点整数部分和补偿后的浮点权重部分计算最终视点值。
作为进一步优化,步骤A中,对每一行的第一个子像素的视点值计算的方法为:
按照求视点公式:N=Mout/P*Vview计算出视点值,对求取的视点值保留八位浮点;其中,N为视点值,Mout为显示屏物理分辨率与光栅倾斜上下移动关系的计算常数,P为光栅物理参数,Vview为视点的个数。
作为进一步优化,步骤C中,对每一行的子像素位置的N的整数部分依次排列并按照循环往复的规律进行估算。
作为进一步优化,步骤D中,所述浮点权重估算的方法为:
从每一行的第二个子像素开始估算,浮点按子像素位置增加,而步长递减到零,再从零递增到此行第一个子像素的值,如此循环,直到此行最后一个子像素位置的浮点权重估算完成。
作为进一步优化,步骤E中,所述对视点估算带来的误差进行逆序同分布规律的补偿,具体包括:
将步骤A中求取的视点值除保留的八位浮点数外需要舍去的浮点部分定义为F,将第一个子像素的F按照与步骤D中的平均分布方式逆序同分布的方式,对同位置的子像素浮点权重进行补偿,得到最终的浮点部分为NF
作为进一步优化,步骤F中,所述根据估算的视点整数部分和补偿后的浮点权重部分计算最终视点值的方法为:
通过公式N=NZ*NF+(NZ+1)*(1-NF)计算得出最终的上屏多视点融合的视点值。
本发明的有益效果是:提高了算法的效率,可以低成本硬件实现浮点型裸眼3D显示系统,节约成本。
具体实施方式
本发明中的裸眼3D算法,从HDMI或者LVDS通路的图片数据输入进来开始,每一个步骤都硬件实现,利用DDR3存储数据实现帧间切换,图片的重组及画质后处理算法,全部硬件实现。这其中难度最大的就是浮点型裸眼3D屏的整型处理和截取一定精度浮点数据处理的算法的硬件实现,以及针对相应的浮点处理后的误差补偿硬件实现算法的优化,具体步骤如下:
1、对子像素(x,y)精确视点值的计算:
这里x代表输入图片横向坐标位置,y代表输入图片纵向坐标位置。根据祼眼3D屏的参数(光栅倾斜角度、光栅贴片倾斜距离),对输入图片每一行的每一个RGB数据进行视点计算,按照求视点的公式:N(x,y)=Mout/P*Vview计算出视点值,它必定是一个小数, 其小数部分即视点权重取精度1/256,即保留八位浮点数;
这个计算包含乘、除等复杂运算,硬件实现方法占用资源多,但本发明方法只计算每一行第一个子像素,其它绝大部分子像素不需要计算,直接结果就是使不可能硬件实现的裸眼3D算法变得可以低成本硬件实现了;上式中,N为视点值,Mout为显示屏物理分辨率与光栅倾斜上下移动关系的计算常数,P为光栅物理参数,Vview为视点的个数。
2、对每一行第一个子像素计算得到的视点值N进行舍去处理得到其整数部分,即N减去N的浮点部分得到的整数值NZ
3、对每一行的子像素位置的N的整数部分进行估算,比如:若该画面是采用8个视点融合而成的,则按照{1,2,3,4,5,6,7,8,1,2,3,4,5,6,7,8……}的方式排列估算;
4、对每一行的子像素位置的N的浮点部分进行排列,规律按将此行第一个子像素计算得到N的浮点部分进行平均分布排列;
5、对每一行的子像素位置的N的浮点视点权重的计算:从每一行第二个子像素开始估算,浮点按子像素位置增加而步长递减到零,再从零递增到此行第一个子像素的值,如此循环,直到此行最后一个子像素位置估算完成;
6、对估算的视点值进行补偿:将步骤1中计算的N值除保留的八位浮点数外,将超过8位,需要舍去的浮点部分值定义为F,具体补偿时,将第一个子像素的F按照与步骤D中的平均分布方式逆序同分布的方式,对同位置的子像素浮点权重进行补偿,得到最终的浮点部分为NF
7、通过公式N=NZ*NF+(NZ+1)*(1-NF)计算得出最终的上屏多视点融合的视点值。
本发明对比软件计算视点的方法,得到结果会有稍许的误差,但是结合人眼的视觉特征,把误差控制在人眼能接受的范围内,采用可硬件实现的系统构架,提高了算法的效率,让昂贵的裸眼3D显示方案成本得到很大程度的降低。

Claims (6)

  1. 裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,包括以下步骤:
    A、根据裸眼3D屏的参数,对图片每一行的第一个子像素的视点值进行计算;
    B、对每一行的第一个子像素的视点值N进行舍去处理得到其整数部分;
    C、对每一行的子像素位置的N的整数部分进行估算;
    D、对每一行的子像素位置的N的浮点部分进行平均分布排列并进行浮点权重估算;
    E、对视点估算带来的误差进行逆序同分布规律的补偿;
    F、根据估算的视点整数部分和补偿后的浮点权重部分计算最终视点值。
  2. 如权利要求1所述的裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,步骤A中,对每一行的第一个子像素的视点值计算的方法为:
    按照求视点公式:N=Mout/P*Vview计算出视点值,对求取的视点值保留八位浮点;其中,N为视点值,Mout为显示屏物理分辨率与光栅倾斜上下移动关系的计算常数,P为光栅物理参数,Vview为视点的个数。
  3. 如权利要求2所述的裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,步骤C中,对每一行的子像素位置的N的整数部分依次排列并按照循环往复的规律进行估算。
  4. 如权利要求3所述的裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,步骤D中,所述浮点权重估算的方法为:
    从每一行的第二个子像素开始估算,浮点按子像素位置增加,而步长递减到零,再从零递增到此行第一个子像素的值,如此循环,直到此行最后一个子像素位置的浮点权重估算完成。
  5. 如权利要求4所述的裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,步骤E中,所述对视点估算带来的误差进行逆序同分布规律的补偿,具体包括:
    将步骤A中求取的视点值除保留的八位浮点数外需要舍去的浮点部分定义为F,将第一个子像素的F按照与步骤D中的平均分布方式逆序同分布的方式,
    对同位置的子像素浮点权重进行补偿,得到最终的浮点部分为NF
  6. 如权利要求5所述的裸眼3D显示系统中可硬件实现的浮点型视点处理方法,其特征在于,步骤F中,所述根据估算的视点整数部分和补偿后的浮点权重部分计算最终视点值的方法为:
    通过公式N=NZ*NF+(NZ+1)*(1-NF)计算得出最终的上屏多视点融合的视点值。
PCT/CN2017/095502 2016-09-20 2017-08-01 裸眼3d显示系统中可硬件实现的浮点型视点处理方法 WO2018054173A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610833722.4 2016-09-20
CN201610833722.4A CN106331678B (zh) 2016-09-20 2016-09-20 裸眼3d显示系统中可硬件实现的浮点型视点处理方法

Publications (1)

Publication Number Publication Date
WO2018054173A1 true WO2018054173A1 (zh) 2018-03-29

Family

ID=57787462

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/095502 WO2018054173A1 (zh) 2016-09-20 2017-08-01 裸眼3d显示系统中可硬件实现的浮点型视点处理方法

Country Status (2)

Country Link
CN (1) CN106331678B (zh)
WO (1) WO2018054173A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106331678B (zh) * 2016-09-20 2019-04-30 四川长虹电器股份有限公司 裸眼3d显示系统中可硬件实现的浮点型视点处理方法
CN107483912B (zh) * 2017-08-31 2019-01-25 成都斯斐德科技有限公司 一种基于浮点型柱状透镜光栅的多视点图像融合方法
CN111104091B (zh) * 2019-12-12 2021-11-26 北京科技大学 一种动态浮点误差分析中精度特定计算的检测和转换方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130265640A1 (en) * 2012-04-04 2013-10-10 JVC Kenwood Corporation Naked-eye stereoscopic display apparatus
CN104796685A (zh) * 2015-03-31 2015-07-22 王子强 一种普适的自由立体显示播放系统
CN104811686A (zh) * 2015-04-14 2015-07-29 西安交通大学 一种浮点型多视点裸眼立体合成图像的硬件实现方法
CN105338336A (zh) * 2015-10-28 2016-02-17 四川长虹电器股份有限公司 基于柱镜光栅与lcd屏幕的立体图像快速合成方法
CN106331678A (zh) * 2016-09-20 2017-01-11 四川长虹电器股份有限公司 裸眼3d显示系统中可硬件实现的浮点型视点处理方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708577B (zh) * 2012-06-15 2016-05-11 中航华东光电有限公司 多视点立体图片的合成方法
CN103813153B (zh) * 2014-01-27 2015-12-30 万象三维视觉科技(北京)有限公司 一种基于加权求和的裸眼3d多视点图像合成方法
CN105898284A (zh) * 2015-11-24 2016-08-24 徐州维林苑文化传媒有限公司 一种判定连续视点无镜立体光栅参数优良性的方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130265640A1 (en) * 2012-04-04 2013-10-10 JVC Kenwood Corporation Naked-eye stereoscopic display apparatus
CN104796685A (zh) * 2015-03-31 2015-07-22 王子强 一种普适的自由立体显示播放系统
CN104811686A (zh) * 2015-04-14 2015-07-29 西安交通大学 一种浮点型多视点裸眼立体合成图像的硬件实现方法
CN105338336A (zh) * 2015-10-28 2016-02-17 四川长虹电器股份有限公司 基于柱镜光栅与lcd屏幕的立体图像快速合成方法
CN106331678A (zh) * 2016-09-20 2017-01-11 四川长虹电器股份有限公司 裸眼3d显示系统中可硬件实现的浮点型视点处理方法

Also Published As

Publication number Publication date
CN106331678A (zh) 2017-01-11
CN106331678B (zh) 2019-04-30

Similar Documents

Publication Publication Date Title
US9407896B2 (en) Multi-view synthesis in real-time with fallback to 2D from 3D to reduce flicker in low or unstable stereo-matching image regions
US9241147B2 (en) External depth map transformation method for conversion of two-dimensional images to stereoscopic images
CN102572486B (zh) 立体视频的采集系统及方法
WO2018054173A1 (zh) 裸眼3d显示系统中可硬件实现的浮点型视点处理方法
CN108174182A (zh) 三维跟踪式裸眼立体显示视区调整方法及显示系统
JP2009077234A (ja) 三次元画像処理装置、方法及びプログラム
US20120218393A1 (en) Generating 3D multi-view interweaved image(s) from stereoscopic pairs
US10462337B2 (en) Liquid crystal display device and image processing method for same
CN111869202B (zh) 用于减少自动立体显示器上的串扰的方法
US20130293533A1 (en) Image processing apparatus and image processing method
JP2014110568A (ja) 画像処理装置、画像処理方法、およびプログラム
CN101895778B (zh) 减轻立体影像鬼影的方法及系统
US10368048B2 (en) Method for the representation of a three-dimensional scene on an auto-stereoscopic monitor
CN104811686A (zh) 一种浮点型多视点裸眼立体合成图像的硬件实现方法
US10672110B2 (en) Information processing system, information processing apparatus, output apparatus, program, and recording medium
CN102497525B (zh) 一种运动补偿去隔行方法
KR20130051709A (ko) 스테레오 카메라 모듈
CN105338336B (zh) 基于柱镜光栅与lcd屏幕的立体图像快速合成方法
CN103002230B (zh) 适应性像素补偿方法
TWI387354B (zh) 色彩飽和度調整方法以及使用該方法的影像顯示系統
TWI478100B (zh) 影像深度估計方法及其裝置
KR20170098235A (ko) 자동 입체영상 디스플레이들의 크로스토크를 감소시키기 위한 방법, 장치, 및 시스템
US10930185B2 (en) Information processing system, information processing apparatus, output apparatus, program, and recording medium
TWI556623B (zh) 三維影像控制方法
CN105354797A (zh) 一种基于l1-l2罚函数的深度图超分辨率重建方法

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: 17852243

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 30.07.2019)

122 Ep: pct application non-entry in european phase

Ref document number: 17852243

Country of ref document: EP

Kind code of ref document: A1