本文介绍了Android 4.3/4.4上的Google PhotoSphere背后的算法的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

Android 4.3/4.4上的球体模式"下的相机应用程序可以将来自不同方向的照片拼接成一个球面全景,质量非常好.

The camera app on Android 4.3/4.4 under the 'Sphere mode' can stitch photos from varied directions into one spherical panorama, with very good quality.

结果说明:

我要做的是为iPad应用程序实现完全相同的功能.

What I'm going to do is to implement an application for iPad with exactly the same functionality.

我尝试了两种方法:

OpenCV 2.4.8提供了一个图像拼接管道,乍一看似乎很有希望.但是,它有两个问题:

OpenCV 2.4.8 provides with an image stitching pipeline which seems very promising at first glance. However, it has two problems:

  1. 该管道处理无序输入的照片,据我所知,它仅使用图像特征匹配来定位照片之间的几何关系,并且当空白照片(例如,白色)上的图像特征提取失败时,该管道的性能很差墙,蓝蓝的天空).但是,iPad可以提供每张拍摄照片的任意空间方向数据(尽管有噪点),但我不知道如何在OpenCV缝合管线中利用这些数据.

  1. The pipeline deals with unordered input photos, as far as I know it only uses image feature matching to locate the geometric relations between photos, and the pipeline performs poorly when image feature extraction fails on blank photos (eg. white wall, blue sky). However, iPad can provide the arbitrary spatial direction data of each taken photo (with noise though), but I don't know how to utilize these data in the OpenCV stitching pipeline.

我找不到像PhotoSphere使用此管道那样生成180x360度全景图的方法.向管道输送全方位的照片只会产生带有弯曲图像边界的大全景图.

I couldn't find a way to generate a 180x360 degree panorama like PhotoSphere does using this pipeline. Feeding the pipeline with photos in all directions only produces a large panorama with curved image boundary.

PanoTools

我还尝试了PanoTools/Hugin库,尽管该库支持预定义的照片方向,但结果非常差且不稳定.另外,我尝试使用 Focal(beta)应用程序在Google Play上,其范围模式也是基于PanoTools/Hugin的,其结果(在球形模式下)似乎并不比我们的好.

PanoTools

I also tried the PanoTools/Hugin lib, although this lib supports predefined photo directions, the result is quite poor and unstable. Additionally I tried using the Focal(beta) app on Google Play, whose sphere mode is also based on PanoTools/Hugin, their results (in sphere mode) seems no better than ours.

我的问题:

  1. 有没有可以做Google PhotoSphere的库?由于我不从事商业用途的开发,因此任何开源库都是可以接受的.
  2. 有没有好的算法(或技术报告,学术论文)?如果有,哪一个是最好的?
  3. Google如何实现自己的PhotoSphere?任何建议表示赞赏.

推荐答案

您可以在此处查看Google的Camera应用程序源代码:

You can take a look at Google's source code for Camera application here :

https://android.googlesource.com/platform/packages/apps/Camera/+ refs

这篇关于Android 4.3/4.4上的Google PhotoSphere背后的算法的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持!

08-06 19:16