stixel world论文总结
1.The Stixel World - A Compact Medium Level Representation of the 3D-World:http://pdfs.semanticscholar.org/2df3/4dbfb4feeb2d7f40e90956ebc8de1f41a5e4.pdf
stixel world开山鼻祖文章
https://zhuanlan.zhihu.com/p/27494151 对stixel world相关的一些概念进行解释
2.Towards a Global Optimal Multi-Layer Stixel Representation of Dense 3D Data：http://www.bmva.org/bmvc/2011/proceedings/paper51/abstract.pdf
多层的stixel world表达，github上有实现
3.Efficient Stixel-based object recognition:https://mydlt.de/david/page/pubs/enzweiler_iv_2012_stixel_class.pdf
根据stixel生成roi
4.Real-time obstacle detection using stereo vision for autonomous ground vehicles: A survey ：http://www.ce.unipr.it/people/smario/papers/itsc2014.surveyOD.pdf
障碍物检测双目方法的一个总结。https://blog.csdn.net/u010665216/article/details/78690225这个博客其实也是根据这篇论文来的。
5.Stixels estimation without depth map computation：http://rodrigob.github.io/documents/2011_iccv_cvvt_workshop_stixels_estimation.pdf Stixels estimation error，这篇文章写了如何评估stixel的准确率
6.Object-Level Priors for Stixel Generation：http://www.visinf.tu-darmstadt.de/media/visinf/vi_papers/2014/cordts-gcpr-preprint.pdf 这个是在multilayer基础上做obstacle优先，看思路能不能借鉴
7.Semantic Stixels: Depth is Not Enough
8.The Stixel World: A Medium-Level Representation of Traffic Scenes：https://arxiv.org/pdf/1704.00280.pdf
stixel跟分割结合
9.Detecting Unexpected Obstacles for Self-Driving Cars: Fusing Deep Learning and Geometric Modeling:https://arxiv.org/pdf/1612.06573.pdf
10.国内的一个人改进stixel:http://www.docin.com/p-1298684890.html
11.Improved Stixel Estimation Based on Transitivity Analysis in Disparity Space
12.Exploiting the Power of Stereo Confidences
stixel的一种改进办法
13.https://d-nb.info/1025886399/34 总结性的pdf，有tracking，有gt的estamation
14.Efficient Representation of Traffic Scenes by Means of Dynamic Stixels：http://www.cvlibs.net/projects/autonomous_vision_survey/literature/Pfeiffer2010IV.pdf 这个是分动态和静态，感觉这个是在stixel-world进行了提升，并且作者也是stixel-world的作者
15.Pixels, Stixels, and Objects：http://pdfs.semanticscholar.org/022e/331d0bea13c88801a465dc3db0cf8c8107dd.pdf
Real-Time Category-Based and General Obstacle Detection for Autonomous Driving自己用的
https://zhuanlan.zhihu.com/p/27494151,这个博客总结到:视差图的精度直接影响计算结果的准确性
通过我的实验,将sgbm换成psmnet,检测出来的准确性的确得到了提升
