1. SLAM論文
1.1 SOFT-SLAM:Computationally efficient stereo visual simultaneous localization and mapping for autonomous unmanned aerial vehicles（KITTI 雙目第一名）
1.2 [CVPR2018]CodeSLAM-Learning a Compact, Optimisable Representation for Dense Visual SLAM
1.2.1 Contributions:
l The derivation of a compact and optimisable representation of dense geometry by conditioning a depth auto encoder on intensity images (深度編碼來(lái)表達(dá)稠密幾何結(jié)構(gòu))
l The implementation of the first real-time targeted
monocular system that achieves such a tight joint optimisation of motion and dense geometry.
1.2.2 depth auto encoder result:
1.2.3 Illustration of the SfM system
1.2.4 總結(jié)
在后端優(yōu)化的時(shí)候轮傍，將光度差和重投影誤差一起優(yōu)化，數(shù)據(jù)關(guān)聯(lián)使用文中的共視depth auto encoder
1.3 [CVPR2018]ICE-BA: Incremental, Consistent and Efficient Bundle Adjustment for Visual-Inertial SLAM
1.3.1 Contributions:
l a new sliding window based solver that leverages the incremental nature of SLAM measurements to achieve more than 10x efficiency compared to the state-of-the-arts
l a new relative marginalization algorithm that resolves the conflicts between sliding window marginalization bias and global loop closure constraints
1.3.2 Optimization framework
1.3.3 Relative Marginalization
1.3.4 總結(jié)
本篇文章在local BA的過(guò)程中筹误，選定新的reference keyframe,后面的相機(jī)位姿將相對(duì)于該幀做優(yōu)化浩嫌，這和HoloLens里面選定錨點(diǎn)位置有相似之處蛛淋。
1.4 [CVPR2018] Learning to Find Good Correspondences
1.4.1 Contributions
l being keypoint-based, it generalizes better than image-based dense methods to unseen scenes, which we demonstrate with a single model that outperforms current methods on drastically different indoors and outdoors datasets
l it requires only weak supervision through essential matrices for training
l it can work effectively with very little training data
1.4.2 Result
1.4.3 Network
1.4.4 總結(jié)
提取局部特征點(diǎn)之后俺亮，匹配完扔給神經(jīng)網(wǎng)絡(luò)躲查，它能夠劃分出inlier和outlier酌心，這是一個(gè)分類(lèi)器。
1.5 [CVPR2017]CNN-SLAM: Real-time dense monocular SLAM with learned depth prediction
1.5.1 Contributions
l we illustrate the proposed frameworkfor 3D reconstruction, where CNN-predicted dense depth maps are fused together with depth measurements obtained from direct monocular SLAM
l we show how CNN-predicted semantic segmentation can also be coherently fused with the global reconstruction model
1.5.2 overview
1.5.3 總結(jié)
使用CNN估計(jì)單幅視圖的深度蚤霞，并進(jìn)行語(yǔ)義上的分割酗失，后端優(yōu)化還是傳統(tǒng)的SLAM優(yōu)化方法。
1.6 [CVPR2017]NID-SLAM: Robust Monocular SLAM using Normalised Information Distance
1.6.1 Contributions
l Robust direct tracking using NID
We present a real-time approach for minimising the NID between a candidate image and a key-frame depth map to recover the sim(3) camera pose. In contrast to previous methods we explicitly incorporate depth uncertainty into the NID score
l Multi-resolution tracking using histogram pyramids
We present a novel histogram-pyramid approach for robust coarse-to-fine tracking using NID which increases robustness and the basin of convergence while reducing computation time at smaller scales
l Direct depth map refinement using NID
We present a per-pixel key-frame depth map refinement approach using NID, which allows for map maintenance and depth updates over successive traversals despite appearance changes over time
1.6.2 Pipeline
1.6.3 NIC
1.6.4 總結(jié)
整個(gè)NID-SLAM屬于直接法的一種昧绣，在跟蹤的過(guò)程中规肴，設(shè)計(jì)了上述的NID方法，利用聯(lián)合熵來(lái)表達(dá)相似性，來(lái)代替?zhèn)鹘y(tǒng)直接法中跟蹤部分奏纪。實(shí)驗(yàn)結(jié)果表明鉴嗤，NID-SLAM比ORB-SLAM和LSD的更加魯棒斩启，姿態(tài)估計(jì)效果不相上下序调。
1.7 [TIP2016] Efficient Non-Consecutive Feature Tracking for Structure-from-Motion
1.7.1 Contributions
l Two-Pass Matching for Consecutive Tracking
l Non-Consecutive Track Matching
1.7.2 Feature matching
1.7.3 Matching matrix
1.7.4 總結(jié)
l 在匹配階段，進(jìn)行兩次匹配兔簇，第一次通過(guò)sift匹配獲得少量的較好的匹配結(jié)果发绢，然后通過(guò)RANSAC 算法生成多個(gè)單應(yīng)矩陣，然后通過(guò)單應(yīng)變換將在同一個(gè)平面上的點(diǎn)篩選出來(lái)垄琐，通過(guò)光流做進(jìn)一步篩選边酒，相當(dāng)于將提取出來(lái)的特征點(diǎn)進(jìn)行了多個(gè)平面歸類(lèi)。
通過(guò)連續(xù)幀匹配得到的跟蹤狸窘，對(duì)跟蹤結(jié)果進(jìn)行評(píng)估墩朦，生成matching matrix，它表明了非連續(xù)幀之間的相關(guān)性翻擒，對(duì)相關(guān)性高的幀進(jìn)行匹配氓涣，增加約束。
