Instant Object Detection in Lidar Point Clouds
傳感器：VLP-64
主要工作：完整的激光雷達語義環境感知方案，通過將點雲投影到地面柵格圖，按照點的密度和最高高度、最大高度差，將點雲分為建築物表面、低矮物體、較高物體，而后按照點雲的密度變化將點雲進行聚類，再講點雲按照最大面積的方向，投影為90*90的深度圖，通過CNN進行分類，最后一步是按照環境的語義信息進行分類結果的糾正，修正點雲聚類等造成的錯誤。
數據集為：需要聯系作者獲取密碼方可下載
Pedestrian Recognition Using High-definition LIDAR 發表於2011的IEEE會議論文
文章主要是在三維點雲中進行行人檢測,主要關鍵部分有以下兩點:
- 特征提取:提出了Z方向的分塊點雲密度作為特征,將點雲塊的反射強度作為特征算子的輸入值,考慮了反射強度的信息,對於較遠的目標,集合信息較弱的情況下,效果提升明顯.
- 將之前提取的特征向量作為輸入,采用SVM算法進行分類,文章的效果提升明顯.
The slice feature and the distribution of the reflection intensities are proposed to improve the recognition performance at a long range with low spatial resolution. The slice feature can represent the rough profile of a pedestrian shape effi- ciently and the distribution of the reflection intensities is effective to discriminate the materials of the targets. The quantitative evaluation using real 3D range data confirms that the proposed method achieves higher performance than the Navarro-Serment method. Moreover, the proposed features can improve the classification ability at a range of more than 30 m
智能駕駛中點雲目標快速檢測與跟蹤
本文主要是集中在目標的跟蹤方面,核心過程如下:
- 一是點雲的快速目標檢測,多線雷達的數據處理依靠單線激光雷達的數據跳躍點來描述目標,在極坐標中表示,再投影到地面來進行聚類,這樣,能夠去除相鄰目標的粘連問題.
- 二是在數據關聯時,使用形狀相似和運動模式相似兩張概率來計算其概率值.
- 三是在模型的運動預測過程中,用橢圓表示行人和自行車,用長方形來表示車輛,確定其運動方向.
文章的技術關鍵在於 目標檢測與目標跟蹤的相輔相成,目標檢測輸入前一幀數據的預測結果,在其鄰域內查找目標,從而確定候選,同時,也優化了重疊等導致的信息不全和變化, 另一方面,模型檢測得到的信息,通過對其運動模式的判定,可以提升其數據關聯的准確率.
LIDAR BASED DETECTION AND CLASSIFICATION OF PEDESTRIANS AND VEHICLES USING MACHINE LEARNING METHODS
作者選取了五個特征,從今年的文獻中,主要包含了box的三個方向的特征,幾何體積,和基於反射強度的特征.最終的測試通過三種方法來進行測試,分別是支持向量機,神經網絡和決策樹算法.
In the end, we selected five features found in literature which were the following.The eigenvalues of the covariance matrix for each cluster in x, y, and z directions, the total volume of the object, and the variance in intensity of the cluster.
最終的誤差,作者主要介紹了兩項,一項是低於0.9的閾值導致的誤差,二是點雲分割過程中的閾值設計,過大,會導致多個目標的粘連,過小會導致目標被分割為幾部分.
Towards 3D Object Recognition via Classification of Arbitrary Object Tracks
Improving Vehicle Detection in Point Cloud Data with Novel Features
主要的思路是地面濾除及分割聚類、選取特征、SVM的分類和最后的特征相關性分析
Vehicle and Pedestrian Recognition Using Multilayer Lidar based on Support Vector Machine
特征選取
除此之外,考慮到自行車在正面和行人的相似性,使用了速度作為區分行人和自行車的特征.
Application of Random Forest Algorithm to Classify Vehicles Detected by a Multiple Inductive Loop System
基於隨機森林的基於門禁電磁感應探測回波信號的目標識別系統。
Automatic Vehicle Classification using Roadside LiDAR Data
2019年新發布的論文，主要工作是通過路邊緣的固定激光雷達，進行道路上車輛和行人等的識別，核心流程為：
作者對目前提出的點雲檢測識別特征的Review：
作者實驗測試了四種流行的分類方法：NB朴素貝葉斯、KNN k nearest neighbours 、 Random Forest 、SVM. 最終結論如下：
This research built a comprehensive database including point clouds, manually marked objects, and corresponding pictures, which is open to all researchers for vehicle classification methods development or improvement. Different vehicle classification methods were investigated using the database built in this study. RF has the highest overall accuracy among the investigated methods. The testing results also indicated that the distance between objects and roadside LiDAR can influence the performance of vehicle classification.