Pedestrian Detection using Lidar point cloud which is self labelled using lidar labeller app and is in my workspace

7 次查看（过去 30 天）
gaurav
gaurav 2024-2-6
回答： Jaimin 2024-11-6，10:18
I have lidar labelled data in my matlab workspace and now i want to perform Pedestrian detection algorithm on it but the data is very unstructured no code is working on it so in which fomat should the data be and how to perform the task. i am providing the properties of lidar data which i laelled in lidar labeller app and is in my workspace.
gTruth =
groundTruthLidar with properties:
DataSource: [1×1 vision.labeler.loading.PointCloudSequenceSource]
LabelDefinitions: [1×5 table]
LabelData: [500×1 timetable]
>> gTruth.DataSource ans =
PointCloudSequenceSource with properties: Name: "Point Cloud Sequence" Description: "A PointCloud sequence reader" SourceName: "C:\Users\asus\Desktop\Lidar\Lidar" SourceParams: [1×1 struct] SignalName: "Lidar" SignalType: PointCloud Timestamp: {[500×1 duration]} NumSignals: 1 >> gTruth.LabelDefinitions ans = 1×5 table Name Type LabelColor Group Description ______________ ______ ________________________ ________ ___________ {'Pedestrian'} Cuboid {[0.5862 0.8276 0.3103]} {'None'} {0×0 char} >> gTruth.LabelData ans = 500×1 timetable Time Pedestrian ___________ ____________ 00:00.00000 {3×9 double} 00:01.00000 {3×9 double} 00:02.00000 {4×9 double} 00:03.00000 {4×9 double} 00:04.00000 {4×9 double} 00:05.00000 {4×9 double} 00:06.00000 {4×9 double} : : 08:13.00000 {4×9 double} 08:14.00000 {4×9 double} 08:15.00000 {4×9 double} 08:16.00000 {4×9 double} 08:17.00000 {4×9 double} 08:18.00000 {4×9 double} 08:19.00000 {4×9 double} this is the info about my data so what should be the variable related to lidarpointclouds

请先登录，再回答此问题。

回答（1 个）

Jaimin
Jaimin 2024-11-6，10:18
Hi @gaurav
Given the properties of your “gTruth” object, you have a ground truth dataset with labels specified for pedestrians. Here's how you can proceed:
  • Convert a Timetable to Point Cloud Objects, extract each point cloud frame and convert it into a “pointCloud” object using the “pointCloud” constructor in MATLAB.
  • Extract and format the pedestrian labels from the timetable to ensure they are compatible with your chosen detection algorithm, such as using bounding boxes.
  • Select an appropriate deep learning model to train the data.
For more information about “pointCloud” kindly refer following MathWorks Documentation.
https://www.mathworks.com/help/vision/ref/pointcloud.html
I hope this will be helpful.

类别

在 Help Center 和 File Exchange 中查找有关 Labeling, Segmentation, and Detection 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by