3D Object Detection using YOLO4 | LiDAR Dataset

[@JRFVRSEC]

thanku maam for this useful video, maam i want to classify this object like person, car etc how can i do??

[@man9mj]

Thank you for your content here. What I like about this channel is that it is more about implementation and practice. Thanks for your effort and time.

[@CodeWithAarohi]

Glad you enjoy it!

[@sri-d18]

Hi Ma'am, How can I annotate 3D images, and are there any other models available for predicting defects in 3D images?

[@kratos6619]

Hello ma'am .. can i have the pre trained model please. Need it for a project and i don't have the compute resource to train the model

[@parthasarathyk5476]

Thank you so much mam. I was the first one asked to make a video on 3D object detection.

[@CodeWithAarohi]

Most welcome 😊

[@soheilmajidi9347]

thanks will be training video for custom data ?

[@soravsingla6574]

Code with Aarohi is Best KZbin channel for Artificial Intelligence #CodeWithAarohi

[@vedantpratik9352]

Thanks Mam, Great Video

[@CodeWithAarohi]

Most welcome 😊

[@8eck]

I'm confused. This trained model can work with 2d images and give 3d bounding boxes with their coordinates or that model can work only with Lidar data?

[@teetanrobotics5363]

thank you so much Aarohi for the content. please keep updating the playlists as well

[@CodeWithAarohi]

Sure, I will update the playlists

[@paulosantosbim]

Is there any way to use my Yolov8 fine tuned model for point cloud segmentation ? This model was trained with custom images dataset

[@0GRANATE0]

your camera needs to have lidar? A normal smartphone camera stream (no lidar) is not enough for detection?

[@ramanapraghee3766]

You are actually doing a great job Aarohi👌🏻

[@CodeWithAarohi]

Thankyou

[@gizemce1880]

Hello , l would like to make object detection directly point cloud data , it means l will use camera , is it possible .

[@shprai2496]

Mam I have downloaded torch for CPU and when I'm running the train.py in command prompt it's showing: torch not compiled with cuda enabled What do I do mam

[@LScattabomb]

Also to me

[@utkarshtripathi9118]

Thank u thank u so .much mam for this video 👍👍

[@CodeWithAarohi]

Welcome

[@jamilahmarramduarte1575]

hi mam, how can i use this to self navigate our trash collector robot maam? because were using lidar sensor maam

[@NERELLAAKHIL]

When I try to unzip the above datasets in Google Colab,The RAM is getting finished and runtime disconnection.Can you please tell an alternative

[@CodeWithAarohi]

One alternative approach you can try is to use a cloud storage service such as Google Drive or Dropbox to store the zipped LiDAR dataset, and then mount the cloud storage drive in your Colab notebook. This way, you can directly access the zipped dataset without having to unzip it, and avoid the issues with memory and runtime disconnection. Here are the steps you can follow to mount a cloud storage drive in your Colab notebook: First, upload the zipped LiDAR dataset to your Google Drive or Dropbox account. In your Colab notebook, run the following code snippet to authenticate your Google Drive account: from google.colab import drive drive.mount('/content/drive') Follow the instructions to authenticate your Google account and authorize access to your Google Drive files. Once the authentication is successful, you can navigate to the zipped LiDAR dataset file in your Google Drive folder using the file path: /content/drive/MyDrive/path/to/zipfile.zip You can then read the LiDAR dataset directly from the zipped file, without unzipping it, using libraries like laspy or entwine. Here's an example code snippet to read a LAS file from a zipped dataset using laspy: import laspy import zipfile with zipfile.ZipFile('/content/drive/MyDrive/path/to/zipfile.zip') as myzip: with myzip.open('path/to/lasfile.las') as myfile: las = laspy.read(myfile)

[@shirshakpattnaik2869]

Is it possible to train dataset using yolov7?

[@utkarshtripathi9118]

Mam we want to know difference between YOLO and Detectron2 Means which one is good, which one fast, which one easy to use pls mam 🙏 one video

[@CodeWithAarohi]

Will do after finishing my pipelined videos.

[@utkarshtripathi9118]

@@CodeWithAarohi Thank u so much 🙏🙏 mam..........

[@anuragshrivastava7855]

How u started ur journey, ur experiences in this field if it is possible make a video on it

[@CodeWithAarohi]

Sure, Will do

[@lohit8543]

its based on yolov2 or yolov4?

[@AmanKumarSharma-de7ft]

What's the fps you are getting for the inference and also your system configuration specially the gpu? So the we have an idea of the performance

[@CodeWithAarohi]

GPU: RTX 3090

[@soravsingla8782]

Excellent video. Really amazing

[@CodeWithAarohi]

Thank you very much!

[@aloktiwari5556]

Ma'am after running test.py. my results folder is not created. Please help me with that, and can you please make a video on how to analyze these results

[@mdabuahammedbabu2458]

same here. did you find any answer for that?

[@shefak95]

Did you find a solution for that problem?

[@francescopiomanna4621]

same here, any solution?

[@NERELLAAKHIL]

@CodeWithAarohi I have a query, What do we have inside .bin files in velodyne folder in dataset. How to find the lidar 3D coordinates of detected objects.

[@CodeWithAarohi]

The .bin files in the Velodyne folder of a dataset typically contain raw point cloud data captured by a lidar sensor. Each .bin file represents a single 360-degree scan of the environment, with each point in the point cloud representing a measurement of the distance and angle of a reflective object in the environment.

[@CodeWithAarohi]

To find the lidar 3D coordinates of detected objects, you'll need to perform object detection on the point cloud data. Object detection involves identifying and classifying objects in the point cloud, and then extracting their 3D coordinates.

[@NERELLAAKHIL]

@@CodeWithAarohi Yes mam, But in this case ,first we are mapping the lidar coordinates to 2D coordinates in form of bird eye view and then detecting objects. Therefore,I am not able to extract 3D coordinates of detected objects.Can you please help with it?Also,Can we get the 3D Lidar coordinates using Voxelnet

[@ayush88]

@@NERELLAAKHIL Hi, did you try with voxelnet ?

[@madhurgarg4114]

How can I do this in my local CPU ?

[@devbgms7032]

Great work , could have an idea regarding on custom data creation in kitti format , is it possible , i hope you can answer this. thank you.

[@ThingswithSid]

cant see image_2 folder

[@MuhammadAdnan-tq3fx]

is it any way to convert the 2D data set into 3D dataset. i want to perform 3D object detection on my dataset?

[@CodeWithAarohi]

To convert a 2D dataset into a 3D dataset for object detection, you can use stereo imaging to estimate depth from multiple angles, employ photogrammetry techniques, or integrate depth information from sensors like LiDAR. Advanced AI methods like monocular depth estimation can also be used to infer depth from single 2D images.

[@Mr.Esmaily]

Thank you so much 👍

[@CodeWithAarohi]

You are welcome

[@ramanapraghee3766]

You are actually doing a great work 👌🏻. Also you can explain this architecture of this model it would be helpful

[@CodeWithAarohi]

Sure I will do a video on architecture.

[@ramanapraghee3766]

@@CodeWithAarohi expecting as next video

[@ShubhamTiwari-ks2qg]

@@CodeWithAarohi when are you uploading video maam ?

[@AmitKumarMallickmes]

Madam, How to do it on the custom dataset? Please make a video on the custom dataset. Thank You

[@CodeWithAarohi]

Will try

[@seratmostaqueem2329]

hi ,great work but can I try it with just a part of data not all data ,I mean i use about 1000 image for image _2 and velodyne but i have error is that the reason ? the error is BrokenPipeError: [Errno 32] Broken pipe

[@ShubhamTiwari-ks2qg]

maam when I was training(local computer) it got stopped in between. how to let it train? did You used high computational machine? please maam help me out .....Its my major project

[@CodeWithAarohi]

I have used RTX 3090. You can reduce the batch size and then try.

[@ShubhamTiwari-ks2qg]

@@CodeWithAarohi what according to you should be good one?

[@akhilnerella5491]

Hello Mam, Can this YOLo4 be used for real time i.e. Lidar mounted car that can detect objects in real time . Also there are two types of images in results.Can you please explain them(point cloud image and image).Which one is the lidar detected image?

[@akhilnerella5491]

Is that just for visualisation?

[@akhilnerella5491]

Also,Here in the paper where is the LiDar mounted as we are getting the bird eye map

[@nakkashravani2277]

I still work in this type of projects..

[@abhisheksumeettoppo5208]

Ma'am can you upload the trained model for this project

[@tanudewangan4310]

Hello, if I want to do 2D object detection from the KITTI dataset then what files I need to download.

[@zeynepakbalk7002]

same question ?

[@kwameampofo3595]

Can you upload a video on how to use custom dataset?

[@CodeWithAarohi]

Sure

[@kwameampofo3595]

@@CodeWithAarohi Thank you

[@rafaelparzeller2974]

@@CodeWithAarohi would also love to see a tutorial for training with a custom dataset :)

[@sabahatrizvi7260]

Great tutorial. Can we do this for custom dataset? If yes, so can you tell what different steps would need to be done as part of training?

[@mrmeow1]

yes. is your dataset annotated?

[@abby.studios]

@@mrmeow1 how to annotate such a 3d File?

[@athletester4472]

@@abby.studios 3d Bat

[@pawankumardoddamani1493]

Hello mam, I have a query could you please come up with a video mam from scratch coding for the yolo v5 object detection instead of using github link for yolo v5 installation?

[@CodeWithAarohi]

I will try

[@Kishi1969]

Great Job....Can you give us the link where we can get the dataset please ?

[@CodeWithAarohi]

www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d

[@comp-it1913]

I have my own lidar point clouds dataset in text format but not labelled. Can you please let me know how to cluster and track objects through JPDA and DBSCAN algorithms in Matlab. can you make a tutorial on it .

[@CodeWithAarohi]

HI, I work with python. Never worked with matlab.

[@comp-it1913]

@@CodeWithAarohi can you make a video on that in python

[@ruwang3132]

Do you have any progress on this point? DBSCAN works? or there are some other methods?

[@ruwang3132]

@Code With Aarohi @COMP-IT Do you have any progress on this point? DBSCAN works? or there are some other methods?

[@ChinnuChinnu-l6f]

Great. It would be better if you could share the result too.

[@CodeWithAarohi]

Ok next time

[@chittaranjanjena8749]

Hello, please can you provide any sensor fusion model on Lidar poaints and 2d-images? any model for object detection using lidar pcd in YOLOv5 or YOLOV7 also help full..

[@sahand9238]

6:04 You forgot to mention we have to download the images too.

[@CodeWithAarohi]

Oh Sorry for that

[@subhanzaheer5138]

which link is for images .. can you please tell me

[@abiralaashish8798]

can you share the .pth file?

[@gehadzakria7849]

Thanks, but the version of torch==1.5.0 no exit

[@CodeWithAarohi]

Install python 3.6.8 and use older version on pip then you can use torch 1.5.0

[@gehadzakria7849]

@@CodeWithAarohi Thanks I will do that

[@Jai_Lopez]

great job........but can you say KITTY one more time ;)

[@rajuvarma-zk2zw]

Thank you madam

[@CodeWithAarohi]

You’re welcome 😊

[@bhanureddy2087]

Settings up the environment is the hard thing and you ignored that

[@CodeWithAarohi]

I agree and I apologize for the inconvenience. You can mail me at [e-mail] I will send you the requirements.txt there

[@maheshkumar_2017]

hi aarohi can we communicate?

[@CodeWithAarohi]

You can mail me at [e-mail]

[@lipa7415]

i get this error "torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate MiB" i tried train 1000 img IS it my 1660 ti is to weak batch 16 epoch 30? pls help

[@CodeWithAarohi]

Try working with batch size 8

[@muhammadsuhail8814]

great work mam do you have a student whatsapp group kindly send the link to contact easily and I mail you mam please check

[@CodeWithAarohi]

No, I don't have any such group.