Hey Coding Library, nice to meet you! I just found your channel, love what you're doing! I like how clear and detailed your explanations are as well as the depth of knowledge you have surrounding the topic! Since I run a tech education channel as well, I love to see fellow Content Creators sharing, educating, and inspiring a large global audience. I wish you the best of luck on your UA-cam Journey, can't wait to see you succeed! Your content really stands out and you've put so much thought into your videos, I applaud you on that! Cheers, take care, and keep up the great work ;)
Thank you very much for the nice words, i really appreciate that! 🙂 Cool that your are creating videos aswell and good luck on your journey too. Feel free to join the discord server so we can stay in touch!
Hello, first of all thanks a lot for this video and all the others, you help us so much and present things like no one does, so I want to say that I really appreciate it! Then I have 2 questions to ask you : 1) Is it possible to modify the shape detection in order to detect a square? If yes by modifying what? 2) I did not understand which the color detection parameters were and how to change the color detection range, can you tell me where to look at? Again, thanks a lot, I love what you're doing!!
Hi, thanks for the video. question for you. I have images from a stereo camera. I know the distance between optics. How can I take measures of the features in one couple of images? thanks in advance
Thanks for watching! What u ask for is pretty much what i do in this video here. In this video i only use 1 feature doe, but u can have as many as u want
hello bro, excuse me , i need your help about many cuestions, i make a project for my work and i want use 3 depth cameras but, i dont know so much about that. do you would give me your email or another social media for comunicate Sorry for my bad English
Wow that's the work I was looking for, a real motivation! Keep it up man. I just want to ask you if you can help find the focal length and field of view of my camera ( laptop camera) .Thanks in advance
Thank you very much! I have a video about a computer vision project here in the tutorial. In that video i show how to calculate the focal length and the field of view aswell🙂 that project is about depth estimation and object detection with only 1 camera
@@NicolaiAI Is there any place where I can chat with you , I am nearly finishing a final year project where I use CNN for object detection and then I estimate the distance using triangulation , but I want to ask you something about the camera calibration. Regards, Elio
Will it be possible to use my own yolov8 object detection or do i then need to build it in a segmentation model for the best results? I also need it to track with id's
Yeah you can definitely run your own custom models and then just extract the detections and use that for the depth calculations. Both detection and segmentation can be used and trackers on top of that. Ultralytics also have trackers directly with a few lines of code. Def check out their videos on their UA-cam channel
I now use the center of the segmentation masks to calculate the desparity. is this a good point or is it better to use the center of the detection boxes? what will be most accurate? and does the circle around the center dot have a certain function? I'm not using it at the moment. I mainly use the center point (my objects are pyramid/trapezium shaped)
Hi Nicolai. Which dual camera do you recommend, I am using a RPi 4, and I´m not sure did a little research and I found StereoPi and some Arducam options. Thanks in advance.
I dont calibrate the cameras in this video here, so that file is just empty, but should be the camera parameters and distortion parameters. If you want to learn how to calibrate a camera and get the parameters i have made a video about that aswell. Thank you🙂
According to your formula you can get Z=b*f/Diaparity. But there is a contradiction in dimension. Because Z, b and f are all in millimeter units, and Disparity is in pixel units. How can this be explained?
Hi Nicolai Thanks for the video and the explainations nice work! One question, how should I define the baseline? Is there any best practice or formula I can use? My use case is to create a data set for volume prediction of cups pictured from above and fixed position. So max distance/depth to the table/ cup is always the same.
good morning, thank you for this great tutorial, although i have a question, did you ignore the calibration in this project? cause it worked well although you didnt uncomment the calibrate function, thank you very much for responding asap ☺
No unfortunately u will need two cameras to do stereo vision and depth estimation with that. I have another video here in the computer vision tutorial where i have a project where i do it with one camera like you have, but it is not as precise as this one with two cameras 🙂
@@NicolaiAI Great work man ! I really appreciate your knowledge ! Can you please share the link of the video where you talk about dense stereo vision (detecting distance of multiple objects). Thanks in advance!!😃😃
It should be on automatically, i think it takes some time before it's fully processed. Maybe a couple of hours. I just uploaded it 1 hour ago and just published. Thank you for watching! 🙂
Great tutorial! It working fine. Just wanted to ask a question, to calculate X and Y coordinates, should I use the equations X=ZxL/f and Y = ZyL/f. If so, then the yL= circle_left [1] or it needs to be calculated?
I just use 2 standard webcams, the cameras really depends on the application. If it's for personal projects webcams are fine but in production u can get stereo cameras but they are very expensive. An example could be the OpenCV oak d
Yes u would be able to do that and calculate it. But It would require that u know a distance between two points in world coordinates and then u can use geometry. Then u know the two sides of the triangle and the unknown is the distance between those two points. u can reproject the 2d points to 3d, if u are thinking about doing some project where u want to get the speed of objects. U can do that with monocular camera aswell unless u need to know where the objects are located precisely
Hi, i have been following your videos since a while now. I would like to know how can i extract the shadow and only a shadow from a picture with openCV ? i would like to automatize my umbrella move in order to have the best shadow possible
Join My AI Career Program
www.nicolai-nielsen.com/aicareer
Enroll in My School and Technical Courses
www.nicos-school.com
Hey Coding Library, nice to meet you! I just found your channel, love what you're doing!
I like how clear and detailed your explanations are as well as the depth of knowledge you have surrounding the topic! Since I run a tech education channel as well, I love to see fellow Content Creators sharing, educating, and inspiring a large global audience. I wish you the best of luck on your UA-cam Journey, can't wait to see you succeed! Your content really stands out and you've put so much thought into your videos, I applaud you on that!
Cheers, take care, and keep up the great work ;)
Thank you very much for the nice words, i really appreciate that! 🙂 Cool that your are creating videos aswell and good luck on your journey too. Feel free to join the discord server so we can stay in touch!
@@NicolaiAI No problem, glad I could help out! I'll definitely check out your server too :)
Hello, first of all thanks a lot for this video and all the others, you help us so much and present things like no one does, so I want to say that I really appreciate it!
Then I have 2 questions to ask you :
1) Is it possible to modify the shape detection in order to detect a square? If yes by modifying what?
2) I did not understand which the color detection parameters were and how to change the color detection range, can you tell me where to look at?
Again, thanks a lot, I love what you're doing!!
What if i use different stereo camera between? For example, one is pi camera module3 else is pi camera module 3 wide lens. What do i have to change?
@NicolaiAI can i use MYNT EYE S SDK for this??
How can i calculate de f_pixel if the cameras are different?
Hi, thanks for the video. question for you. I have images from a stereo camera. I know the distance between optics. How can I take measures of the features in one couple of images? thanks in advance
Thanks for watching! What u ask for is pretty much what i do in this video here. In this video i only use 1 feature doe, but u can have as many as u want
I also have a video about depth maps and another one only with the theory of stereo vision that i would recommend!
hello bro, excuse me , i need your help about many cuestions, i make a project for my work and i want use 3 depth cameras but, i dont know so much about that. do you would give me your email or another social media for comunicate
Sorry for my bad English
If u are a member u can find my email under the community section
Wow that's the work I was looking for, a real motivation! Keep it up man. I just want to ask you if you can help find the focal length and field of view of my camera ( laptop camera) .Thanks in advance
Thank you very much! I have a video about a computer vision project here in the tutorial. In that video i show how to calculate the focal length and the field of view aswell🙂 that project is about depth estimation and object detection with only 1 camera
@@NicolaiAI Is there any place where I can chat with you , I am nearly finishing a final year project where I use CNN for object detection and then I estimate the distance using triangulation , but I want to ask you something about the camera calibration. Regards, Elio
I have a discord server in the description :)
@@NicolaiAI Since I have the same problem regarding the focal length, I wanted to ask if you can share the link to this video here. Thanks.
Thanks for your tutorial. The instruction really nice and clear for me. Keep doing well!
Awesome! Im glad u like it
Will it be possible to use my own yolov8 object detection or do i then need to build it in a segmentation model for the best results? I also need it to track with id's
Yeah you can definitely run your own custom models and then just extract the detections and use that for the depth calculations. Both detection and segmentation can be used and trackers on top of that. Ultralytics also have trackers directly with a few lines of code. Def check out their videos on their UA-cam channel
I now use the center of the segmentation masks to calculate the desparity. is this a good point or is it better to use the center of the detection boxes? what will be most accurate? and does the circle around the center dot have a certain function? I'm not using it at the moment. I mainly use the center point (my objects are pyramid/trapezium shaped)
Hi Nicolai. Which dual camera do you recommend, I am using a RPi 4, and I´m not sure did a little research and I found StereoPi and some Arducam options. Thanks in advance.
U can also take a look at econ systems. I have a video about their cameras here on the channel as well
@@NicolaiAI thanks for your answer I will take a look at econ systems
Is the variable f or alpha used in the code? I couldn't find where
Alpha is not used in the code since we already know the focal length, but alpha can be used to calculate the focal length, f
hi, where can i find the "calib_parameters.py" file which you mention in "calibration.py" :( ? Thnx
I dont calibrate the cameras in this video here, so that file is just empty, but should be the camera parameters and distortion parameters. If you want to learn how to calibrate a camera and get the parameters i have made a video about that aswell. Thank you🙂
According to your formula you can get Z=b*f/Diaparity. But there is a contradiction in dimension. Because Z, b and f are all in millimeter units, and Disparity is in pixel units. How can this be explained?
Which camera are you using ? can put the link for so that we can buy it .
I'm just using two cheap standard webcams. If u want a good and fixed setup i would recommend oak d from opencv
Hello @nicolaiAI, you have not used focal length (f) in find_depth function. Why so ?
Hi Nicolai
Thanks for the video and the explainations nice work!
One question, how should I define the baseline? Is there any best practice or formula I can use?
My use case is to create a data set for volume prediction of cups pictured from above and fixed position. So max distance/depth to the table/ cup is always the same.
Sorry, Quick question, is the program accurate for other objects?
good morning, thank you for this great tutorial, although i have a question, did you ignore the calibration in this project? cause it worked well although you didnt uncomment the calibrate function,
thank you very much for responding asap ☺
i have just webcam on my laptop...so can i do this project???/
No unfortunately u will need two cameras to do stereo vision and depth estimation with that. I have another video here in the computer vision tutorial where i have a project where i do it with one camera like you have, but it is not as precise as this one with two cameras 🙂
@@NicolaiAI o thanks for reply bro🙂🙃😑😐
Great Tutorial.. can you please make a video to estimate depth of multiple object.
Thank you for watching! I already have a video about dense stereo vision and deep learning for depth. U can use that for depth to multiple objects
@@NicolaiAI Great work man ! I really appreciate your knowledge ! Can you please share the link of the video where you talk about dense stereo vision (detecting distance of multiple objects). Thanks in advance!!😃😃
hello sir, i just calibrated my stereo cameras but my disparity matrix second and third column is infinite how can i i overcome this problem?
Can you share the code how to calculate the distance of any object?
please implement some deep learning paper related to deep map and view synthesis. thanks
Nice! Thanks for this!
Thanks for watching!
My friend please activate the subtitles.
It should be on automatically, i think it takes some time before it's fully processed. Maybe a couple of hours. I just uploaded it 1 hour ago and just published. Thank you for watching! 🙂
@@NicolaiAI On the contrary thanks to you. This tutorial you put in is a gem
Thank you! I appreciate that
Great tutorial! It working fine. Just wanted to ask a question, to calculate X and Y coordinates, should I use the equations X=ZxL/f and Y = ZyL/f. If so, then the yL= circle_left [1] or it needs to be calculated?
did you ever resolve this ? I have the same question
your camera name????
I just use a budget webcam in this video. So nothing special. U can use whatever u want
Nice! Thanks for this!
Thank you! Any recommendations on what camera to use? Maybe put a link here?
I just use 2 standard webcams, the cameras really depends on the application. If it's for personal projects webcams are fine but in production u can get stereo cameras but they are very expensive. An example could be the OpenCV oak d
Now we got depth, can we use time somehow to obtain speed ?!!!
(1/FPS) gives time taken for each frame and when divided by difference b/w depths of object in successive frames gets us speed right ?.
Yes u would be able to do that and calculate it. But It would require that u know a distance between two points in world coordinates and then u can use geometry. Then u know the two sides of the triangle and the unknown is the distance between those two points. u can reproject the 2d points to 3d, if u are thinking about doing some project where u want to get the speed of objects. U can do that with monocular camera aswell unless u need to know where the objects are located precisely
Hi, i have been following your videos since a while now. I would like to know how can i extract the shadow and only a shadow from a picture with openCV ? i would like to automatize my umbrella move in order to have the best shadow possible