Project14 | Photography: We're Giving Away Raspberry Pi HQ Camera Kits for Projects that Use Them!

Hello Everyone, For my entry I will be creating a VEHICLE-X : Real-Time Vehicle Detection, Tracking, Number Plate Detection, and Speed Estimation using Raspberry Pi HQ Camera project

Real-Time Vehicle Detection, Tracking, Speed Estimation, and number plate detection.

As we are going for a walk, escorting our kids to school, or going to our work in the morning, we’ve all experienced unsafe, fast-moving vehicles operated by inattentive drivers that nearly mow us down.

We feel almost powerless. These drivers disregard speed limits, crosswalk areas, school zones, and “children at play” signs altogether. When there is a speed bump, they speed up almost as if they are trying to catch some air!

So here comes the technology of camera in action.

Hello Everyone, my name is Akshita Gupta and for my entry, I will be creating a VEHICLE-X project that will be a real-time vehicle detection, tracking, number plate detection as well as speed estimation.

To build the VEHICLE-X, I will use the following components

• Raspberry Pi 4
• 7" touch display for a proper preview
• Raspberry HQ Camera

Implementation

1. Pi HQ Camera will detect vehicles in video using a MobileNet SSD.
2. Tracks the vehicles
3. Estimates the speed along with the number plate of a vehicle and stores the evidence in the cloud.
4. I will also configure the Raspberry pi with openVINO environment to improve the fps rate.

Advantages-

1. The smart system can help track people, vehicles traveling from Point A to Point B.
2. The solution encompasses a mechanism by which law enforcement agencies can track if there is a deviation from a specific route, the speed of the vehicles can also be detected.
3. This solution will ensure that vehicles do not cross a speed limit to minimize accidents with animals, who cross the road, especially during monsoon.

1. For nighttime use, we may need infrared cameras and infrared lights and/or adjustments to our camera parameters.

This is a small image of the solution which I have performed using a pre-recorded video and have used only Open Cv for it without a pi camera.

I hope you all liked it

Akshita Gupta

Eagerly waiting to create this.....

i have tried out similar application on Raspi 4. with object detection + tracking, the speed will be very slow, slower than 1fps

Have you used GPU? When I tried motion detection using my inbuilt camera it gave 32 fps. Hope so I will make it better.

you mean you use your PC or raspberry pi?

with raspberry pi, even if it has a small gpu, you can't use it for object tracking + detection, i guess no library for non NVidia gpu

If you are facing fps problem then I suggest you to configure your raspberry pi4 with OpenVino Environment. it works in my case

With my application, tracking window movements Window opening monitor with ArUco - Installing OpenCV on the Raspberry PI 4B, if I display the images using QT I I get as much 1.5 fps.

My application does not require using QT or displaying images only for demonstration purposes, so for the final version I will compile OpenCv without QT.

great, will love to see your project

1fps??? Wow. I have a NCS2 I got from hackster last year so I plan on using that to speed some things up, hopefully.

having said that the frame rate does depends of the size of the video. i haven't tried with openvino nor NCS2 so can't comment much. i did try with Xilinx Kria, i would say its speed is good

Thanks for replying.

I'll probably try on 5-10FPS with 480-720 p(it's not like I'll be doing facial recognition)

Yeah try with OpenVino

Yeah try with OpenVino