Advice on monitoring station build

Thread Starter

Les1951

Joined Apr 6, 2021
6
Hi guys
I am starting on a project to build a small monitoring station it’s purpose is to monitor occupants of passing cars and record the number of people that look at an advertisement, the information needs to be captured by a suitable camera and either stored locally on a storage device or sent directly to a remote storage, then either can then be accessed and analysed.
My first thought was to use Raspberry pi as the main board but there are several alturnative manufacturers out there which may have a more suitable board.
The project would require a camera capable of capturing good quality pictures or video and be able to connect to a cellular network using a 4G SIM card, support a storage device (if needed ) and have a low power requirement so that it can be powered by battery, which could be charged by solar panel.
Can anyone suggest a board that would support as much of these requirements as possible on board and have an easy to use OS which would make any needed add on boards or peripherals as easy as possible to program and get up and running
I would greatly appreciate any advice regarding the most suitable choice of main board and any other advice to help me to get this project going.
Thank you
 

Yaakov

Joined Jan 27, 2019
5,476
How far away will the camera be from the vehicle?
What sort of lighting will be available (will the occupants be in shadow? Will there be direct sunlight? Reflections?)
How do you intend to analyze the images to determine if they are looking at the advertisement?
 

eetech00

Joined Jun 8, 2013
3,087
Hi guys
I am starting on a project to build a small monitoring station it’s purpose is to monitor occupants of passing cars and record the number of people that look at an advertisement, the information needs to be captured by a suitable camera and either stored locally on a storage device or sent directly to a remote storage, then either can then be accessed and analysed.
My first thought was to use Raspberry pi as the main board but there are several alturnative manufacturers out there which may have a more suitable board.
The project would require a camera capable of capturing good quality pictures or video and be able to connect to a cellular network using a 4G SIM card, support a storage device (if needed ) and have a low power requirement so that it can be powered by battery, which could be charged by solar panel.
Can anyone suggest a board that would support as much of these requirements as possible on board and have an easy to use OS which would make any needed add on boards or peripherals as easy as possible to program and get up and running
I would greatly appreciate any advice regarding the most suitable choice of main board and any other advice to help me to get this project going.
Thank you
I would look first at the system that will be using/analyzing the data, what format it requires, then choose the appropriate camera and camera interface with the required remote connectivity.
 

LowQCab

Joined Nov 6, 2012
1,913
I wouldn't do it, in many places Video without consent is illegal.
Are You really going to watch Cars go by for ~12-Hours ???
How will You hide it so that it doesn't get vandalized, especially with a Solar-Panel ???
.
.
.
 

LesM

Joined Dec 20, 2021
2
How far away will the camera be from the vehicle?
What sort of lighting will be available (will the occupants be in shadow? Will there be direct sunlight? Reflections?)
How do you intend to analyze the images to determine if they are looking at the advertisement?
 
Last edited by a moderator:

LesM

Joined Dec 20, 2021
2
How far away will the camera be from the vehicle?
What sort of lighting will be available (will the occupants be in shadow? Will there be direct sunlight? Reflections?)
How do you intend to analyze the images to determine if they are looking at the advertisement?

Hi Yaakov
Thanks for you input
Clearly these are issues that will need to be addressed and will probably be easier once there is a prototype to see how each of the theses problem affects it, the camera will probably need to have a wide view, perhaps with an adjustable lens or if possible self focusing, problems like storage and sending the information hopefully will be quite easily addressed by the type of main board used, many have usb, memory card and SIM card sockets, Wi-Fi and camera connection’s.
Has anyone used or known of boards that work well are straight forward to setup and have these connections on board or can use add-on expansion cards and can use a computer OS to program their use, any first hand advice on boards and add-ons will really help me avoid waisted time.
Thank you in advance for any further suggestions.
 

Yaakov

Joined Jan 27, 2019
5,476
Hi Yaakov
Thanks for you input
Clearly these are issues that will need to be addressed and will probably be easier once there is a prototype to see how each of the theses problem affects it, the camera will probably need to have a wide view, perhaps with an adjustable lens or if possible self focusing, problems like storage and sending the information hopefully will be quite easily addressed by the type of main board used, many have usb, memory card and SIM card sockets, Wi-Fi and camera connection’s.
Has anyone used or known of boards that work well are straight forward to setup and have these connections on board or can use add-on expansion cards and can use a computer OS to program their use, any first hand advice on boards and add-ons will really help me avoid waisted time.
Thank you in advance for any further suggestions.
I asked those questions because they speak to feasibility. The choice of SBC (Single Board Computer) is not the primary issue at the start, rather the possibility of succeeding at the project goals. The questions I asked (and probably some I didn't) are physical constraints that could be show stoppers no matter what you use for SBC.

For example, unless you are very close to the road, a wide angle lens will not provide you with the resolution you need to distinguish the gaze of the occupants. In any case, you will need a very high resolution sensor. If you are farther away, you will need a telephoto lens to get the same coverage, and still have the same resolution issue.

The cameras that are "easy" to use with SBCs are not high resolution. Cameras sufficient for your application are like to be very costly and not "easy" to implement for your project.

As far as lighting is concerned, this is also a physical limitation. Were it my project, I would be out there with some a couple of different camera and lens combinations on site, at different times of day, trying to get photos that would work for the application. A program like The Photographer's Ephemeris available as a web app with a free account, can help you determine the sun angle at different times of day to plan the survey.

Determining what constraints physics places on you for this project is, to me, far more important that the SBC choice at the outset.

Good luck.
 

MisterBill2

Joined Jan 23, 2018
11,556
I predict no success at all, because of the variability of the images of faces with the reflection from the glass windows and the motion of the vehicles. The software will need to have both a frame capture function and also an image analysis function and that will not happen with a small computer package. Consider how many pixels will be needed to detect a face looking, and the need to decide when to look. And we have no hint as to the distance between camera and car.
How fast is the traffic moving as it passes? That will have an effect on the images as well.
In addition to those issues there is probably no way to verify that a given face is looking at the advertisement.

Your best hope is to record the video with high resolution and then have a human evaluate the video after the fact. Use an hour of video taken at the time of day when those targeted by the ad will be passing.
 

MisterBill2

Joined Jan 23, 2018
11,556
The same scheme has been used in grocery stores to determine the effectiveness of displays. But the cameras looked down and saw where people stopped and never saw faces. And from 8 or 15 feet adequate vision is simpler. Of course the publication I was reading was about marketing, not engineering and so I can't vouch for all the details being correct.
Posing this project for analysis would be a good classroom exercise in evaluating feasibility and return on investment. But it really sounds more like a concept created by an MBA than anybody aware of reality.
 
Top