long distance objec...
 
Notifications
Clear all

long distance object detection

6 Posts
3 Users
0 Likes
1,531 Views
queenidog
(@queenidog)
Eminent Member
Joined: 3 years ago
Posts: 20
Topic starter  

I want to count the number of vehicles that cross on highway, past my driveway.  I've tried lidar (LED version, they should name it LEDAR) and found the distance is limited.   Tried higher quality SR08 ultrasonic which seemed okay for first 10-15 feet but nothing after that.   I have to span two lanes of traffic+distance to my property or about 25 feet.   So then I got a Doppler radar which seemed to work: even gave me readings of the grass moving in wind ACROSS the road.   So I stuck with this for last couple months, added RTC and SD card for datalogging and now I'm testing it and not happy.   It is too sensitive...Testing in my basement all I do is turn my head and I get a reading, but I don't seem to have a low range.  I expect a value of 0 (volts or A/D bits from 0-1023) if nothing is moving.  Okay, I realize Doppler is for moving objects but I was hoping a vehicle moving in front of unit would disturb a field I could measure.  Am I on the right track?  Maybe a microwave sensor that is NOT Doppler?

I've considered hacking a laser (true laser) tape measure  and bouncing light off a reflector on other side of the road, then back.  My intention is this: detect a vehicle, time how long vehicle is detected and from that determine if it was a motorcycle (short time), car/suv (med time) or bus/big truck (long time).  The laser idea is the best because of simplicity but I worry about sunlight washout, people seeing a laser light and freaking (it's a class II laser, red, bright), and the need for exact alignment.  Microwave seemed the best because it could reflect off body of a vehicle no matter what color, sunlight bright or not. 

Any suggestions?   


   
Quote
Robo Pi
(@robo-pi)
Noble Member
Joined: 4 years ago
Posts: 1909
 
Posted by: @queenidog

I've considered hacking a laser (true laser) tape measure  and bouncing light off a reflector on other side of the road, then back.  My intention is this: detect a vehicle, time how long vehicle is detected and from that determine if it was a motorcycle (short time), car/suv (med time) or bus/big truck (long time).  The laser idea is the best because of simplicity but I worry about sunlight washout, people seeing a laser light and freaking (it's a class II laser, red, bright), and the need for exact alignment.  Microwave seemed the best because it could reflect off body of a vehicle no matter what color, sunlight bright or not. 

I was going to suggest using a reflector on the far side of the road if possible.    You could do this with light.  It wouldn't even necessarily need to be a laser.   There might also  be the possibility of  using a light source on the far side of the road instead of a reflector and just recording when that light source is blocked. 

Posted by: @queenidog

I worry about sunlight washout

If you actually have a light source on the far side of the road you could have it inside a deep tube maybe even with a dark black  circular shield around the front of it.   Then aim your detector at that black shield so that it detects the light in the central tube.   Alternatively you  could use a similar set up with a reflector having the reflector deep within the tube to avoid any unwanted light reflections.

Another option that crossed my mind was to use RF.   Just have a highly focused RF beam aimed at a parabolic reflector on the other side of the road.   These are just ideas offered as food for thought.

I would personally favor the independent reflector or source across the road rather than trying to detect the actual vehicles.  I would simply trust that system to be more accurate.  But with today's sensors measuring the vehicles directly might be doable.   I'm just offering ideas for thought. Nothing more.

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Noble Member
Joined: 4 years ago
Posts: 1654
 
Posted by: @queenidog

I want to count the number of vehicles that cross on highway, past my driveway. 

Any suggestions?   

Vision is one possibility if you have the computing power and sufficient frames per second.

Your issue I assume is you don't have permission to place reflectors and such like in a public space.

 


   
ReplyQuote
queenidog
(@queenidog)
Eminent Member
Joined: 3 years ago
Posts: 20
Topic starter  

I do have a RPI3 with cameras and have seen an app that detects and tracks vehicles.   REally cool but I dislike RPI3 O/S: too hard to use, too many updates and for this project, overkill.

 


   
ReplyQuote
queenidog
(@queenidog)
Eminent Member
Joined: 3 years ago
Posts: 20
Topic starter  

@robo-pi

Thanks robo,   good ideas.  Yes I want simple.  I like your idea of focused light (ie down a tube) to shelter it from sunlight.  I'll have to try my tf-mini lidar again.   I CAN add something across the road...we live in country and there is a ditch there with a post for a sign that I can attach to.   Can't leave a light there though...power issues.  This has to run on battery power.   

Stay tuned.


   
ReplyQuote
robotBuilder
(@robotbuilder)
Noble Member
Joined: 4 years ago
Posts: 1654
 

Video would be a bit of an overkill if you have access to public land to place your devices which is why I asked the question.
Breaking an infrared (invisible) laser beam should give a clean reading?
If you use two beams you should be able to measure the speed, direction and width of the vehicle.
Only complication is if two vehicles are passing each other in opposite directions?


   
ReplyQuote