Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[MD Corie]

I'm not sure whether this belongs in Security or Automation, but here goes:

The second biggest problem I have with outdoor motion sensors (the biggest being improper triggering) is trying to align the sensor's effective field of view with the area that I actually want it to cover, both in terms of depth and width.  So far, the only way I know of is to use the trial and error method, making various adjustments to the orientation of the motion sensor and/or to any added masks or shields.  This process is full of uncertainty of results, and is quite tedious - and sometimes never-ending.

So, my question is:  Is there any more effective, more definite/reliable, and less time-consuming, method for getting a motion sensor "aligned" to the desired detection area - so that it reacts properly to any motion in the desired area, while ignoring all motion outside of that area?

Also, what is the best/most reliable way to induce triggering in order to test the alignment?  (I've simply been wandering around in the desired area, and this does not seem to produce consistent results in any way, shape, or form, so I'm wondering whether there is something -maybe a light or reflector?- that I should use in order to get definite triggering results).

I sure would appreciate any/all helpful hints, because I'm thoroughly frustrated by my attempts to date.

Thanks in advance!

[Brian H]

The X10 motion sensors use a Passive Infrared Detector to detect motion. By a change in detected heat in predetermined time frame.
From the X10 wiki on automation sensors.
http://kbase.x10.com/wiki/What_is_Passive_Infrared_Detection
http://kbase.x10.com/wiki/Optimizing_Motion_Sensor_Detection

If you are using security sensors there is a section in the wiki for those also. They also use a PIR detector.
http://kbase.x10.com/wiki/Security_Motion_Sensor

[MD Corie]

The X10 motion sensors use a Passive Infrared Detector to detect motion. By a change in detected heat in predetermined time frame.
From the X10 wiki on automation sensors.
http://kbase.x10.com/wiki/What_is_Passive_Infrared_Detection
http://kbase.x10.com/wiki/Optimizing_Motion_Sensor_Detection

If you are using security sensors there is a section in the wiki for those also. They also use a PIR detector.
http://kbase.x10.com/wiki/Security_Motion_Sensor

Thanks, but those deal with performance issues (which I also have here - but that's another whole rat's nest).  In this post, I'm only asking about effective ways to determine/adjust/limit the area that a motion sensor will react to.  The two ways that I am aware of to affect the "viewed" area are to alter the orientation of the motion sensor module, and to strategically "mask" portions of the sensor in order to shield it from motion that might occur in surrounding area(s) that are not of interest.  The problem here is the extremely tedious and unreliable nature of attempting to do these adjustments via trial and error, using myself as the "target" to be detected.  So, I'm hoping there is some more systematic procedure, as well as some more reliable way to trigger the sensor during testing in order to accurately define the boundaries of the viewed area.  Any suggestions?

[everydayflyer]

Issue is that the sensitivity and range(field) will change as the temperature and conditions change. The greater the difference between ambient temp. and body temp. the greater the range and sensitivity. Also concrete (driveways),buildings,grass and other vegetation will heat at different rates depending on exposure to sun and wind.

[MD Corie]

Issue is that the sensitivity and range(field) will change as the temperature and conditions change. The greater the difference between ambient temp. and body temp. the greater the range and sensitivity. Also concrete (driveways),buildings,grass and other vegetation will heat at different rates depending on exposure to sun and wind.

That's quite true, and I presume that "alignment" would have to be done at some "nominal" time... but here, I'm not worried so much about performance as I am about simply trying to match the sensors to their desired coverage area - in other words, to "adjust" each sensor's field of view so that it can react only to movement within its desired area(s).  This need is partly to eliminate triggers from motion that occurs in nearby areas that are not of interest, and partly to avoid "overlap" of adjacent sensors (in other words, so that only one sensor will react to movement in any given area).

Now, I suppose macros might be used cleverly to enhance the detemination of what part of a common area that motion is occurring in via some sort of "triangulation" - but I think that might require even more precise determination of each sensor's actual field of view (?)

Anyway, the need here is to figure out how to restrict the detections to only motion in specific areas, mainly because the motion sensors' field of "view" is wider than the field of view of the associated camera.

[dave w]

Now, I suppose macros might be used cleverly to enhance the detemination of what part of a common area that motion is occurring in via some sort of "triangulation" - but I think that might require even more precise determination of each sensor's actual field of view (?)

Anyway, the need here is to figure out how to restrict the detections to only motion in specific areas, mainly because the motion sensors' field of "view" is wider than the field of view of the associated camera.

There is no magic formula. It is trial and error. Mounting high and pointing downward helps control the field of vision. However, as everydayflyer pointed out; what is true today will change tomorrow. What is true during the day will change at night. It is a five buck PIR detector with range parameters that will change daily.

[beelocks]

It is a five buck PIR detector with range parameters that will change daily.

They're stable for that long? 

[dave w]

It is a five buck PIR detector with range parameters that will change daily.

They're stable for that long? 

Yes, I know, sometimes I tend to be overly optomistic. 

[MD Corie]

Now, I suppose macros might be used cleverly to enhance the detemination of what part of a common area that motion is occurring in via some sort of "triangulation" - but I think that might require even more precise determination of each sensor's actual field of view (?)

Anyway, the need here is to figure out how to restrict the detections to only motion in specific areas, mainly because the motion sensors' field of "view" is wider than the field of view of the associated camera.

There is no magic formula. It is trial and error. Mounting high and pointing downward helps control the field of vision. However, as everydayflyer pointed out; what is true today will change tomorrow. What is true during the day will change at night. It is a five buck PIR detector with range parameters that will change daily.

I'm sadly well aware of the "instability" of the performance parameters of these sensors, but just to be sure I'm on the same page here, does this instability also affect the area that the sensor will respond to?  (I mean, other than adverse effects on the range component of the covered area, will it significantly alter the width aspect of the field where a sensor will detect motion, too?)

What I'm trying to do here is to "calibrate" the motion sensors so they will "see" (generally) only the same area as their associated cameras can view.  The problem is that the sensors now pick up motion in the adjoining sensors' "areas", too - and so it's hard to discern which camera to turn on.

I'm assuming the only way to establish these zones is to use trial and error to some extent, but I'm hoping there is some way to definitely trigger a sensor from a given spot in its field of view.  So far, it seems the "warm body" approach (i.e. - me) does not give very reliable triggering, so it's hard to establish the current boundaries of a sensor that way.  So, I was hoping there might be some "tool" or technique that would give more reliable triggering.  (I've tried off-the-wall stuff like shining a flashlight at the sensor, and even using a driveway reflector to "flash" the sensor, but nothing really seems to work consistently).

[dave w]

 

[MD Corie]

Gee, thanks for the "help".    I take it you're claiming the motion sensors are junk, and not only does their performance change over time with changing environmental conditions, but their behavior also changes at random from minute to minute?  If so, then there's no hope of using them to control cameras to monitor any real time events?  In that case, I guess my question has to become one of what sort of sensor can be used to detect intruders reliably in the viewed area and trigger cameras to observe them in a timely manner?

[beelocks]

Trying to align something you can't see with something else that you can't see is going to be very difficult at best (mindless optimism, it's actually just on the wrong side of impossible).

The X10 motion sensors are about twice as vague as you've found them to be - often I cannot get a reliable read with a ***eye sensor across a standard doorway under fairly stable light and heat conditions. Once you start mounting them outside, subject to light and temperature changes you might as well just give up, stand outside and watch what's going on for yourself instead of trying to do it electronically.

I gave up on camera systems using motions sensors a long time ago.
If you only want to detect motion within the view of the camera, then you might like to look at a 'real' surveillance system that uses pixel changes from the camera to start events.

My opinions are based on experience and do not necessarily reflect those of others on this forum who will likely disagree with everything I say (depending upon prevailing ambient light/temperature conditions) 

I know there are many here who use the motion sensors reliably, but you'll probably find that their requirements tend to be a little more vague than the exact critical sensing conditions that you're trying to achieve .

[MD Corie]

Trying to align something you can't see with something else that you can't see is going to be very difficult at best...
...
I know there are many here who use the motion sensors reliably, but you'll probably find that their requirements tend to be a little more vague than the exact critical sensing conditions that you're trying to achieve .

My own experience confirms that this scenario is far from ideal... unfortunately, it is what I have, and my only real options are to either give up entirely, or try to wring whatever functionality I can out of what I have.

Just to clarify the picture a bit, I'm not using ****Eye motion sensors, but rather the Floodlight motion sensors (model PR511, IIRC).  I don't know how the types compare, nor whether these sensors have the same characteristics as the ****Eye types, but I do know they differ operationally in a couple of ways.

By the way, I have also used three different EagleEye sensors indoors (although they are supposedly made for outdoor use).  They generally work fairly well for controlling area lights (at least until their batteries start getting low, then all bets are off), but they are still noticably less "positive" in their responses when compared to non-X10 motion-sensing bulb sockets that I have.

Anyway, back to the issue of triggering surveillance cameras with the floodlight sensors that I have, I'm guessing a "tool" that would be helpful when trying to determine the areas "covered" by the motion sensors would be something that I could use that would positively trigger the sensor from a given spot, regardless of any nuances of performance variations in the sensor.  (In other words, something that would overcome any borderline sensitivity issues due to environment or whatever, and no fooling ensure a yes/no detection from a given spot in the area of interest).

What I'm hoping to accomplish is to positively map the edges of the sensor's horizontal (side-to-side) field of view - which I am assuming to be more of a "mechanical" limit and thus unaffected by performance variations due to current environmental factors.  (If this assumption is invalid, then the whole exercise is moot).

The end result that I seek is to be able to tell which camera to turn on when any given sensor triggers - as opposed to the vaguery that occurs when a sensor can detect areas covered by more than one camera, and thus it becomes impossible to determine which camera to turn on.  Again, my assumption is that it should be viable to mask or orient the sensors so that they also can "see" only an approximation of the 60-degree field of view that the cameras see - and will ignore motion that occurs outside of such regions.

To me, this is a different problem than what has been discussed here - that is, my interest is in limiting detections to only certain areas, rather than addressing the concern about whether or not detections will actually occur within such areas, depending on environmental factors, phase of the moon, or whatever...  In other words, I'm not trying to deal with ensuring that "real" detections occur within the view of the sensors, but rather to limiting the area that they can react to.  Unfortunately, this still presents the need to ensure detections in order to map the boundaries... hence my original question of whether there are any good tools or techniques that would facilitate this boundary mapping, given the variability of the sensors' performance.  My thought was that some readily-detectable IR source could be moved from the sides of the detection area towards the middle until the sensor triggered, thus determining the edge of view.  Unfortunately, I have not come up with any IR source that seems to produce any consistent results... so either my trigger source is not adequate, or else my procedure is bogus somehow...  so I'm hoping someone may have some better ideas.

[dave w]

Gee, thanks for the "help".    I take it you're claiming the motion sensors are junk, and not only does their performance change over time with changing environmental conditions, but their behavior also changes at random from minute to minute?  If so, then there's no hope of using them to control cameras to monitor any real time events?  In that case, I guess my question has to become one of what sort of sensor can be used to detect intruders reliably in the viewed area and trigger cameras to observe them in a timely manner?

No. The detectors are not junk, neither the xxxEyes, nor the floodlight PIRs. But what you want to do with PIR motion detectors isn't possible, and that has been pointed out several times in this thread.  Did you bother to read the Wiki sites you were provided earlier in this thread?  One of the sites gives typical field of view for X10 PIRs and can be overlays for your cameras...BUT...

The PIRs depend upon seeing a MOVING heat differential. THIS INCLUDES EDGE OF FIELD DETECTION. A warm body moving into the field of view will be detected sooner on a cool day than on a warm day.  The edge of field detection is dynamic, and heavily influenced by ambient temperature, the size of target, amount of sunlight (background IR), etc. What you want to do:  find precises and dependable, repeatable edge of field detection points using PIR motion detectors, just isn't possible.

What you want is a video switcher/multiplexer which will output a trigger when movement is detected in the video frame.

[MD Corie]

Gee, thanks for the "help".    I take it you're claiming the motion sensors are junk, and not only does their performance change over time with changing environmental conditions, but their behavior also changes at random from minute to minute?  If so, then there's no hope of using them to control cameras to monitor any real time events?  In that case, I guess my question has to become one of what sort of sensor can be used to detect intruders reliably in the viewed area and trigger cameras to observe them in a timely manner?

No. The detectors are not junk, neither the xxxEyes, nor the floodlight PIRs. But what you want to do with PIR motion detectors isn't possible, and that has been pointed out several times in this thread.  Did you bother to read the Wiki sites you were provided earlier in this thread?  One of the sites gives typical field of view for X10 PIRs and can be overlays for your cameras...BUT...

The PIRs depend upon seeing a MOVING heat differential. THIS INCLUDES EDGE OF FIELD DETECTION. A warm body moving into the field of view will be detected sooner on a cool day than on a warm day.  The edge of field detection is dynamic, and heavily influenced by ambient temperature, the size of target, amount of sunlight (background IR), etc. What you want to do:  find precises and dependable, repeatable edge of field detection points using PIR motion detectors, just isn't possible.

What you want is a video switcher/multiplexer which will output a trigger when movement is detected in the video frame.

If I could afford one, I'd be interested in trying out a video detector.  But for my application, it does not appear that it would be practical.

Anyway, for some reason, I cannot seem to get across what I really want to do with these sensors (as opposed to what people seem to insist that I want to do).  Yes, I bothered to read the wikis... but I wonder if anyone bothered to read what I am actually trying to accomplish;  it sure doesn't seem that way.   For example, I don't necessarily care what the normal field of view of a sensor is, if I'm trying to mask it in order to be consistent with the field of view of its associated camera;  instead, I need to find out where the sensor is "seeing" when I apply the masking adjustments.  My assumption is that the sensor physically cannot see beyond the masks, regardless of what the evironmental conditions may be doing to the sensor.  I also assume that a strong enough mapping IR source ought to overcome any sensitivity issues with the detector and trigger the sensor when it moves from behind the mask into the non-masked area.  If these assumptions are incorrect, then I surely don't understand why.

I don't pretend to have any great expertise with these sensors - not IR sensors in general, for that matter - but things are just not making sense.  For instance, what consequence is there in regards to mapping the boundary even if the performance/sensitivity of the sensor might vary over time as the ambient conditions change?  I don't see why long-term variances would be an issue for short-term mapping.

It also sounds to me like "precision" that I am "requiring" is being over-estimated:  I really only need to segregate the detection areas into about 6-8 "quadrants" around the building, and be able to tell which quadrant motion has been detected in - without getting triggers on two or more sensors for any given motion.  For example, if motion is occuring to the northwest of the building, I need it to trigger only the northwest sensor, and not the north, northwest and west sensors.  So, even if the actual boundaries are "sloppy" by a few degrees left or right, it should be possible to separate the fields of view adequately to get distinct detections, even if it may mean that "dead" areas might develop between the sensors' fields of view under various environmental conditions.  It's not my expectation that one sensor's field will end at 45 degrees and the next will pick up at 46 degrees, although there are a couple of places where areas would need to get relatively "tight".  Rather, my expectation would be to restrict the sensors to "view" an area limited to about 50-60 degrees, and roughly in alignment with the 60 degree field of the associated camera - instead of allowing the sensor to trigger on motion that occurs both within and outside of the 60 degree field of the camera, and especially not within the field of a different camera.  I'm just finding it difficult to understand why this would be such a challenge.