Unfortunately the design of the basic PIR can not discriminate between different heat sources. It reacts to changes in the heat signatures in the field of its view. If they are static it remains passive and if it detects a change it reports that change. One would need some sort of temperature discriminator that responded only to temperatures in the range of human body temp to be able to see humans and ignore animals of other body temps and objects.
Welcome to the element14 community. There have been a lot of conversation here on this same topic. Like jw0752 mentions above, PIR sensors in and of themselves do not discriminate the source of heat they are seeing. There are some ways that a standard PIR sensor could give you some more information about a source, but most PIR motions sensors are designed to look only at large changes in heat.
Having said that, there are some ways to get more information. One is to use a comparative approach, where the PIR sees a reference signal and compares it to the potential human signal. I have seen this done where a shutter is placed inline with the field of view. The shutter is a wheel like structure that is 50% open (to view the heat source in question) and 50% closed to sense an ambient condition that rotates, periodically obscuring the PIR detector. The electronics compare the live to ambient signal to develop a difference level (think of a sinusoidal waveform where the peak to peak voltage is difference value). This level could be used to determine the strength of the heat source. Too low or too high of a signal would imply a non-human heat source, where a signal between the extremes might be more likely to imply a human heat source. The gain of the circuit would need to be determined such that the human temperature range was always near the middle of the reading range. This approach does suffer in that it is very sensitive to the distance the human heat source and the PIR sensor, i.e. sensitivity decreases as the distance increases.
Ok, sorry I couldn't resist being little facetious... as this was the first thing that popped into my head. As in, I very much doubt there is any sensor out there that can readily determine the genetic makeup of a human.
So I assume you are asking whether a PIR can differentiate the heat signature of an adult human form (as in adult body mass walking upright) from something else. And the anwer is... sort of... as in, a pyroelectric sensor (PIR) also requires the right type of fresnel lens to help determine this. If you use a "curtain" type fresnel lens, which has a narrow vertical field of view combined with a wide horizontal field of view, this lens can be positioned in such a way as only detect objects of a certain height, for example. If you then combine this technique with filtering or the methods, as described above, then you have a good chance of differentiating. Another option to improve results, is to also use two PIRs at different heights, and use as a comparitor where both need to trigger concurrently, to derive that it is a human form being detected.
A simple PIR sensor cannot uniquely identify a human. A PIR sensor detects long wave infrared energy. Many things generate this type of energy. All warm-blooded animals and all sources of heat generate this type of energy. A human has a characteristic temperature that is different from most other emitters and humans have a defined size range that limits the set of possible sources, so a more complex PIR sensor that can measure temperature and size may be able to distinguish a human, but it would likely require a full thermal camera and significant image processing to get any kind of accuracy.
You have started a very interesting conversation. Another aspect to this subject is that the radiation given off by a source would have a characteristic bell curve with a distribution of frequencies that peak and the nominal temperature. In the case of a dog for example there is only a 2 degree C difference from human temperature. It would take a much more sophisticated sensor than the ones incorporated in a PIR to see the difference. There have however been some interesting idea put forth and you have the basis here for some good experiments.
Another good point. On the same concept I referenced above (shuttering), we used notch filters on top of an array of PIR elements to narrow our view to certain areas of the IR spectrum (both long and short waves) to create a number of detectors, which we used to detect specific gases (in our case measuring the absorption caused by the presence of gases instead of directly reading a warm body).
