A passive sensor does not need any additional energy source?
There are no strict, formal definitions of "passive sensor" and "active sensor". These terms can have different meanings based on the context in which they are used. Both of the quoted paragraphs are basically correct but they have different contexts.
Regarding the "additional energy source", this would typically be a battery or similar power supply. The word "additional" here means a source in addition to any energy provided by the sensor itself. In many cases the sensor itself provides no energy so the "additional" energy source is "in addition to the sensor" not "in addition to a another energy source".
Adding my two cents. Maybe there are no formal definitions, but the one given in your textbook seems pretty formal to me, so let's stick with that.
And it is quite straightforward - no additional energy should be applied to the sensor in order for it to be considered "passive".
The keyword here is "additional". The caveat is that the physics laws require some energy to be applied to the sensor in order for it to detect something. In fact, energy IS the thing that is being sensed.
The thermocouple converts thermal energy into voltage, photodiode (in photovoltaic mode) converts light energy, and piezoelectric sensor converts mechanical energy. You can add a simple coil in the magnetic field to this list, which converts motion into voltage. And so on.
An "active" sensor also senses energy (once again, there is no way around this). However unlike "passive" sensor it does not generate voltage on its own. Instead it changes its property (usually resistance) in response. And at this point you need additional energy to read this change and amplify it to some useful level.
The same photodiode in photoconductive mode becomes active sensor, because you need to apply voltage to read its resistance. Hall effect sensor requires current across it to be able to detect magnetic field. MEMS sensor requires electrical excitation to detect motion.
Your second source sounds confusing, and in fact it is. That is because it mixes complex sensor systems into the picture, and tries to categorize them by simple sensor criteria.
Now, the problem is that in addition to simple sensors described above we also call "sensors" some complex devices of which the sensor itself is just a tiny component.
For example, garage door sensor built on photodiode in photovoltaic mode will be an active "sensor", even though diode itself is passive sensor. Why? because you have to apply energy to the circuit that detects voltage on a diode, amplifies it and switches external relay. Not to mention energy consumed by the laser, which can be considered a part of the whole "garage door sensor" device.
Same with metal detector. The coil that generates voltage in response to magnetic field is passive sensor by the definition above. However in order to induce magnetic field in a metal you need to generate alternating magnetic field of your own, and for that you need additional energy. So, if you take the entire metal detector device and call it a "sensor" then it becomes active by the same definition.
I hope the above makes things clearer. Having said that [insert mischievous grin here], a bonus question for you:
Is laser sensor card an active or passive device?
On one hand, it detects IR radiation without any energy applied to it. On the other, it has to be recharged (i.e. energy must be pumped in) by visible light before use.
UPDATE:
After reading all the answers and comments here I believe the root of the confusion and the fuel of all the discussions is in the definition of a "sensor" itself.
Let's get this straight:
some devices can function by themselves, directly converting measured physical phenomenon (energy) into useful output signal (e.g. voltage, visible light, mechanical movement). By the textbook definition these all are passive sensors.
some devices cannot produce meaningful output signal without additional energy input, whether directly used for excitation or indirectly for required amplification circuits. In this case we should call the device a "sensing element" (not a sensor!), which makes the containing device an active sensor.
finally, even passive sensors can be equipped with the additional circuitry to simplify their usage (various current/voltage transducer modules, for example), which makes them active again.
So, if we count only sensing elements as "sensors", then classification is straightforward. If we allow complex devices to be called "sensors" even though the actual sensing element is just a tiny part of them, then majority of the modern sensors will be active devices.
Writers can divide the sensors to classes as they want. I would make my division as follows:
Active sensors make some actions which affect the environment. They for example send radio waves or apply a force. The response is analyzed and the information is extracted.
Passive sensors only take what the environment offers and they change their state along the changes in the environment. For example a light dependent resistor changes its resistance along the variations in light. Or a dynamic microphone gives a voltage which depends on how loud voices there exist. The interesting data can be sometimes extracted easily or it can need complex decision making. But the most complex decision making system doesn't make a sensor active as long as the sensor doesn't send anything which is used to test the environment.
Active sensors surely need some energy source to be able to make physical work. That energy can be collected from the environment, it can be taken from a source which is assembled in the sensor box or there can be supply lines from an external source such as a computer or AC mains.
Passive sensors need energy to transfer their state to the information user which can well be a device. That user can supply everything which is needed to read the state (for example some voltage to cause a current which depend on the resistance which depend on lightness, temperature humidity or other interesting quantity). Or the sensor can have the tools which use energy and convert the state to more robust or more informative form of data. It can for ex. have a numerical display or to WiFi connected computer. That energy can as well be collected from the environment, stored in the sensor box or supplied with wires.
A dynamic microphone collects from the environment the energy which carries the state to the user. The sensed sound directly makes the needed electricity. But that starts to be false as soon as one decides that a preamp is needed for more robust signal and within the same breath he says that the preamp is a part of the sensor.
Intruder alarm sensors for example can as well be passive as active and both need energy to do their job. PIR sensor does not send anything, but their photodiodes need quite complex amplification, filtering and decision making circuits. In addition a clear intrusion alarm signal must be formed. There are also radar based intrusion alarms which I consider active.
Even a door switch can be active or passive. It's active for ex. if it contains a spring which moves a contact if a door opens. It has got its energy beforehand loaded to the spring when the door was closed. A passive door switch hasn't that spring, the contact must be moved by the door to both directions.
As said everyone can write his definitions as he wants. Nobody cares as long as the writer doesn't fool the readers or collide against law.