The first thing to realise is that the patterns in the two periods are identical - the difference being that the ionisation energies in period 3 are all lower than those in period 2.

Explaining the general trend across periods 2 and 3
The general trend is for ionisation energies to increase across a period.
In the whole of period 2, the outer electrons are in 2-level orbitals - 2s or 2p. These are all the same sort of distances from the nucleus, and are screened by the same 1s2 electrons.
The major difference is the increasing number of protons in the nucleus as you go from lithium to neon. That causes greater attraction between the nucleus and the electrons and so increases the ionisation energies. In fact the increasing nuclear charge also drags the outer electrons in closer to the nucleus. That increases ionisation energies still more as you go across the period.
In period 3, the trend is exactly the same. This time, all the electrons being removed are in the third level and are screened by the 1s22s22p6 electrons. They all have the same sort of environment, but there is an increasing nuclear charge.
Why the drop between groups 2 and 3 (Be-B and Mg-Al)?
The explanation lies with the structures of boron and aluminium. The outer electron is removed more easily from these atoms than the general trend in their period would suggest.
Be 1s22s2 1st I.E. = 900 kJ mol-1
B 1s22s22px1 1st I.E. = 799 kJ mol-1
You might expect the boron value to be more than the beryllium value because of the extra proton. Offsetting that is the fact that boron's outer electron is in a 2p orbital rather than a 2s. 2p orbitals have a slightly higher energy than the 2s orbital, and the electron is, on average, to be found further from the nucleus. This has two effects.
*
The increased distance results in a reduced attraction and so a reduced ionisation energy.
*
The 2p orbital is screened not only by the 1s2 electrons but, to some extent, by the 2s2 electrons as well. That also reduces the pull from the nucleus and so lowers the ionisation energy.
The explanation for the drop between magnesium and aluminium is the same, except that everything is happening at the 3-level rather than the 2-level.
Mg 1s22s22p63s2 1st I.E. = 736 kJ mol-1
Al 1s22s22p63s23px1 1st I.E. = 577 kJ mol-1
The 3p electron in aluminium is slightly more distant from the nucleus than the 3s, and partially screened by the 3s2 electrons as well as the inner electrons. Both of these factors offset the effect of the extra proton.
Why the drop between groups 5 and 6 (N-O and P-S)?
Once again, you might expect the ionisation energy of the group 6 element to be higher than that of group 5 because of the extra proton. What is offsetting it this time?
N 1s22s22px12py12pz1 1st I.E. = 1400 kJ mol-1
O 1s22s22px22py12pz1 1st I.E. = 1310 kJ mol-1
The screening is identical (from the 1s2 and, to some extent, from the 2s2 electrons), and the electron is being removed from an identical orbital.
The difference is that in the oxygen case the electron being removed is one of the 2px2 pair. The repulsion between the two electrons in the same orbital means that the electron is easier to remove than it would otherwise be.
The drop in ionisation energy at sulphur is accounted for in the same way.
Trends in ionisation energy down a group
As you go down a group in the Periodic Table ionisation energies generally fall. You have already seen evidence of this in the fact that the ionisation energies in period 3 are all less than those in period 2.
Taking Group 1 as a typical example:

Why is the sodium value less than that of lithium?
There are 11 protons in a sodium atom but only 3 in a lithium atom, so the nuclear charge is much greater. You might have expected a much larger ionisation energy in sodium, but offsetting the nuclear charge is a greater distance from the nucleus and more screening.
Li 1s22s1 1st I.E. = 519 kJ mol-1
Na 1s22s22p63s1 1st I.E. = 494 kJ mol-1
Lithium's outer electron is in the second level, and only has the 1s2 electrons to screen it. The 2s1 electron feels the pull of 3 protons screened by 2 electrons - a net pull from the centre of 1+.
The sodium's outer electron is in the third level, and is screened from the 11 protons in the nucleus by a total of 10 inner electrons. The 3s1 electron also feels a net pull of 1+ from the centre of the atom. In other words, the effect of the extra protons is compensated for by the effect of the extra screening electrons. The only factor left is the extra distance between the outer electron and the nucleus in sodium's case. That lowers the ionisation energy.
Similar explanations hold as you go down the rest of this group - or, indeed, any other group.
Trends in ionisation energy in a transition series

Apart from zinc at the end, the other ionisation energies are all much the same.
All of these elements have an electronic structure [Ar]3dn4s2 (or 4s1 in the cases of chromium and copper). The electron being lost always comes from the 4s orbital.
Note: Confusingly, once the orbitals have electrons in them, the 4s orbital has a higher energy than the 3d - quite the opposite of their order when the atoms are being filled with electrons. That means that it is a 4s electron which is lost from the atom when it forms an ion. It also means that the 3d orbitals are slightly closer to the nucleus than the 4s - and so offer some screening.
You will find this commented on in the page about electronic structures of ions.
As you go from one atom to the next in the series, the number of protons in the nucleus increases, but so also does the number of 3d electrons. The 3d electrons have some screening effect, and the extra proton and the extra 3d electron more or less cancel each other out as far as attraction from the centre of the atom is concerned.
The rise at zinc is easy to explain.
Cu [Ar]3d104s1 1st I.E. = 745 kJ mol-1
Zn [Ar]3d104s2 1st I.E. = 908 kJ mol-1
In each case, the electron is coming from the same orbital, with identical screening, but the zinc has one extra proton in the nucleus and so the attraction is greater. There will be a degree of repulsion between the paired up electrons in the 4s orbital, but in this case it obviously isn't enough to outweigh the effect of the extra proton.