A car brakes with velocity v(t)=20−4t m/s and stops at t=5 s. On a velocity-time graph this is a line from 20 down to 0, and the region under it is a triangle with area 21⋅5⋅20=50. That area is the distance travelled, 50 m, because each thin strip of the region is a velocity times a short time, which is the distance covered in that moment. The definite integral ∫05(20−4t)dt adds up all those strips.
For curves with no simple shape, an antiderivative does the adding. F(t)=20t−2t2 has derivative 20−4t, and F(5)−F(0)=(100−50)−0=50, the same answer. Only the change in F matters, which is why any +C cancels and why the order is always top bound minus bottom bound.
The integral counts area below the axis as negative. For ∫02πsinxdx, the hump from 0 to π has area 2 above the axis and the dip from π to 2π has area 2 below it, so the integral is 2−2=0. That 0 is a net result, like walking 2 km east and then 2 km back west. The total area, like the distance walked, is ∫02π∣sinx∣dx=4.
Think of it like this. A definite integral works like a month's bank statement. Deposits and withdrawals both happen along the way, and the net change is deposits minus withdrawals, whatever the opening balance was. Area above the axis is a deposit and area below is a withdrawal. To find how much money moved in total, you would add both kinds instead.
Core idea. A definite integral is one number — the signed area under f from a to b — computed as F(b)−F(a).