We have a (point) light source at position (xL, yL) with yL > 0, and a finite series of line segments, all of finite non-zero length, given by the coordinates of their two endpoints. These endpoints are all different. The line segments are all situated above the x-axis (y > 0).
The segments cast their shadows onto the x-axis. We assume that the shadows of two segments either do not overlap at all, or have an overlap that has some non-zero width (they do not just touch). We also assume that for each segment its shadow is more than just one point, i.e., there is no segment that is directed toward the light source. The height of the light source (yL) is at least 1 unit larger than the y-coordinates of the endpoints of the line segments. This guarantees that indeed each line segment has a bounded shadow on the x-axis.
The collection of shadows divides the x-axis into dark and lighted areas (intervals). The problem is to determine the number of lighted areas — which is at least 2 (if there is at least one line segment, otherwise it is 1).
In the picture below the three line segments A, B and C cause three lighted areas, as indicated.