Is the energy harvester a magnet that moves back and forth within a coil of insulated wire? If so, you should be able to float the two terminals, Connect them to the junction of D1 and D4 and to the junction of D2 and D3. Then "ground" if necessary either Node 1 for negative output from Node 2 with respect to "ground," or "ground" Node 2 for positive output from Node 1 with respect to "ground". Either way, your 'scope should show full-wave rectification with respect to "ground".
Note: you cannot measure a "floating" input signal without a differential input to your oscilloscope. Sometimes this can be accomplished by using two channels on the oscilloscope. You apply one of the two floating signal wires from your energy harvesting device to each channel. Set one channel to invert its signal, then subtract the two channels. There must be a "ground" return from the energy harvester source for the two channels, easily synthesized by connecting, say, 100 kΩ resistors to each energy harvester output and then connecting the junction of these two resistors to your circuit "ground". The actual value of these two "virtual ground" resistors is not critical, but their series combination does place an additional load on the energy harvester, so use as large a value as practical, depending on the input impedance of the oscilloscope. This input impedance is typically one megohm shunted by a few picofarads of capacitance. The resistors for the "virtual ground" should be about one tenth of this value or less.
Actually, if you connect the two resistors as described above to create a "virtual ground," you don't need to have the channel differencing method to monitor the output of the energy harvesting device. You can use either channel on the oscilloscope to monitor half of the output.