Which nature of light do the retinal cells observe?

Limiting the discussion to the rhodopsin, the detection of light is often described as "photon-rhodopsin interaction", thus stressing the particle nature. From ref. 1,

"... the receptor (rhodopsin) is located in high concentrations in

specialized organelles (rods and cones) connected to a

synaptic junction in the retina, enabling them to transmit the

signal (light) into the central nervous system. Upon

interaction with a photon, rhodopsin, in response to the

photoisomerization of the chromophore retinal, undergoes a

conformational change. This enables many heterotrimeric

G-protein molecules (transducin) to couple to an activated

rhodopsin, then exchange their bound GDP for GTP and

split into their α- and βγ-subunits. In the visual system, the

α-subunit triggers an enzymatic cascade, which in the end

leads to an electrical signal and neuronal stimulation. The

immense amplification of the signal through the enzymatic

cascade allows even single-photon events to be registered..."

What this boils down to is that the interaction is not exactly like the photoelectric effect, whose main result is the simple displacement of an electron from an atom, because it involves the complex reconfiguration, binding and splitting of molecules followed by a chemical cascade. But you might liken it to a photoelectric tube (photon detector) in a quantitative sense because of the large energy gain the process realizes.

There's also an argument for the wave nature, in color vision. The opsin portions of the red, blue and green pigment molecules in the cones are filters that determine the part of the spectrum applied to the rhodopsin-like portion of the molecule, and thus the cone's peak response color (ref. 2). Color (and wavelength and frequency) are generally considered in the context of waves.

It involves absorption of a photon by an electron, with the photon's energy transferred to the electron. So the actual process involves discreet "parcels" of energy - particles. However, when discussing the energy of the photons, we still use the terms wavelength and frequency. It's easier to categorise the range of photon energies by wavelength and frequency, than by its energy in electron volts, joules or other numerous units of energy.