The GRANDProto300 (GP300) array is a pathfinder for the Giant Radio Array for Neutrino Detection (GRAND) project. Serving as a test bench, the GP300 array is expected to pioneer techniques of autonomous radio detection including identification and reconstruction of nearly horizontal cosmic-ray (CR) air showers, and shed light in understanding the interesting `transition region' from the galactic to extragalactic CR sources. An offline analysis of signal identification over ambient noise is crucial at this stage, where very relaxed self-triggering thresholds of radio antennas will be used for study purposes. In this work, we show results and efficiency of signal identification with classical approaches using a wide set of simulated realistic signal templates and also validated against measured background recorded by deployed prototypes.