Besides its ubiquitous fame in optoelectronics, halide-perovskites (HPs) have also carved a niche in the domain of resistive switching memories (Re-RAMs). However owing to material and electrical instability challenges that encompass HP thin-films, rarely perovskite Re-RAMs are used to experimentally demonstrate data processing which is a fundamental requirement for neuromorphic applications. Here, in a first, lead-free, ultrahigh density HP nanowire (NW) array Re-RAM has been utilized to demonstrate image processing via design of convolutional kernels. The devices exhibited superior switching characteristics including high endurance of 5×10^6 cycles, ultra-fast erasing and writing speed of 900 ps and 2 ns respectively and > 5×10^4 s retention time for the resistances. The work is bolstered by in depth mechanistic study and first principle simulations which bear evidence of electrochemical metallization triggering the switching. Employing the robust multi-level switching trait, image processing functions of embossing, outlining and sharpening were successfully implemented.