This scoping review protocol aims to map the extent, range, and nature of existing literature on AI-assisted three-dimensional crime scene reconstruction using two-dimensional crime scene sketches and photographic evidence. Crime scene reconstruction is increasingly moving toward digital, spatial, and computational methods, while artificial intelligence, computer vision, photogrammetry, and structure-from-motion techniques are creating new opportunities for transforming conventional scene documentation into three-dimensional forensic visualizations.Guided by JBI scoping review methodology and reported in accordance with PRISMA-ScR, the review will identify input modalities, computational methods, reconstruction workflows, output forms, validation approaches, reported limitations, and forensic relevance across forensic science and technically relevant adjacent domains. The review will consider literature involving 2D sketches, scene diagrams, floor plans, photographs, multi-view images, video-derived frames, and hybrid visual inputs. Relevant computational approaches may include artificial intelligence, machine learning, deep learning, semantic segmentation, object detection, scene-layout estimation, photogrammetry, structure-from-motion, generative modelling, NeRF-based reconstruction, and sketch-to-3D or image-to-3D workflows.The review will use a Population/Concept/Context eligibility framework and will search Scopus, Web of Science Core Collection, PubMed/MEDLINE, IEEE Xplore, and ScienceDirect, with supplementary searching through Google Scholar, backward reference checking, and forward citation tracking. Retrieved records will be deduplicated using Mendeley Reference Manager, screened independently by two reviewers, and charted using a piloted data charting form. Findings will be synthesized through descriptive numerical summaries, tabular evidence mapping, and narrative synthesis.This protocol is expected to provide a transparent roadmap for identifying methodological trends, evidence gaps, validation issues, and future research directions in AI-assisted 3D crime scene reconstruction. The findings will support future experimental studies, forensic tool development, and the possible design of a focused systematic review.