Using DeepPVMapper's detections as an independent, remote-sensing-based ground truth, this study audits the RNI and RTE's internal grid-connection data — and finds a systematic downward bias in both.
Read the audit →Papers and projects that build on or relate to DeepPVMapper — production estimation, TSO benchmarking, adoption determinants, explainability, and external mapping & GIS projects using the registry.
Browse related work →Popular-science coverage and press articles about DeepPVMapper and the registry audit, including a three-part series in PV Magazine France.
Read the coverage →The detailed architecture, training setup, and validation methodology behind the detection pipeline — tiling, classification, segmentation, and characteristic extraction.
Read the pipeline →