Wastewater pipe defect rating model for pipe maintenance using natural language processing

Introduction Closed-circuit video (CCTV) inspection has been the most popular technique for visually evaluating interior status of pipelines in recent decades. Certified inspectors prepare pipe repair document based on CCTV inspection. The traditional manual method assessing structural wastewater conditions from documents takes a long time and is prone to human mistakes. automatic identification necessary texts received little attention. Computer Vision Machine Learning models failed estimate damage because they are not entirely understood have difficulty providing high data needs. Hence, problems physically consistent findings due their Currently, very small curated annotated image set with well-defined, precisely labeled categories test models. Methods This study provides valuable determine defect rating by developing an automated framework using Natural Language Processing (NLP) small, images, data, more text data. used this broken into grammatical units NLP technologies. next step analysis entails words find frequency defects then classify them respective ratings maintenance. Results discussions proposed model achieved 95.0% accuracy, 94.9% recall, 95% specificity, 95.9% precision score, 95.7% F1 showing potential be large-scale accurate efficient pipeline failure detection improve quality pipeline.