This project aimed to: 1) construct a cohort of no less than 10000 cases of f-ILD (including pneumoconiosis ≥3000 cases) with continuous regular follow-up to reveal the clinical phenotypes closely related to the development, progression and prognosis of pulmonary fibrosis; 2) systematically evaluate the safety and effectiveness of frozen lung biopsy, surgical lung biopsy/thoracoscopic lung biopsy and other techniques, and to optimize the histological diagnosis method of f-ILD; 3) construct a set of artificial intelligence (AI) evaluation system for quantitative evaluation of pulmonary fibrosis and its severity, and develop application software; 4) excavate and verify important molecular targets for the formation of pulmonary fibrosis and identify biomarkers; 5) combined with clinical phenotype, imaging, pathology and biomarkers to establish f-ILD early recognition and progress model, intervention strategies, guidelines and consensus, and applicated nationwide.