Abstract: The integration of high proportional new energy to a new type power systems and the implementation of the "double carbon" goal have promoted the rapid development of the photovoltaic industry. Currently, China's photovoltaic installation capacity has ranked in the forefront of the world. Facing challenges of transmission jamming, insufficient fault identification precision, and low data processing speed that cannot meet the requirements of high precision and fast fault identification, this paper proposes an intelligent photovoltaic video inspection technology with high precision and fast fault identification. It comprehensively improves fault identification precision and speed with the combined power of anti-interference inspection data transmission, flexible computing resource allocation, and fault identification algorithm optimization. On this basis, an intelligent photovoltaic video inspection system is constructed to realize all-weather and all-round intelligent monitoring of photovoltaic power stations. Compared with the inspection system based on parallel scheduling and complex neural network, the proposed video inspection system increases transmission rate by 97.20% and 85.18%, reduces video stream processing delay by 40.16% and 23.78%, and improves the identification precision of snail stripe fault by 22.12% and 8.77%. The intelligent photovoltaic video inspection system with high precision and fast fault identification can substantially improve identification precision, realize reliable transmission and fast processing of inspection data, and improve inspection efficiency, security, and reliability.