The invention provides a deep learning technology-based field investigation method and system. The method comprises the steps of S101, obtaining a deep learning model and a proof photo obtained by training a deep learning platform based on a sample set; S102, obtaining an identification type result of the proof photograph through a deep learning model, judging whether the proof photograph of the same target pattern spot has feature consistency or not according to the identification type result of the proof photograph, if yes, executing S103, and if not, executing S104; S103, analyzing and judging by using a rule engine according to the identification type result of the proof photograph in combination with the current land utilization situation so as to obtain an investigation result of thetarget pattern spot; and S104, determining that the proof photograph does not have feature consistency. According to the method, the land type corresponding to the proof photo can be automatically recognized through the deep learning model, recognition is more accurate, repeated work is avoided by reminding the abnormal photo which does not meet the requirement, the survey result is obtained through the rule engine, the speed is high, and errors are not likely to happen.