Gamma radiation dose measurement method based on machine vision

The invention discloses a gamma radiation dose measurement method based on machine vision, comprising a gamma radiation noise detection module and a gamma radiation dose estimation module, the gamma radiation noise detection module uses MobileNet V3 as a network backbone and LR-ASPP as a detection head, and the gamma radiation dose estimation module estimates the gamma radiation dose based on the MobileNet V3. According to the gamma ray noise detection method, the number and the form of gamma ray noise in an image are accurately recognized, and meanwhile, a lightweight neural network model (Mobile-Nuclear Net) with execution performance and detection precision of a space and channel attention mechanism (CBAM) is added, so that the use of noise information which cannot be utilized originally in the image becomes possible through accurate recognition of the number and the form of the gamma ray noise in the image. Due to the fact that the gamma ray intensity received by the camera is strongly correlated with the