The invention discloses a digital holographic image 
reconstruction method based on 
deep learning, and the method comprises the following steps: S1, collecting an original image, and carrying out the sampling of the original image; s2, discarding part of data after sampling is completed, and compressing the data; s3, 
data transmission is performed after compression is completed, and the data is transmitted to a lower computer needing to be played; s4, the lower computer decompresses the compressed packet to which the data is transmitted; s5, after compression is completed, 
image restoration iscarried out, and the image is displayed; the structure is scientific and reasonable, safe and convenient use, a 
data acquisition card is subjected to A / D conversion and quantization post-
processing, acomputer technology is introduced, and numerical calculation of a computer is utilized to simulate an 
object wave model function and an optical interference function, so that image 
data conversion isfacilitated, the conversion efficiency is improved, the transmission speed is reduced after compression, data is transmitted to a cloud end, and safety and reliability are achieved.