Ca II absorbers, which trace cold, dusty, and metal-rich gas in star-forming regions, provide a powerful probe of the interstellar medium and galactic evolution. However, their intrinsic weakness and scarcity have long limited statistical studies, with only about 1,000 systems previously identified. This work introduces a dual convolutional neural network (CNN) framework that combines Ca II detection with Fe II cross-verification to minimize false positives and improve reliability. Applied to over 100,000 quasar spectra from the Sloan Digital Sky Survey (SDSS) Data Release 16, this method identified 1,646 Ca II absorbers—expanding the known catalog by more than a factor of three. Statistical analyses reveal that stronger absorbers are associated with dusty, turbulent disk environments, while weaker absorbers trace galactic halos. Using strong systems as probes, this study finds evidence of disk size growth over cosmic time, offering new constraints on the evolution of galactic structure and dust-rich environments.