Digital holographic on-chip microscopy achieves large space-bandwidth-products (e.g., >1 billion) by making use of pixel super-resolution techniques. To synthesize a digital holographic color image, one can take three sets of holograms representing the red (R), green (G) and blue (B) parts of the spectrum and digitally combine them to synthesize a color image. The data acquisition efficiency of this sequential illumination process can be improved by 3-fold using wavelength-multiplexed R, G and B illumination that simultaneously illuminates the sample, and using a Bayer color image sensor with known or calibrated transmission spectra to digitally demultiplex these three wavelength channels. This demultiplexing step is conventionally used with interpolation-based Bayer demosaicing methods. However, because the pixels of different color channels on a Bayer image sensor chip are not at the same physical location, conventional interpolation-based demosaicing process generates strong color artifacts, especially at rapidly oscillating hologram fringes, which become even more pronounced through digital wave propagation and phase retrieval processes. Here, we demonstrate that by merging the pixel super-resolution framework into the demultiplexing process, such color artifacts can be greatly suppressed. This novel technique, termed demosaiced pixel super-resolution (D-PSR) for digital holographic imaging, achieves very similar color imaging performance compared to conventional sequential R,G,B illumination, with 3-fold improvement in image acquisition time and data-efficiency. We successfully demonstrated the color imaging performance of this approach by imaging stained Pap smears. The D-PSR technique is broadly applicable to high-throughput, high-resolution digital holographic color microscopy techniques that can be used in resource-limited-settings and point-of-care offices.