Is it possible to combine decimation and low pass filtering in one step? Not necessarily only for images but also for general signals.
Yes, that's what people usually do when they implement downsampling: since of the output of the anti-aliasing filter, you throw away N-1 samples, why even calculate these?
The trick is to decompose your filter into polyphase components, which enables you to run the resulting filter operation only once per output of the downsampling, instead of once per input. There's plenty of reference implementations - from GNU Radio's decimating FIR filters, to rescalers in image processing hardware.