Doubly robust estimation of causal effects for random object outcomes with continuous treatments