First of all, ensure your data meets the necessary assumptions for accurate analysis and then use multinomial logistic regression, not binomial regression. Binomial logistic regression is limited to dependent variables with two categories so binomial might be not appropriate.
Yadpiroon Siri Thanks for your answer and sorry for misleading you. I should have written independent variable rather than dependent variable. The dependent variable is dichotomous (Return vs. not Return), while the independent variable comprises 3 categories: urban, rural and camp.