How to compute the likelihood that the pixel i belongs to class sea or land

I have remote sensing image and I constructed two GMM models for each class(sea and land) by fitgmdist function.how to use constructed GMMs to compute the probability of pixel belong to sea class or land class?

GMModel_sea = fitgmdist(sea_data,2);
GMModel_land = fitgmdist(land_data,5);
prob= pdf(GMModel_sea ,sea_data);

I have used pdf function to compute the likelihood that the pixel(i) belongs to class sea or land. But I am getting very low values lower 0 and values greater than 1. Why is it happening?