In disaster risk management, a parametric insurance is a contract where the buyer enters into a protection that will payout under predefined conditions, comparing the value of a parameter or an index to a trigger. A disaster index estimating damages can be built by comparing satellite images from pre-disaster and post-disaster using Convolutional Neural Networks (ConvNets). Standardization of these financial instruments could guarantee a timely, diligent and transparent process to release needed funds after the catastrophe, particularly critical for developing countries.