Implementation of a DPU-Based Intelligent Thermal Imaging Hardware Accelerator on FPGA

Thermal imaging has many applications that all leverage from the heat map can be constructed using this type of imaging. It used in Internet Things (IoT) to detect features surroundings. In such a case, Deep Neural Networks (DNNs) carry out visual analysis tasks which provide system with capacity make decisions. However, due their huge computational cost, networks are recommended exploit custom hardware platforms accelerate inference as well reduce overall energy consumption system. work, an adaptive is proposed, intelligently configure itself based on battery level. Besides achieving maximum speed increase equals 6.38×, proposed achieves significant reduced by 97.81% compared conventional general-purpose CPU.