tempcal: change to double as float was insufficient for calculation

cmake/configs/nuttx_px4fmu-v2_default.cmake

src/modules/tempcal/matrix_alg.cpp

src/modules/tempcal/matrix_alg.h

src/modules/tempcal/polyfit.cpp

src/modules/tempcal/polyfit.h

src/modules/tempcal/tempcal_main.cpp