In order to obtain the shapes, sizes, fall velocities, orientations of raindrops in the natural condition, starting from the particle imaging velocimetry transient visual measurement technology, an image restoration algorithm based on point spread function and a velocity calculation algorithm based on auto-correlation image are investigated, and a natural precipitation imaging velocimetry system based on a CCD sensor is designed and developed in this paper. The diameters, terminal velocities, axis ratios, orientations and their distributions of raindrops are investigated by the field experiment. The velocities of raindrops increase exponentially with the diameter, and axis ratios of raindrops decrease linearly with the diameter, the orientations of raindrops are easily affected by the wind, the average canting angle for all drops tends to be nearly symmetric 2.1° with a standard deviation of 11.5°. The empirical relations between velocity and diameter, between axis ratio and diameter are fitted by nonlinear fitting and linear fitting of observed data. In this paper we further provide microphysical characteristics of raindrops with local features. compared with those from the existing model. These conclusions are of important significance for the research on cloud precipitation physics, ground calibration of weather radar, assessment of rain induced microwave attenuation, etc.