Quantitative investigation of insect activity in their natural habitat is a challenging task for entomologists. It is difficult to address questions such as flight direction, predation strength, and overall activities using the current techniques such as traps and sweep nets. A multispectral kHz remote detection system using sunlight as an illumination source is presented. We explore the possibilities of remote optical classification of insects based on their wing-beat frequencies and iridescence features. It is shown that the wing-beat frequency of the fast insect events can be resolved by implementing high-sampling frequency. The iridescence features generated from the change of color in two channels (visible and near-infrared) during wing-beat cycle are presented. We show that the shape of the wing-beat trajectory is different for different insects. The flight direction of an atmospheric insect is also determined using a silicon quadrant detector.