Several factors to consider when collecting outdoor datasets
Everyone knows that deep learning is greatly affected by data. A good data set may allow us to get a better model. Therefore, when we collect or make a data set, we should consider as many influencing factors as possible .
Take the outdoor dataset as an example. When we are collecting an outdoor data set, we should collect as many images of different scenes as possible to make our trained model have stronger generalization ability. Generally, we will consider the following aspects.
- Weather
Rain, snow, fog, dust, thunderstorms of varying intensities - Lighting
Uniform lighting of varying intensities
Exposure of varying intensities - Camera Shake and Low Quality Image
Blurring, Noise, Low Resolution - Season
Spring, Summer, Autumn, Winter - Urban, rural and wild
- Different landforms
Deserts, lakes, forests, grasslands, mountains - day and night
- Shooting angle, shooting distance and shooting height
over!
ps: I look forward to your continued additions!