In the project MicroVision, we will collect data (mainly image data, possibly other data such as GPS, LIDAR) in traffic, from sensors installed on a micro-mobility vehicle (such as an electric scooter). The purpose of the data collection is to gather mainly image data to train object classification and detection model for different road users. Existing models can already classify and detect certain, common road users like pedestrians, or vehicles like cars or bicycles. However, e-scooters, for instance, are usually not classified in publicly available models, and neither are e-scooter and bicycle riders, i.e., the vehicle and the rider combined. This is where this project comes into play, by collecting and annotating data of such cases, training object detection and classification models, and releasing the anonymized training data and trained models online. Specifically, we will collect image (and possibly LIDAR, GPS, and other telemetry data) with an equipped vehicle in real traffic. The images will be annotated to allow training of machine-learning models, that, together with the data, shall be published online for others to re-use. We thereby contribute to research and development that aims at improving micro-mobility, in particular its safety.