The Moroccan Video Intelligent Transport System dataset for vehicle detection

Intelligent transportation systems (ITS) are composed of functionalities that aim to reduce traffic jams and detect anomalies. These functionalities include vehicle detection and classification that represent critical tasks and require a high degree of accuracy to ensure the extracted information's reliability. Therefore, we present a key data set related to the detection of vehicles in Morocco, nicknamed MoVITS. The data provided in this paper is collected using a stereo-based vision system, and it is useful for vehicle detection and classification by using different methods for many purposes. This data set is intended to support researchers, stakeholders, and the general public interested in ITS. The vehicle video dataset incorporated with deep learning, transfer learning, and fine-tuning can provide outstanding vehicle detection and classification results.

智能交通系统（Intelligent Transportation Systems，ITS）由旨在缓解交通拥堵、检测异常状况的各类功能模块组成。车辆检测与分类是其中的核心任务，需达到极高精度，方可保障所提取信息的可靠性。据此，本文提出一款面向摩洛哥地区车辆检测任务的核心数据集，昵称为MoVITS。本文所提供的数据通过基于立体视觉的视觉系统（stereo-based vision system）采集获取，可通过多种方法适配多场景下的车辆检测与分类任务。本数据集旨在为关注智能交通系统的研究人员、行业利益相关者以及社会公众提供支撑。这款集成深度学习、迁移学习（Transfer Learning）与微调（Fine-tuning）技术的车辆视频数据集，可实现优异的车辆检测与分类效果。