基于深度学习实现人行横道检测的设计

Design of Crosswalk Detection Based on Deep Learning

在分析驾驶场景中的人行横道和道路路面等道路元素视觉特点的基础上,采用经典特征提取和深度学习相结合的方法,从深度学习网络结构出发提出了一种能够在不同分支输出具有不同空间分辨率和不同场景特征的进行人行道检测的深度学习网络,并在所建的数据集上对所提方案进行实验,验证了所提算法的有效性。

Based on the analysis of the driving scene of crossings and road pavement road elements such as visual features,on the basis of the classical feature extraction and the combination of deep learning method,starting from the deep learning network structure,A deep learning network for pavement detection with different spatial resolution and different scene features is proposed,and data set is built on experiment in this paper.The proposed network verifies the effectiveness of the proposed algorithm.