HSV Semantic Segmentation on Partially Facility and Phanerophyte SunShine-Shadowing Road

The previous work reported the importance of visual intelligence on Self-Driving Sweeping Bot (SDSB) to reach the pedestrian security and the sweeping efficiency, in which the machine vision on object detection and semantic segmentation in terms of vehicle, pedestrian, rubbishes detection and road segmentation would be further investigated. Subsequentially, this work proposed an optimizing HSV encoding framework embedded with the morphology operation to examine the semantic segmentation on campus sweeping road. Furthermore, considering the complex weather effects caused the road performed with partially sunshine shadowing effects by facilities and giant tree plants, such that led to the inconsistent color presentation in a road. The experimental results indicated that the proposed framework given the promising results to three categorized sunshine-shadowing images groups.