Abstract
The real-time pedestrian detection algorithm requires the model to be lightweight and robust. At the same time, the pedestrian object detection problem has the characteristics of aerial view Angle shooting, object overlap and weak light, etc. In order to design a more robust real-time detection model in weak light and crowded scene, this paper based on YOLO, raised a more efficient convolutional network. The experimental results show that, compared with YOLOX Network, the improved YOLO Network has a better detection effect in the lack of light scene and dense crowd scene, has a 5.0% advantage over YOLOX-s for pedestrians AP index, and has a 44.2% advantage over YOLOX-s for fps index.
Highlights
In variety of realistic application of multi-object detection, accurate statistics of traffic flow play a crucial role in the self-driving system and intelligence transportation, which includes the real time detection task on the intersection
Use support vector machines (SVM)[4], Adaptive Boosting (Adaboost)[5] and other classifiers to judge whether there is a object in the region
There are based on candidate regions two stage target detection algorithms Faster-RCNN (Faster region-based Convolutional Neural Networks)[6], based on the regressive one-stage target detection algorithm SSD (Single Shot Detection)[7], and YOLO(You Only Look Once)[8], etc
Summary
In variety of realistic application of multi-object detection, accurate statistics of traffic flow play a crucial role in the self-driving system and intelligence transportation, which includes the real time detection task on the intersection. In the face of complex and changeable natural environment, it is difficult to achieve satisfactory results either through visible light information or infrared information Such as in fog, rain, fog, weak light, pedestrians in visible light image target invisible or ambiguous situations, and the infrared image can improve the quality of such a case the image [9], but because of its rich cannot describe the pedestrian characteristics of contour and the characteristics of color information, in a crowded scenarios can lead to more residual and checked by mistake. The experimental results show that compared with similar models such as YOLOX[13], the improved YOLO5s[14] is more robust under weak light scene and dense crowd scene
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
