Not wearing seatbelts or using phones while driving is illegal in many countries, but such violations still remain common. In recent years, there has been a growing interest in using object detection technologies to automatically detect such violations through traffic surveillance systems.

This study evaluated how camera angles, both vertical and horizontal, influence the detection of in-vehicle objects, such as seatbelts and mobile phones, by using a YOLOv11s model trained specifically for this purpose. In addition, the impact of CLAHE as a preprocessing technique was explored across four enhancement levels on both day- and nighttime images.

The results showed that CLAHE had a small effect on overall performance, where a clip limit of two was the best configuration. Camera angles were found to primarily affect object visibility rather than detection accuracy, where seatbelts were more prone to being obscured in some angles. The conclusion was that object visibility, camera placement, image clarity, and lighting conditions are all critical factors in developing traffic surveillance systems.