Object detection is one of the most important aspects of computer vision which is essential for many industrial and safety-critical applications. It involves identifying and localizing objects within an image or video frame. Compared to basic image classification which merely assigns a label to an object, object detection uses bounding boxes to precisely locate and classify the objects. This research examines the application of You Only Look Once (YOLO) on safety equipment detection at construction sites. YOLO has transformed object detection with its unified methodology that strikes an appropriate balance between speed and accuracy. While monitoring and identification of employees, machinery, and safety hazards on construction sites is crucial for preventing accidents. YOLOv8 is evaluated in this study with various color spaces to determine how well they perform on the dataset. The potential of the model is evaluated using critical measures such as F1-score, recall, precision, and mAP. The findings show that the model performed best in RGB color space with Precision of 1.0, recall of 0.85, F1-score of 0.85, and mAP of 0.87. It offers significant improvements in accuracy and inference speed over previous versions, which makes it especially well-suited for challenging detection tasks in uncertain circumstances.