The detection of potholes on Indian roads is crucial for improving road safety and reducing vehicle damage, especially in countries like India where road conditions are often affected by various environmental factors. This paper presents a robust pothole detection system that utilizes the environment-specific YOLOv8n pothole detection models specifically optimized for Indian weather conditions. The dataset used for training the model includes data collected under different environmental conditions such as morning, afternoon, evening, and night. To address lighting and visibility challenges, contrast enhancement and histogram equalization were employed. The pre-trained weights of YOLOv8n were fine-tuned to improve the performance of pothole detection across these varied conditions. The model demonstrated high performance in real-time detection, achieving a precision of 96.60% in morning conditions, 95.47% in the afternoon, 94.91% in the evening, and 88.34% at night. Recall rates ranged from 87.92% to 74.98% across different conditions, with a mean average precision of 96.09% in well-lit conditions and 88.46% at night. This work contributes to road maintenance automation by improving detection accuracy, reducing false positives, and providing reliable results under challenging weather and lighting conditions.