Road traffic accidents remain a critical public health and economic issue in Cambodia, with traffic violations at signalized intersections contributing significantly to incidents and congestion. This study examined traffic violations at five intersections in Phnom Penh and assessed the performance of existing ANPR cameras in detecting key violations, including red-light running, motorcycle helmet non-use, car seatbelt non-use, and mobile phone use while driving. Data were collected through manual counts, ANPR-generated reports, and interviews with traffic police. Results show that motorcycles accounted for the majority of violations, with ANPR systems reliably detecting most externally observable violations but underperforming for in-cabin behaviors such as seatbelt and phone use. With an average error rate of 9.3%, the ANPR systems detected 10.2% of total traffic volume as violations, slightly below the 11.3% observed manually. This study indicates that the ANPR system captured over 90% of violations for the included types, demonstrating high reliability and effectiveness for automated traffic monitoring and enforcement. The study demonstrates the potential of ANPR systems to support automated traffic enforcement and data-driven policy for urban road safety improvements.