According to the industrial accident statistics published by the Japanese Ministry of Health, Labour and Welfare, approximately 2,000 fatal and non-fatal forklift-related injuries occur annually. Of these, the number of fatal accidents per year has remained in around 30 for the past few years. Especially big body industrial vehicles such as large capacity forklifts have large blind spots and the forklifts and equipped with safety mirrors and on-board cameras. However, a satisfactory effect cannot be achieved unless the forklift operator pays close attention to these functions. The accident analysis results by the Japan Industrial Vehicles Association also indicate that there are always a certain number of accidents that could have been avoided if the operator was able to check the surroundings more properly. The demand of our customers for safety operation support systems is increasing every year. Therefore, the demand for a safe driving support system from customers to our company is increasing year by year.We therefore developed a technology that uses deep learning to enable real-time automatic human detection from images generated by the cameras on board the forklift and the operator is alerted accordingly. This human detection system for forklifts has been made commercially available. The system configuration and features are presented in this report.