Logistics applications involve many labor-intensive operations which are error-prone. Foreffective management, human activity analysis can provide useful information, which helps (1) to enable much more flexible and efficient process checking, (2) to assess workers' safety, and (3) to reduce effect by human error and increase productivity. However, having round-the-clock manual monitoring is impractical. In this respect, employing new machine vision and sensing technologies in logistics applications will be helpful. This project aims to develop machine vision and sensing tools to automate human activity analysis for process checking and safety monitoring. Specifically, we will develop machine vision and sensing techniques for the following procedures: a. Pose tracking for activity analysis, e.g., toy manufacturing; b. Pose comparison for vocational training, e.g., load lifting. This project has been defined by wide request of local industry. With a number of new technologies incorporated into different aspects of logistics applications, the overall productivity and safety will be greatly increased.