This paper describes a method that can be used to monitor the human posture in industrial environment. The method is based on the simultaneous use of a 3D camera (Microsoft Kinect V2) and a wearable motion capture system that uses inertial measurement units to identify the body posture (Notch Wearable). The data fusion at feature level allows overcoming the intrinsic limitations of the two methods, deriving from occlusions and electromagnetic interferences, respectively. This work describes the algorithms implemented for the data validation, the calibration of the two measurement systems performed in controlled environment and experiments performed in tailoring operations. The first step of our analysis was the implementation of algorithms used to validate the skeleton measured by the Microsoft Kinect, that is unreliable in presence of occlusions. The calibration of the Kinect and Notch systems was performed by asking to different subjects to repeat movements of the upper limbs and evaluating the measurements repeatability and reproducibility. Finally, experiments were performed in different tailoring operations and the cumulative probability density function of different body angles was evaluated to compute the biomechanical loads of the workers.