To estimate the moisture in an entire arbitrarily shaped workpiece, an automatic system should continuously sense moisture at all points of the workpiece. This paper proposes a design method for a microwavebased moisture sensing system for granular materials in arbitrarily shaped containers based on scanning in a moisture sensing box. The size of the moisture sensing box is first determined as the maximum size of any granular material containers based on initial user requirements. The moisture sensing system physically consists of a pair of antennas for microwave transmission and reception and a pair of distance sensors scanning throughout the maximum scanning range of the moisture sensing box to find the attenuation and workpiece thickness at all points of the workpiece, and these sensed microwave attenuations and workpiece thickness measurements are then used to calculate the moisture at all points of the workpiece. The average moisture of all points is finally taken as the estimated moisture of the whole workpiece.To evaluate the performance of the proposed system, experiments have been performed with granular materials in containers of various shapes, and the estimation results show the proposed method of continuous moisture scanning of a workpiece can improve 7.77% approximately compared with the antenna array approach.

No abstract

Because safety is important in everyday life, according to the National Retail Federation, organized retail theft accounts for an estimated 30 billion dollars each year in the U.S. Therefore, this paper proposes the detection system of theft walking patterns using motion-based Artificial Intelligence from a stationary camera. This method uses object detection by the Yolov5 model for localization of the objects, which is less affect to a background problem to detect the moving objects from the RGB video. After that, the coordinates of bounding boxes from Yolov5 are used to calculate the centroid of the box; then, it is used to make the motion of walking pattern behaviour. Finally, the sequences of the motion patterns are used with a bi-directional long-short term memory network, which is used for the time-independent pattern using 10-fold cross validation at training size 50% of total dataset. To evaluate the performance of the method, the datasets consist of normal walking people, which have 30 videos, and theft walking people, which have 30 videos. The accuracy of the system is approximately 97%.