The sensors will determine the activity base on movement and heat signature of the rodent, it will then trigger an alert of the trigger to the gateway through wireless communication, which with then forward the data to the cloud and to the back-end.

Direction analytics were developed to provide information for rodent pathway derivation and estimation of rodent population counts. This was applied on to the frames with 'blobs' detected in the thermal videos captured in the SIR sensor (Figure1).

The moving direction of rodent was classified into five categories; up, down, left, right, and no motion. No-motion in this instance was defined as no noticeable movement of rodent blob detected by algorithms during the captured frame intervals. Referring to Figure3, assuming that the blog centroid coordinates in the first and last frame of a successive sequence as denoted as (x1,y1) and (xn,yn) respectively. the distance and orientation of the moving vector of the blobs are obtained as:

Based on the computed relative orientation angle θ, the moving direction was classified into one of the 5 possible directions (Table 1, Figure 3). The SIR were installed in such a way that the path along its Up and Down direction is primary travelling path to be monitored.