Smart Snickometer for assisting Cricketing Umpires

This paper proposes a novel technique for designing a smart Snickometer for assisting on field cricketing umpires and evaluates its performance for combination of different classifiers and features. Snickometer are used to analyze weak signals coming from bat/pad/gloves when incoming ball hits any part of them. The sound generated are captured from high sensitive microphone attached to the stumps and are synced with video for offline analysis. A decision is made by simultaneous analysis of both video and oscilloscope in slow motion. The traditional method involves manual interpretation and may lead to erroneous judgment. The entire verification process sometime takes too much time. To overcome the limitation of the current working model, a Smart Snickometer is proposed. Data are collected in real time environment. Feature extracted from these data are fed to different classifiers and their performance is reported. Best classification accuracy of 95.83% is achieved with both K-Nearest Neighbour (KNN) and Support Vector Machine (SVM). Maximum Amplitude and Peak Energy are best performing features.