Submitted to: Agroacoustics Symposium
Publication Type: Proceedings
Publication Acceptance Date: 6/8/1995
Publication Date: N/A
Interpretive Summary: Not Applicable
Technical Abstract: The computer-based ALFID System quantifies infestation of internally feeding larvae in a grain sample using spatial localization of sounds produced by insects in the sample. Localization is based on the relative arrival times of each emanating sound to sensors arrayed in a sample container. A previous implementation, using hardware threshold detection to ascertain arrival times, demonstrated conceptual feasibility but its accuracy was limited due to low signal to noise ratios and differential distortion of sounds propagating through the grain. This is addressed by an implementation which employs acquisition and software analyses of signals from the array of sensors. Cross-correlation analysis of adjacent sensors' outputs provides time-delay relationships which characterize each sound based on its source's location. A cluster analysis of these cross- correlation results is used to group multiple sounds emanating from the same source. Each valid group of sounds then indicates the presence of on insect.