| The Canada Grain Act imposes a zero tolerance for stored-product insects in grain. In Canadian Grain Commission inspection offices and terminal elevators, the Berlese funnel in which heat is used to extract insects from grain samples is the insect detection method for the incoming and export grain. This method is slow and inaccurate (especially for internal grain feeders) and can lead to viable insects being missed and subsequent cross contamination of stored grain in elevators. Dead and internally developing insects are potential sources of contamination in grain products and contribute to fragments in flour.; The efficiency of the soft X-ray method to detect insect infestations in grain, identify different grain types, and differentiate wheat kernels of different moisture contents was evaluated in this study. Infestations caused by different life stages of Cryptolestes ferrugineus (Stephens), Tribolium castaneum (Herbst), Plodia interpunctella (Hubner), Sitophilus oryzae (L.), and Rhyzopertha dominica (F.) in wheat kernels were detected using the soft X-ray method. Canada Western Red Spring wheat kernels uninfested and infested by different life stages of the insects were X-rayed at 15 kV potential and 65muA current. Algorithms were developed to extract histogram features, histogram and shape moments, and textural features using co-occurrence and run length matrices from the X-ray images. A total of 57 extracted features were used to identify uninfested and infested kernels using linear and quadratic function parametric classifiers, a non-parametric classifier, and a multi-layer feed forward back propagation neural network (BPNN).; The non-parametric classifier and BPNN correctly identified 95% of Canada Western Amber Durum wheat, Canada Western Red Spring, Canada Western Soft White Spring, barley, and corn kernels. The linear-function parametric classifier and BPNN identified more than 84% of infestations due to C. ferrugineus and T. castaneum larvae. The infestations by C. ferrugineus pupae and adults were identified with more than 96% accuracy and 97% of kernels infested by P. interpunctella larvae were identified by both the linear-function parametric classifier and BPNN. Kernels infested by different stages of S. oryzae and R. dominica larvae were identified with more than 98% accuracy by the linear-function parametric classifier and BPNN. The linear-function parametric classifier and BPNN performed better than the quadratic-function parametric classifier and the non-parametric classifier for the identification of infested kernels by different insects.; Using the Berlese funnel method, 67, 51, and 81% of first, second, and third instars of C. ferrugineus respectively, were extracted in 6 h. The same infested kernels were all categorized as infested by the trained BPNN. The soft X-ray method detected the presence of live larvae inside infested kernels. The linear-function parametric classifier identified kernels infested by external and internal grain feeders with more than 86% accuracy with the false positives of 34% identified as infested by the external grain feeders. The classification accuracies by the linear-function parametric classifier were 74 and 94% respectively, for uninfected and infested kernels (pooling kernels infested by external and internal grain feeders). |
