الملخص الإنجليزي
Insect infestation is a serious problem in stored dates and it significantly affects the marketability of dates in domestic and international markets. At present, infestation is detected by.manual inspection method. It is not possible to detect the internal infestation by the current method. An accurate detection method for internal and external infestations preferably at early life stages of the insect would be highly beneficial to take corrective measures such as fumigation or cleaning. Therefore, the objective of this study was to investigate the capability of X-ray imaging in detecting internal infestations caused by saw-toothed beetle (Oryzaephilus surinamensis Linnaeus) in Fardh dates. Artificially infested dates were incubated at 30 + 1°C and 70 = 5 % relative humidity. A total of 40 sound and infested dates were x-rayed for different life stages (egg, larvae, pupae, and adult) of O. surinamensis. An algorithm was developed to extract 44 features (histogram and textural attributes) from the X ray images. Linear discriminant analysis (LDA) was used to discriminate the sound and infested dates using extracted features. The histogram features played an important role in the classification of sound and infested dates. An overall classification accuracy of 80% was achieved when life stages were treated separately (5 classes: sound, egg, larvae, pupae, and adult). While, discrimination efficiency increased to 87.5% when there were 2 classes (sound vs. infested dates; larvae, pupae and adult). In pairwise discriminant analysis, the highest classification accuracy (96.9%) was achieved between sound dates and dates infested with adults. Further studies are required to investigate the effect of moisture content and hardness of dates in the discrimination between the sound and infested dates.
