Your Cart

Your cart is empty

Home  >  Volume 45

48. AUTOMATED DETECTION OF SOLANUM LYCOPERSICUM (TOMATO) SIZES USING IMAGE PROCESSING TECHNIQUE by Abayomi-Alli A.,Arogundade O. T., Abayomi-Alli O. O., Oyemade S. V.1 and Sijuade A. A. Volume 45 (March, 2018), pp 357 – 368
Sale price: $5.00


The size assessment of Tomato fruit like other horticultural product is carried out by human inspection which suffers a high degree of subjectivity due to psychological factors. In this study, an Automated Tomato Size Detection System (ATSDS) was proposed and developed using image processing techniques. Primary data source was from 594 tomato fruit images of varying sizes collected in a controlled environment. ATSDS successfully carried out image segmentation and area extraction to obtain the Area of Tomato Fruit (ATF) and categorize the tomato fruits into very large, large, medium, small and very small Tomato Size Category (TSC) being Tomato Fruit Sizes (TFS) of 5,4,3,2 and 1, respectively. Results from ATSDS on the tomato images showed that average TFS=2.68, with 4 very large tomatoes, 75 large, 267 medium, 224 small and 24 very small tomatoes, respectively. Nine observers were assigned to manually conduct a size assessment of the images using the Mean Opinion Score (MOS) to obtain a Subjective Tomato Fruit Size (STFS). Experimental results showed average STFS=2.72, with 26 very large, 84 large, 266 medium, 214 small and 24 very small tomatoes. TFS and STFS had a linear correlation result of 0.97. For future direction, the automated tomato size detection system should be wrapped into a model and made into a mobile app.

Keywords:  Automatic, Correlation, Detection, Mean-opinion-Score, Segmentation, Tomato, Visual inspection.