Multispectral inspection of citrus in real-time using machine vision and digital signal processors
Derechos de accesoopenAccess
MetadatosMostrar el registro completo del ítem
Citrus are one of the major fruits produced in Spain. Most of this production is exported to Europe for fresh consumption, where consumers increasingly demand best quality. Nowadays, Spanish producers have to compete with other countries with lower production costs. Moreover, inspection and classification tasks in these countries are made manually, which is subjective and varies among different experts or along the day. For these reasons, automatic inspection means, as machine vision, are a priority in Spain, in order to ensure products with an excellent quality. Current commercial sorters based on machine vision only solve the problems that require less computing time, as for instance, sizing or classification in colours. Sometimes they work with low resolution images, in order to achieve high processing speeds. However, this approach reduces the accuracy of the system when estimating the size of the fruit. Another important fact that needs consideration is the possibility of detecting defects on the skin surface using wavelengths that are outside the visible spectrum. This work includes the development of a multispectral camera, which is able to acquire visible and near infrared images from the same scene; the design of specific algorithms and their implementation on a specific board based on two DSPs that work in parallel, which allows to divide the inspection tasks in the different processors, saving processing time. The machine vision system was mounted on a commercial conveyor, and it is able to inspect the size, colour and presence of defects in citrus at a minimum rate of 5 fruits/s. The hardware improvements needed to increase the inspection speed to 10 fruits/s are also described. The experiments, carried out with oranges, mandarins and lemons, demonstrated that the software is able to single the fruit before estimating the size, which is calculated with an error less than 2 mm. To check the performance in colour estimation. mandarins in different maturity grades were used. Results compared with human classification allow 94% coincidence in the worst case (when the fruit is changing colour from green to orange). The system is also capable of correctly classifying lemons and mandarins, attending to the external defects in 93 and 94% of the cases, respectively, following the Spanish citrus standards. (C) 2002 Elsevier Science B.V. All rights reserved.