Kazunori Ninomiya

Automated fruit grading system using image processing

This paper describes the operations and performance of an automated quality verification system for agricultural products and its main features. The system utilizes improved engineering designs and image-processing techniques to convey and grade products. Basically two inspection stages of the system can be identified: external fruit inspection and internal fruit inspection. Surface inspection is accomplished through processing of color CCD images while internal inspection employs special sensors for sugar and acid content. An X-ray sensor is incorporated to detect biological defects.

Features Extraction for Eggplant Fruit Grading System Using Machine Vision

Machine vision based grading for agricultural crops has been well developed and accepted as an attractive grading method. However, machine vision based grading for eggplant fruit is not available yet. This study reports on the attempt to develop an eggplant grading machine using six CCD cameras as the sensing device. Feature extraction algorithms were developed to extract eggplants features, i.e., length, diameter, volume, curvature, color homogeneity, calyx color, calyx area, and surface defect. The system could acquire six images per fruits covering the entire surface of the eggplant fruits. An agreement rate of 78.0% was achieved in the feasibility study where the machine vision based grading was compared with manual grading. The throughput of the developed system was 0.3 second per fruit. Details of the system, an outline of the algorithm, and performance results are reported in this article.

Machine Vision Algorithm for Robots to Harvest Strawberries in Tabletop Culture Greenhouses

A strawberry harvesting robot consisting of a four DOF manipulator, an end-effector with suction pad, a three camera vision system and a rail type traveling device was developed as a trial to conduct experiments in a tabletop culture greenhouse. In order to harvest the strawberries with curved or inclined peduncles, a wrist joint which can rotate 15 degrees to the left or right from its base position was added. On the algorithm side, peduncle inclination angle was measured by the center camera. Harvesting experiments show that it was possible to precisely harvest more than 75% of fruits which were not occluded by other fruits with the developed robot. Experimental data also show that peduncle length, color and inclination pattern change with the seasons. Complex situations often exist in the real field conditions such as limited visibility of back end strawberries, occluded fruits, obstructions and complex peduncle patterns. Further studies are desirable to automate the harvesting task using a robot.

Surface Gloss Measurement on Eggplant Fruit

Surface gloss is one of the important quality factors affecting consumers opinion on purchasing eggplant fruits. The current manual gloss grading method is subject to human bias and the outcome is often very ambiguous. This study developed a sensing device consisting of two CCD monochrome cameras and three long tungsten lights and an associated quantitative method of measuring surface gloss of eggplant fruit using the device. Distinguishing various glossiness states of the fruit were made possible by analyzing the specular reflectance of the fruit surface. Specular reflection of the light on fruit surface was measured by digital image analysis method. Gradient filter was applied on the gray-level image for measuring magnitude of intensity changes. The algorithm consistency and validation test on the samples showed the algorithm was feasible of quantitatively assessing eggplant fruit surface gloss with an accuracy rate of 0.806. Details of the sensing devices, outline of the algorithm and preliminary results are presented in this article.

Application of NIR-color CCD camera to Eggplant Grading Machine

An on-line eggplant grading machine was developed to inspect and grade fresh market eggplant in an agriculture cooperative located at Okayama, Japan. Two machine vision systems, which made up from 6 color CCD cameras and 4 monochrome CCD cameras were used for acquiring digital image of the eggplant. Eggplant fruits are graded when the fruits were conveyed through these cameras on a special designed rotary tray. 180 ° vertical turn of the rotary tray in between these camera boxes enable the inspection of the eggplant’s entire surface. It was found that disorientated and disposition fruit on the rotary tray affect the grading process. Eggplant fruit is dark purple in color with extremely low spectral reflectance in the visible spectrum. Consequently, defect detection on the eggplant fruits and extraction of fruits feature from the low color contrast background are difficult. A new NIR-enhanced-color CCD camera (380nm-1400nm) was studied in overcome the problems mentioned above. Experimental result showed that this new camera was able to extract the fruit’s feature from dark background and detection of low-contrast-defects was found possible too.

Strawberry harvesting robot for fruits grown on table top culture

We have been developing strawberry harvesting robot for harvesting in a table top culture greenhouse. A strawberry harvesting robot consisted of a 3 DOF manipulator, an end-effecter, machine vision system and a traveling device. An end-effecter, which consisted of a suction head connected to a blower and tow fingers could hold on fruit by suction head and cut and grasp peduncle by two fingers rotate according to peduncle inclination. Machine vision system, which consisted of three identical color cameras (stereo vision system and center camera) and five LED lighting device could calculate fruit’s location of 3 dimension and recognize target fruit and peduncle details. From harvesting experiments at May 2007, it was observed that 38% of fruits were harvested. Corresponding problems were miss-stereo matching and false recognition of fruit and peduncle.

Feasibility on the quality evaluation of agricultural products with terahertz electromagnetic wave

This study is aims at an application of terahertz (THz) waves to nondestructive evaluation of agricultural products. An experiment system with THz electromagnetic waves working in reflection geometry was built and applied to the quality evaluation of tomato fruits. This system was capable of nondestructive detection of defects derived from moisture difference on the thin surface layers of the tomato, which was difficult with visible or NIR light. From the experimental results, it was found that the assessment of the internal damage is feasible. Especially, it was clear that the difference between the damaged parts and normal parts could be easily distinguished by using THz waves in reflection. The complex refractive index of sugar solutions was also measured using the THz Time Domain Spectroscopy (THz-TDS) for obtaining its correlation with the measurable sugar content (Brix %). From the result, it was observed that higher sugar concentration relates to a lower absorption of the THz waves and that the real part of the complex refractive index did not change significantly, contrary to what is observed in the visible region.

Patterns of Fluorescence Associated with Citrus Peel Defects

Unshu citrus were sorted by fluorescence imaging in a commercial packinghouse and undamaged-appearing unshu that had been rejected by the packinghouse due to fluorescence appearing on their peel were studied. We examined the various visible patterns, based upon fluorescence and microscopic images, to provide a categorization of physical reasons for the observed fluorescence. The categorization classes were: 1) slight physical damage: thin scar, hole and flow, shrunken at calyx; 2) rubbing against decayed fruits; 3) green spots; and 4) rind puffing. The percentage of observation for each of the four classes was 22 %, 15 %, 42 % and 21 %, respectively. Storage of the classes indicated that, except for the green spot class, the injured area expanded quickly and caused the fruits to rot within a week.

Machine Vision System for Detecting Fluorescent Area of Citrus Using Fluorescence Image

Abstract This research is carried out to develop a machine vision system which could identify the fluorescence area on injured or defective citrus surfaces. The target objects whose surfaces were injured by needle insertions were acquired by a camera VGA using UV lamps (radiating Blacklight and UV-B wave-lengths) and white LEDs. Because damaged citrus peel includes fluorescent substances, it was easy to discriminate fluorescence parts from healthy parts. The results showed that the blacklighting system is practical and feasible, and that the proposed algorithm of fluorescence area detection is effective for some varieties of citrus.

Development of Multi-Product Grading System

It is anticipated that grading system of agricultural product, which gives us many kinds of information such as size, color, shape, defect, and internal quality, will be important from the view point of traceability in future. Many grading systems have been developed and practically used for fruits and vegetables in Japan. They play roles to be substituting for human labor with precision by use of machine vision, NIR analysis, and automation technologies. Japanese farmers often produce various agricultural products in small quantity for short seasons annually, while the high capacity grading systems are used for specific products therefore the annual operating period is very less. It is becoming obvious that the multi-product grading system is needed. In this paper, a multi-product grading system is developed. Initially, fruits or vegetables such as tomato and orange will be arriving randomly through the conveyor up-to a CCD color camera fixed at a height of 80 cm. Based on the Information through this camera robot will move to pick the object by using suction pad. Then, robot carries the fruit to grade it by using two CCD color cameras. One is positioned horizontally and the other one vertically upside-down. The horizontal camera is used to predict the color percentage, bruise and other defects on the surface of the fruit. Whereas bottom camera does the processing for the bottom part of the fruit, such as calyx and any defect affecting the grade of the products. Surface and bottom information of fruit will be given as input to the developed neural network. Three grading levels i.e., grade A grade B and grade C are assigned as output to the neural network. Thus three boxes were kept in specific order to place the fruits. This paper shows that neural network can be applied as a tool to grade the fruits with good accuracy. Future study is needed to improve the proposed grading model, to develop the grading methods for eggplant and process the information to implement the traceability of fruits in the system.