Delivering Better Quality Rice: Color Sorters Made Possible through Mechatronics

One effect of climate change on rice is immature milky-white grains (called shirata in Japanese). When rice plants are exposed to high temperatures between ear emergence and harvest, some grains fail to fully develop and turn a chalky white color. What is more, damage from stink bugs and other factors can cause some grains to turn black. The presence of these defective grains in yields can lead to deductions during grading inspections. In fact, even just two or three defective grains out of 1,000 can result in a downgrade to second grade rice, lowering its selling price. For rice farmers, preventing the inclusion of defective grains is critical for operations, and increasing profitability demands production of high-quality, high-value rice, even when yields remain the same.

• 

  Good-quality brown rice (left) and defective rice for sorting (right)

At the same time, Japan’s total number of farm households continues to decline, and the workforce is steadily aging. Rice cultivation is usually limited to a single harvest per year, with processes such as harvesting, drying, and hulling concentrated into a short period of time. With the limited workforce, there is a growing need for more efficient and precise sorting of defective grains, as well as automation to make this possible.

In recent years, more and more abandoned rice paddies have been taken over by core farmers within the local community. Even if each field is relatively small, managing an increasing number of scattered fields can lead to inconsistencies in rice quality. As a result, the need to properly sort rice from these paddies and ensure a stable supply of high-quality products is expected to grow. Furthermore, by linking sorting data with FMIS, farmers can improve planning for future harvests.

Rice color sorters were originally designed for use in rice milling facilities. Kubota began developing and marketing its large-scale KG rice color sorter in 1998. Following the liberalization of rice sales in Japan in 2004, a new trend emerged: More farmers began selling their rice directly through independent distribution. Sensing the needs of rice farmers, Kubota developed a rice color sorter with the price, capacity, and processing capacity that made it accessible to individual farmers. In 2006, it introduced this system to the market as the KG-A series.

This had a huge impact, enabling farmers to establish their own full-fledged production workflows for drying, hulling, color sorting, weighing, and bagging. Since then, further improvements have been made, such as significantly reduced warm-up times through LED lighting and increased processing capacity, and the system has supported the dawn of the era of independent rice distribution.

• 

  KG-A Series

Rice color sorters have a three-step process: feed, detect, and eject. Rice is fed through the machine, a camera identifies defective grains, and bursts of air blow them out. However, this process has presented a drawback known as carryover, in which good grains are sometimes blown out with defective grains. To minimize this loss, farmers carried out secondary (return) sorting, feeding the rejected grains back into the machine to recover any good rice.

In addition, it was often difficult to remove all defective grains in a single pass. Sometimes two separate sorting processes were necessary: First to remove dark discolored grains, and the second to remove the immature milky-white grains (shirata).

As a result, secondary sorting became standard practice at farms to maintain high yields with less loss of quality rice. For farmers handling rice by the ton, this double sorting process required significant working hours and extra time. Recognizing that eliminating this secondary process was key to saving labor and improving efficiency, Kubota set out to develop a solution. Its efforts led to the rice color sorter in use today.

The key function of a rice color sorter is the accurate detection of defective grains. Previous machines performed sorting with two cameras, but achieving high-precision detection of both discolored rice and immature grains simultaneously required the use of three cameras. Much of the defective rice that remained unsorted was discolored grains, and this occurred when the darker portion of the grain was not facing the reflective camera. Therefore, Kubota added one more reflective camera to enable simultaneous detection from the front and back of the grains. The concept was to develop a detection system comprised of three cameras and two LED lights, green and red. This would be able to detect both discolored and immature grains with high accuracy in a single pass.

While this detection system improved accuracy, its biggest drawback was its cost, which had risen beyond the point where farmers could easily adopt it. To both eliminate the need for secondary sorting and make the technology more affordable, Kubota set out to develop its own digital line sensor camera. Drawing on experience gained in the late 1990s, when the company manufactured such cameras in the R&D process for rice color sorters for industrial use, it was able to manufacture all components in-house, minimizing the burden on farmers.

The next challenge lay in the mechanical design. The camera needed to be positioned and secured within a limited space while meeting a range of operational requirements. The optical axis, which is the path connecting the camera to the rice grains, is highly sensitive to the point that even a 0.1-degree deviation in the angle of an internal mirror can affect detection accuracy. However, while rice color sorters are subject to vibration while in use, they are frequently moved in agricultural settings and may be used in environments with significant temperature fluctuations. It was found that fixing the camera too rigidly could cause the mirror to crack under vibration or other factors. To address this, Kubota devised a mounting method that utilizes the restoring force of springs. Elasticity was intentionally incorporated into the design, including rubber in the sealant, and this resolved the issue.

In addition, rice dust (bran, etc.) is a major issue for cameras. To prevent contamination, the rice flow path and the camera housing were completely separated by glass. An automated cleaning system was incorporated to blow air across the glass surface every 20 minutes, ensuring consistently high detection accuracy even through glass.

This reflects Kubota’s hands-on approach to development, grounded in a thorough understanding of real-world conditions. The development team visited several farms and integrated mechanical and electronic technologies to deliver equipment that is both highly durable and easy to maintain.

The next challenge was addressing the issue of carryover, in which good grains are unintentionally ejected. Kubota hypothesized that by precisely targeting the center of each defective grain as it flows, the unintentional ejection of the good rice around it could be avoided. Previously, cameras could only detect the presence or absence of defects. The new system adds the capability to identify the beginning and end of each grain. In doing so, the system recognizes each grain separately and can calculate the center position of any grain judged to be defective. This way, it can target only the defective grains with high precision.

Detection of defective rice using three cameras and use of the air ejector had to be performed within a narrow space of about 30 mm where the chute carrying the rice is momentarily interrupted. First, the discharge position of the air ejector was fixed. Then, optimal upstream distance at which the cameras should capture the rice was calculated, and the cameras positioned. The control system was then designed to trigger the air ejection at the moment the cameras scan the rice.

Advances in CPU speeds, which enable processing of large amounts of data in short times, made it theoretically possible to shorten the distance between the camera’s detection point and the air ejector. However, given the physical constraints with the placement of optical components, Kubota repeated real-world testing with actual passes of rice, making fine adjustments in increments as small as 0.1 mm to approach the theoretical value as closely as possible.

In addition, maximizing the effectiveness of center-position detection requires physical conditions that prevent grains from overlapping. This depends largely on the design of the chute through which the rice flows, including its installation angle, chute length, groove shape, and surface treatment. Going beyond mere simulations, the team is passing actual rice through the system and identifying the optimal configuration.

For this rice color sorter, which provides feed, detect, and eject operations, Kubota’s full range of technologies have been brought together, including mechanical and structural design to move rice grains without them overlapping, optical and image-processing technology using digital line-sensor cameras, and electrical, electronic, and control engineering to judge images and control the air ejector that removes defective grains.

Through its innovations with this rice color sorter, Kubota has eliminated the need for secondary sorting on farms, significantly improving work efficiency. This has helped farmers provide stable shipments of high-quality rice – and boosted their profits.

Using the data gained from the rice color sorter, farmers can do more than determine rates of defective grains; they can also identify the type of damage, whether from stink bugs, heat stress, or other causes. By linking this data with FMIS, they can visualize defect trends for each field, leading to revisions in cultivation plans, such as optimizing water management and pesticide use for the next season’s harvest. For farmers, utilizing this data provides both medium- and long-term benefits in terms of farm management enhancements: reducing costs and labor, producing higher-value rice, lowering environmental impact, and more.

For rice farmers, the rice color sorter is the final line of defense before their carefully cultivated rice is packaged for sale. Increasing its sorting sensitivity improves quality, but it also raises the rejection rate and reduces shipment volume. And lowering its sensitivity increases volume, but it can also compromise quality and grade. Faced with this dilemma, many farmers struggle to find the right balance for scanning sensitivity, and each year’s sorting work is often adjusted based on intuition.

Kubota is researching and developing a system that will allow customers to adjust sensitivity with ease. Its edamame sorter, inspired by its rice color sorting technology, has already been outfitted with an automatic sensitivity adjustment feature. Kubota will link its wide range of technologies and expertise developed through its agricultural machinery R&D with on-site needs to further enhance performance.

The goal is to enable farming where the rice color sorter can work without humans nearby. Kubota will work to minimize the time and effort needed by farmers while improving sorting accuracy and profitability, contributing to sustainable rice production.