Search

Position and orientation data are critical for guiding and managing cooperative robot tasks in the greenhouse. Errors, however, associated with current positioning and orientation systems, such as indoor GPS compasses or the sensitivity of magnetic direction sensors to magnetic materials, limit their application in greenhouses. In this research, we develop and evaluate a positioning and orientation system using a sound signal and two different receiver architectures in a greenhouse. The results obtained: a positioning accuracy of 23.0 ± 5.5 mm and average orientation accuracy of 3.43 ± 1.67° and 1.58 ± 0.86° for these two methods, demonstrate this spread spectrum sound based system has the potential to be deployed on small distributed greenhouse robots.

A spread spectrum sound-based local positioning system (SSSLPS) has been developed for indoor agricultural robots by our research group. Such an SSSLPS has several advantages, including effective propagation, low cost, and ease of use. When using sound velocity for field position measurements in a greenhouse, spatial and temporal variations in temperature during the day can have a major effect on sound velocity and subsequent positioning accuracy. In this research, a temperature-compensated sound velocity positioning was proposed and evaluated in comparison to a conventional temperature sensor method. Results indicate that this new proposed method has a positioning accuracy to within 20 mm in a 3 m &times, 9 m ridged greenhouse. It has the potential to replace the current system of using the temperature sensors in a greenhouse.

Quadcopters can play an important role in precision agriculture; mostly having been researched and commercialized for the open field conditions. While greenhouse applications have been limited by the positioning system. Conventional GPS positioning systems are difficult to use in greenhouses. Spread Spectrum Sound-based Positioning System (SSSPS) on the other hand, has the potential to be used on a quadcopter in a greenhouse. In this research, we evaluated whether acoustic noise from a quadcopter over wide frequency range would interfere with the SSSPS operation. We set a microphone on the quadcopter, and evaluated the SSSPS system for quadcopter noise tolerance with several signals. The results demonstrate the navigation system can tolerate quadcopter noise over distances up to 15 m between the tweeter and the microphone on the quadcopter with the RMSE of 5.9 mm.

Indoor navigation plays an important role in agricultural robots working in greenhouses. The SS-sound system, where Time of Arrival (TOA) of spread spectrum modulated sound is used to localization, is one of the effective methods for indoor navigation, but there is a called near-far problem where the sound waves emitted from distant anchors are disturbed by the one emitted from nearby anchors. Transmitting the SS-sound from multiple anchors using Time Division Multiple Access (TDMA) is effective to solve the near-far problem. However, the TOA from multiple anchors cannot be obtained simultaneously, so that the localization cannot be performed. To solve the problem, we propose a method for combining the TDMA-based SS-sound system and inertial navigation system. A simple numerical example simulating the movement of a ground-moving robot was presented to clarify the effectiveness of the proposed method.

Traditional greenhouse robots are heavy and expensive, so a small size with low cost platform is needed. When operating small distributed robots in a greenhouse, a positioning and orientation system will be needed for the accurate control. While the traditional methods of determining orientation have limited application in greenhouses due to the errors associated with indoor GPS compasses or the sensitivity of magnetic direction sensors to magnetic materials in the greenhouse. In this research, we develop methods for determining orientation using a sound signal and two receivers. These methods were then experimentally evaluated in a greenhouse with and without manually removing distance reflection data. After deleting these signal reflection errors, the average accuracy of the two methods was determined to be 3.1° and 2.8°, respectively.

Small distributed robots have the potential to be commercialized in greenhouse operations due to their low cost, high efficiency and light weight compared to conventional greenhouse robots. In order to guide the path of these distributed dynamic robots, a positioning system with a centimetre accuracy is needed. While a GPS navigation system does not perform well in such an environment, the passive Spread Spectrum Sound-based Local Positioning System (SSSLPS) using FDMA (Frequency Division Multiple Access) does have the potential to accurately localize robots in such an environment if certain limitations can be overcome. One such issue is the Doppler effect. In this paper a Doppler shift compensation (DSC) algorithm, which applies a carrier frequency detection method, was used for passive SSSLPS (FDMA), and evaluated in terms of its ability to compensate for the frequency shift of the sound signals when measuring distance and position. The results show the applied DSC algorithm can measure distance using a compensated reference signal at 0.1–0.8 m/s constant velocity and 0.06–1.06 m/s2 acceleration. The positioning error of the passive SSSLPS (FDMA) mounted on the crawler robot was 50.3 ± 22.3 mm accuracy, while the failed detection rate was 2.6%. These results suggest the applied DSC algorithm can effectively compensate for Doppler shift when using passive SSSLPS with FDMA for dynamic robot positioning.

To apply Spread Spectrum (SS) sound-based positioning system in green house, measurement accuracy was investigated by using Base- Station (BS) for compensating temperature. In this experiment, normalized correlation value and correlation ratio were used to detect correlation peak of direct wave of SS sound. When 9 positions in actual greenhouse were measured, success rate to detect direct wave of SS sound was 99%. As the result, this correlation peak detection method is useful to detect correlation peak of direct wave. Average measurement error at each position was 55 mm and improved about 20 mm by using BS method. Moreover, when center correlation peak was detected, average measurement error was 15.5 mm. Thus, BS method enables SS sound-based positioning system to be useful in green house.

Conventionally, biochemical analysis is used to estimate the freshness of fish products, but this method is destructive, technically demanding and time consuming. In this study, the fluorescence properties of raw fisheyes of Japanese dace (Tribolodon hakonensis) fish were investigated using spectrofluorometer during 4 °C storage. Subsequently, according to the results of the spectrofluorometer, fluorescence imaging system was established with 365 nm UV LED to capture the fluorescence images using a common color camera. The color components from fluorescence images of pupil and iris were extracted to establish a prediction model for a ‘standard freshness index’ of K value which was acquired by using biochemical analysis using partial least square regression method. From the results, RMSECV of 3.5% and R2 of 0.92 was realized. This study demonstrated that the UV-fluorescence imaging technique has a potential to estimate the fish freshness non-destructively during storage.

Halation is a serious problem for machine vision systems which can cause loss of color information from an image. Images of tomato fruits grown in a greenhouse taken with a different rotational angle polarizing filter (PL filter) were acquired and combined into one resulting image to eliminate multiple halation areas due to sunlight and reflected light. An image acquisition system composed of a CCD camera and a stepping motor was used to acquire the images. The multiple halation areas could be eliminated at a certain angle of incidence and the color representation of the resultant image was improved. The results demonstrate the feasibility of eliminating multiple halation areas using this method.

An image segmentation method to calculate pupil sizes of cattle was developed.Two cattle were identified as severe vitamin A deficiency by our method.Cattle had smaller resting pupil in the dark during vitamin A deficient stage. To produce beef with a higher marbling standard, Japanese black cattle farmers usually actively attempt to lower the serum vitamin A level in the fattening period to a minimum desired level (about 30IU/dL). However, early identification of suboptimal vitamin A deficiency in the cattle is important to prevent them from becoming susceptible to contracting serious diseases. In this study, we investigated the feasibility of using Pupillary Light Reflex (PLR) analysis to identify vitamin A deficient cattle during this fattening period. PLRs of 43 cattle were recorded monthly from June 2012 to February 2013 using a 2CCD camera based handheld machine vision system. A new image processing algorithm to segment the pupil from the background was developed. Compared with manually selected results, the root mean square error associated with the constriction amplitude (CA) acquired by the image processing algorithm was only 2.3%, indicating the effectiveness of this algorithm. No significant differences were found between results of CA from fattening cattle in the high (>60IU/dL) and low (

A spread spectrum sound-based positioning system is developed for an open field environment. Errors associated with wind are a major challenge for achieving high positioning accuracy with such a positioning system. Here, wind compensation using a base station configuration is proposed. Field experiments show that by using the proposed compensation method, the influence of wind can be minimized and high positioning accuracy can be achieved. Within a 30 m × 30 m area, the average two-dimensional (2D) positioning accuracy was within approx. 20 mm, at an updating frequency of 2 Hz. Optimization of the base station position was also investigated.

Department of Farm Power and Machinery, Bangladesh Agricultural University, Mymensingh-2202, Bangladesh(Received: 20 March 2012. Accepted: 10 October 2012)In order to develop fully-automatic grafting robot for tomato seedlings, the machine vision system forsorting grafting seedling and inspecting grafted seedling was studied. In this paper, the devices anddifferent lighting (Front lighting, Backlight, and both Front lighting and Backlight) of machine visionsystem were optimized and selected. The result shows that Blue Backlight without filter device isbest for both sorting and inspecting seedlings. Then the images of sorting and inspecting seedlingswere acquired using the selected machine vision system and processed by the program. Experimentresults show the success rate on processing the images of sorting seedlings can reach 96% andeven the gap of 0.2 mm between scion and rootstock can be found easily when inspecting graftedseedling.

Fluorescence imaging has been used to detect peel defects in many citrus varieties, but not in lemon and yuzu. In this study, we extract and identify fluorescent components from the peel of lemon and yuzu. The characteristics of excitation and fluorescence of these extracted components were determined, and their identity clarified using NMR spectroscopy. Extracted fluorescent components for both lemon and yuzu have a coumarin structure. Two coumarins (5-7-dimethoxycoumarin,5-geranyloxy-7-methoxycoumarin) were identified in lemon and one (5-geranyloxycoumarin) in yuzu. Their excitation and fluorescence wavelength were observed at 320–330 nm and 380–400 nm, respectively. Moreover, when a fluorescence imaging system was set-up with UV-A illumination, it was shown that this systemcan detect peel defects in lemon, and to a lesser extent in yuzu.

A serious problem in Japanese agriculture is the aging and shortage of farm laborers. Agricultural robots are expected to solve or alleviate this problem, as well as provide additional information about agricultural products. In this research, we focus on machine vision for an asparagus harvesting robot. Although TV cameras are often used for machine vision in harvesting robots, it is difficult to use them in direct sunlight, due to fluctuations in illumination intensity and color temperature effects of sunlight. So a laser sensor was used. This sensor detected asparagus spears based on their distance from sensor and their length. Total success rate of detection was 75 % of asparagus spears and 71 % of parent asparagus.

There is a promising potential to use sound-based positioning system as low cost platform for automation and robotic system in agriculture. It can provide high positioning accuracy with relatively low cost. For practical use, it is important to make deployment of this system become simple and easy. Therefore, self-calibration method is highly demanded to replace manual calibration which is laborious, impractical, and time consuming. Here, we proposed a self-calibration method for sound-based positioning system in greenhouse. Experiment and numerical simulation were conducted to evaluate the performance of the proposed method. Experimental result shows that by using the proposed method, calibration of the sound-based positioning system in 7 m x 16 m area can be done with relatively small error, about 2-5 cm. We also investigated the performance of the proposed method under different greenhouse conditions, to see the influence of wind and temperature. Although there are some limitations, this study shows the feasibility to use the proposed system in greenhouse or other places with relatively flat floor.

A study was conducted to investigate the best wavelengths for fluorescence excitation and the resulting fluorescence wavelengths in the range of 300-700 nm with citrus peels taken from 15 varieties, which are concerned with detection of surface defects of citrus fruits. Characteristics of UV absorbance, excitation and fluorescence spectra were observed by variety, and principle components analysis (PCA) and discriminate analysis (DA) were used to categorize the citrus varieties by fluorescence intensity levels in order to give some tips for optimizing the fluorescence imaging based machine vision system. The observed proper excitation wavelength for best fluorescence emmision and resulting peak fluorescence wavelength varied variety to variety and ranged from 350 to 380 nm and 490 to 540 nm respectively. The selected varieties of citrus were categorized successfully into four groups of known fluorescence level, namely strong, medium, weak and no fluorescence groups.

A mobile grading machine for citrus fruits has been developed to collect crop information such as fruit yield, diameter, and sugar content of fruits of each tree part. It consists of a mobile mechanism, a differential global positioning system, a fruit conveyer system, a color camera for machine vision, a NIR spectrometer, and a personal computer for control and database. Preliminary field tests were conducted to investigate basic performance of this machine in a mandarin orange orchard. Using the collecting data of fruit yield and diameter and sugar content of fruit, crop information maps of each tree part was made. These maps indicated that each tree part has large variability of yield, size, and sugar content of fruit.

The relationship between serum vitamin A level and tapetum reflection was investigated using ultraviolet imaging. The tapetum was obtained from eyes of slaughtered cattle in which the serum vitamin A level was below 50 IU/dl. The tapetum reflection was examined in the tapetum lucidum, tapetum nigrum, and optic disc. For each part, the tapetum reflection decreased with increasing serum vitamin A level. A high tapetum reflection value might represent the bleached tapetum symptom that occurs because of the low serum vitamin A level.

Using a machine vision with backlighting LEDs, a tomato-seedling grading and sorting algorithm was developed for a fully automatic grafting robot. A UXGA camera, a blue color backlighting device and light filtering devices were selected to acquire seedling images. The developed algorithm was used to determine the bending, nodes of leaves and stem diameter of seedlings from their images and then to grade and sort them as the initial task of the grafting robot. Results showed that the sorting success rate was 97% and the rest 3% was failed as the target portions of the images of seedlings were covered by irregular arrangement or bended leaves from all sides.

GPS has become a common technology for a positioning system of vehicles outdoor. In agriculture, the RTK-GPS has been used for autonomous tractors and harvesters in researches, and their measurement accuracy was sufficient. In room, however, the GPS measurement errors were larger because of delayed or covered electric waves. Rail or electromagnetic induction type position sensing methods have been used for autonomous mobile robots under greenhouses or other facilities. There are problems that installation costs are expensive and route modification is not easy, when those systems are used in agricultural facilities. In this research, a positioning system using spread spectrum sound is described and noise tolerance and signal identification could be improved by using a spread spectrum technique. So far, a system for measuring distance was developed and estimation of measurement accuracy was conducted. As a result of measurement distance 10m between microphone and speaker under non wind and non temperature gradient, RMSE 3.4mm was obtained. It is also known, however, the measurement error increases under windy and changeable temperature conditions. In this research, a correction method of the measurement error by wind and temperature is developed and estimated.

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.

A mobile grading machine was earlier developed to establish precision citrus fruit production. This machine could travel in citrus orchards and measure appearance and internal qualities of fruit harvested by humans on-site. In addition, it could record harvesting date and fruit locations. It consisted of a mobile mechanism, a differential global positioning system, a fruit conveyor system, two machine vision systems (one for fruit inspection, the other for trees monitoring), a sugar content measurement system using a near-infrared (NIR) spectrometer, and a personal computer for control and database. The color, size, and sugar content of each fruit and leaf color of each tree were measured by the machine vision devices and the NIR inspection system. As fruit features were extracted based on image analysis by the machine vision system, it was observed that the result of size had a correlation with actual citrus size and the result of color was accurately compared with human eyes. However, it was considered that the lighting device positions should be changed for higher quality images, because the edges of the fruit were dark due to radiation angles. This paper reports on moves of the lighting device position and to obtain lower light radiation angles so that the estimate of fruit size could get higher correlation with the actual fruit size. Furthermore, the earlier results of sugar content measurement had a low correlation with the actual brix values because the sensing head of NIR inspection system was pressed to each fruit from the top. The pressing direction was changed to from the side to obtain higher performance of the system. The improvements caused higher correlations of the sensor outputs to the fruit size and sugar content.

Marbling score is an important index of beef quality. Vitamin A (VA) in blood inhibits differentiation of preadipose cells into adipose cells. High VA level lowers the marbling score of beef. Therefore, breeders are trying to optimize the blood VA level to improve marbling score (ca. 30 IU/dl). Monitoring the blood VA level is essential, however, a currently available method of blood VA level is blood sampling which is expensive, invasive, and time-consuming. Therefore, an alternative method is desirable. In the previous study, there was a linear relationship between the blood VA level of cattle and the light intensity at 470nm which was reflected from cattle eye. The objective of this study was to investigate the relationship between the blood VA level and the color of cattle eye, especially the pupil color. The final goal of this project is to develop a non-invasive monitoring system of VA level in cattle blood using image processing technique. The image of cattle eye was captured using a CCD camera and analyzed with WinRoof software. The relationships between the blood VA level and the mean RGB value or chromaticity value, “rgb„, or HIS value of pupil were examined for twenty-five cattle with different blood VA level which were ranged between 15 IU/dl and 120 IU/dl. The image of cattle retina was also captured and analyzed in the same way. The blood VA level of test cattle was ranged between 15 IU/dl to 105 IU/dl. Comparing the pupil data with the retina data, Intensity value (I) showed a similar tendency for both pupil and retina images.

There are many types of citrus fruit grading machine with machine vision capability. While most of them sort fruit by size, shape and color, detection of fruit rot remains challenging because their colors are similar with normal parts. Objectives of this research were to investigate if fluorescence would be a good indicator of the fruit rot, and to develop an economical solution to add the rot inspection capability to an existing machine vision fruit inspection station. A machine vision system consisting of a pair of white and UV LED lighting devices and a color CCD camera was proposed for the orange fruit grading task. Since the time lag between the color and fluorescence image captures was short (14 ms), it was possible to inspect color, shape, size, and rot of a fruit on the move before it leaves an existing industrial inspection chamber.

We investigated the pattern of stem elongation in response to temperatures in the light and dark periods in 6 bedding plants and 2 potted plants. The plants that responded well to different day and night conditions (DIF) could be classified into 2 groups. In the first group comprising plants Northpole and Verbena, the stem elongated mainly in the light period, and in the second group comprising plants chrysanthemum and Habotan, the stem elongated mainly in the dark period. Moreover, in the case of Zinnia, which exhibited an opposite reaction to DIF, stem elongation was observed in the dark period as the temperature increased. In the case of French marigold, Snapdragon, and Stock, which were not affected by DIF, stem elongation was not observed even when the temperature in the light and dark periods was varied.

In this study, sound-based ranging system in greenhouse environment with compensation of measurement error caused by multipath effect using Artificial Neural Network (ANN) was proposed. Greenhouse environment has special characteristic which is different with condition where the similar system that previously developed were applied. There are challenges need to be handled in developing accurate ranging system in greenhouse such as high humidity, temperature gradient, wind, and obstacles. In this study, error compensation was performed by first estimating the measurement error using some features extracted from the cross-correlation wave of the received signal by using ANN. Then, the estimated value was used to compensate the measurement error. The experiment result showed the feasibility to apply the proposed method to improve accuracy of sound-based ranging system in greenhouse environment.

Defective unshu oranges ( Citrus reticulate Blanco var. unshu ) were sorted based on fluorescent imaging technique in a commercial packinghouse but fresh appearing unshu were rejected due to fluorescence appearing on their peel. We studied the various visible patterns based on colour, fluorescence and microscopic images, where even areas of the peel that are not obviously damaged can have fluorescence, to provide a categorization of fluorescence reasons. The categorization corresponded to: 1) hole and flow; 2) influenced by damaged or rotten fruits that have released peel oil onto it; 3) immature or poor peel quality; 4) whitish fluorescence due to agro-chemicals and 5) variation of the growing season. The identification of such patterns of fluorescence might be useful for citrus grading industry to take some initiatives to make the entire automated system more efficient.

Most of mandarin oranges (Citrus Unshu) are manually harvested and sorted sorted based on size, color, shape and internal qualities at cooperative grading facilities in Japan. 80% of the oranges are marketed as fresh fruits. Since higher quality oranges can be given higher prices, it is important to produce high quality fruits and to grade them properly. It is known that fruit quality in each tree is different from place to place because environmental conditions are different even in a same orchard. There are not many researches on this variability from a viewpoint of precision agriculture in Japanese orange production. In this research, the 2-D map and 3-D map were developed in order to create the visualized maps of features on color, size, sugar content, yield, canopy size, and others for optimum tree management by measuring the qualities of the harvested fruits using a mobile citrus fruit grading machine. This machine could travel in citrus orchards and measure appearance and internal qualities of fruits harvested by human in front of citrus trees. 2-D maps can show a relationship between 2 types of data. 3-D map can show the relationships among 3 features. It was observed that 2-D maps were effective for overlaying aerial images and a soil information map and that 3-D maps were effective for tree management changing with complicated environmental conditions. In conclusion, field maps using circles and cylinders were effective for tree management in Japanese agriculture.

There are a lot of researches on position determination methods in room such as wireless LAN, RFID, infrared ray, TV camera, sound wave and etc. Among them, it is considered that a positioning system using sound has some merits of high measurement accuracy and low cost. In our research group, a positioning system using spread spectrum sound has been developed. So far, a system for measuring distance was developed and evaluation of its measurement accuracy was conducted. As a result of measurement under a condition of no wind, no temperature gradient and no noise, RMSE of 10m distance between a microphone and a speaker was 3.4mm. However, there are various noises, winds and temperature conditions in actual facilities such as fruit grading facilities, greenhouses, and dairy farms. It is, therefore, important to improve noise tolerance and to compensate changeable sound velocity by wind and temperature. In this paper, evaluation of noise tolerance was conducted by using three type sounds.

Serum Vitamin A level is important indication to control the beef quality in Japan. The beef quality depends on serum vitamin A (VA) level. From 16 months to 24 months, serum VA level should be maintained low level and kept at about 30 IU/dl to produce the high quality meat.It is need to control serum VA level without cattle diseases. Generally, serum vitamin A level is mainly detected by blood assay. However, it is time-consuming, expensive, and stressful to the cattle. Therefore, a non-invasive, rapid, and simple method for monitoring serum VA level is desirable to produce the high quality beef. In the previous study, pupil reflection was related to the serum VA level. The photosensitive pigment rhodopsin related to serum VA has an absorption about 380nm, so it would appear that there is the relationship between pupil reflection in 375nm-wavelength and VA level in blood. The objective of this study was to predict serum VA level with pupil reflection using ultraviolet (UV) camera. Materials were eye balls from 32 live Japanese black cattle and 22 slaughtered cattle. The positive correlation between pupil reflection and VA level was observed in live cattle, whereas the negative correlation between tapetum reflection and VA level in slaughtered cattle was found. These results suggested that LED light doesn’t reach tapetum and it seems to be difficult to get the tapetum information from pupil images.

Recently, tomato production in large-sized greenhouse facilities has been popular in Japan. In large-scale greenhouse, however, harvesting tomatoes by hand is tough work. Therefore, robotic agricultural system has been addressed. In order to promote robotic agricultural system, it is important to develop the machine vision system.