It is well known that machine vision technique is very useful for inspecting size, shape, color and damage. Recently, machine vision technique for sorting and inspecting fruit vegetables has been studied by many scientists. Due to shape irregularity of some fruit vegetables, the traditional image recognition and image understanding methods can not judge the shape satisfactorily and therefore, the shape judgment for sorting those fruit vegetables is still done manually on the site. For this reason, in this study image processing, neural network and fuzzy were applied to precisely judge the shape to develop an automatic sorting system for fruit vegetables.
This report, presents a new method of feature extraction, and developed sorting system and its software for fruit vegetables. The technology of image processing and neural network were applied in the developed software and the sorting system. Because of the learning capability, the new sorting system can sort more than one variety of fruit vegetables with high accuracy compared with the manual operation. The experiment results of its application to four varieties of strawberry will be presented in the second report. Moreover, the research results of the sorting system as applied to green pepper and other fruit vegetables will be presented in the near future.

View full abstract

The shape judgment on four varieties of strawberry was carried by sorting system software which applied the neural network reported in first report (part 1). The varieties were 'Reiko', 'Toyonoka', 'Wyoho'and'Akihime'. The patterns for every variety were made to be learned by the software. The results of shape judgment of strawberries by using the learned software was 94%-98% at the agreement rate. Also the image area and the weight of strawberries showed significant relationship that the correlation coefficient was above 0. 93. Therefore, it was thought that the strawberry size could be judged from the projection area. The software for the sorting system could judge the shape of strawberry by learning the standard pattern of strawberry for each variety. Moreover, the developed software proved to be versatile, hence it is a multipurpose sorting system for vegetable fruits.

View full abstract

A new procedure for reducing the noise on time-series data is presented. Lowpass filters commonly are used for noise reduction, but high-frequency components such as plant elongation responses to sudden changes in environmental conditions also are removed. This phenomenon complicates analysis of data depicting plant responses to changes in environmental conditions.
The procedure we developed consists of moving average noise reduction combined with numerical differentiation. The moving average is an all-encompassing method of noise reduction and is applied easily, the length of the symmetric moving average operator is adjusted according to a transition in the original time series. A second derivative is used to detect the transition in the original time series. It is assumed that a point in the original time-series with a large second-derivative value is likely the transition point. The length of the symmetric moving average operator is decreased when the second derivative is large and increased when the second derivative is small. Parameters of the developed procedure are the minimum and maximum values of the moving average operator length. These values affect noise reduction levels independently in the transition and stable domains.

View full abstract

A mathematical model to estimate the micropropagated plug production cost using axillary bud propagation was proposed on the basis of data on Prunus verecunda culture. No significant difference of the production cost was observed between shoot number 11 and 17 in a culture vessel at the process of rooting.

View full abstract

The previously described cost simulation model was used to improve micropropagation of Brassica campestris cv. Kukitachina in vitro. The simulation was based on data obtained from different practical experiments including the use of simple medium and food containers, rooting from excised single shoots and acclimatization at a time, and so on.
There was no significant difference between the effect of Otsuka hydroponic culture nutrient or Hyponex with additives and the modified Murashige and Skoog's medium on successive shoot multiplication. Also the reagent grade sucrose could be replaced with table sugar, granular sugar or reagent grade glucose without adverse effect on the successive shoot multiplication. The above results suggested that cheap carbon sources could be successfully used in micropropagation. Also in comparison with culture bottles, food containers were equally effective for successive shoot multiplication. The shoots multiplied three times every 20 days. Rooting from excised shoots and acclimatization of in vitro plants were possible in nursery boxes and seventy per cent of regenerants were successfully planted out.
The production cost of in vitro plants at 10, 000 level production, using the above modified procedure was calculated to be 61. 0 yen per plant. This represents a 39% cost reduction when compared with the previously described laboratory procedure. Improvement of micropropagation by simulation of production cost was discussed.

View full abstract

A prototype system has been developed, for the purpose of realizing a fruit harvesting system, here by us. This prototype system consists of 1) a robot with an articulated arm that has a small CCD camera in its hand and 2) real-time image processing technology.
This single camera can measure the three-dimensional position of a target in space by means of moving the camera after catching the target in its sight. The parameters of the camera system can be determined automatically by using image processing to measure a standard mark put on a known position. The robot system has the ability to learn the visual features of a target, extracting them from an image for example that of a tomato or an apple. Using these features the target can be recognized.
Results obtained from harvesting trials using tomatoes and apples have shown that prospects of practical speed and precision of harvesting are evident.

View full abstract

Growth and maturity judgment for leafy head vegetable, such as cabbage, lettuce, and so on, is determined through vision by estimating head size. Image processing can be a vital tool for measuring parameters in agricultural production. Image processing technique for detecting the head of cabbage have been developed but the noise caused by outer leaves and adjacent cabbages resulted in failure of detection. Therefore, noise elimination techniques have been developed in this study based on the assumption that the top view shape of a cabbage is circular. Two techniques, Region of Interest Method and Hough Transformation Method, was investigated. These two methods showed the feasibility of estimating head size and the simplification of the process using a premise of cabbage characteristics. The Hough Transformation Method gave a higher precision of head detection than the Region of Interest Method, although the computed head size is slightly larger than the actual size. Processing speed of the Hough Transformation Method, however, was rather low. This latter performance, therefore, indicated the difficulty of its application to real time processing such as harvest mechanism.

View full abstract

Shoots of five horticultural crops were cultured in the following systems, 1) the “Culture Pack” -Rockwool (CP-RW) system, a prototype of film culture system, 2) its practical model (PCP-RW system), and 3) a conventional culture system (BOS·AGAR). The aim is to investigate the efficiency of film culture systems on the in vitro production of plantlets. Regardless of the culture medium used and plant species examined in this study, increased growth of plantlets was obtained in both film culture systems as compared to those in the conventional system. This suggests that PCP?RW system is more efficient and has a greater practical value in micro-propagation of many crops than the conventional culture system (BOS·AGAR).

View full abstract

The purpose of this study was to determine the effect of water quality on the germination and the growth of hypocotyl in plants. The spring water of Tetuta village in Okayama prefecture (M-water) influenced germination and elongation of the hypocotyl in the Kaiware radish. Compared with tap water (Kumamoto City), M-water gave a high rate of germination. The growth of hypocotyl with distilled water was significantly different (P=5) from that with M-water on 9 days after plating. The enzyme activity in the seed was higher with M-water than with tap and distilled water.

View full abstract

The influence of water on plant tissue is discussed. The results of the cytological observation are as follows :
The average length of guard cells grown with desalinated water, tap water and M-water (Tetuta village) was 22.2, 25.7 and 27.9μm, respectively.
Epidermal cells of the hypocotyl cultivated with M-water were much larger than those cultivated with desalinated water and tap water.
Observation under an optical microscope, revealed that the vessel trachea in the vascular bundle had grown more vigorously in the M-water than in the desalinated water or tap water.
The concentration of auxin examined by P-test was higher in the plants cultivated with M-water.

View full abstract