Search

In the surface mount technology, Ball Grid Array (BGA) has been used in a production of PC boards, because of their excellent characters such as high density of the lead pin pitch, better lead rigidity and self-alignment during re-flow processing. This paper deals with the development of image analysis for the detection of defects at BGA solder joints in PC boards by using X-ray images. In the conventional IC boards, it is possible to detect defects of solder joints by visual inspection, because the lead of IC package is set on its outside. However, we can't detect visually defects at BGA solder joints, because they are hidden under the IC package. In a production line, the inspection of BGA in PC boards depends on the function test of electric circuits in the final process. To improve a cost performance and the reliability of PC boards, an inspection of BGA is required in the surface mount process. Types of defects at BGA solder joints are solder bridge, missing connection, solder voids, open connection and miss-registration of parts. As we can find mostly solder bridge in these defects, we pick up this to detect solder bridge in a production line. The problems of image analysis for the detection of defects at BGA solder joints are the detection accuracy and image processing time according to a line speed of production. To get design data for the development of the inspection system, which can be used easily in the surface mount process, it is important to develop image analysis techniques based on X-ray image data. At the first step of our study, we attempt to detect the characteristic of the solder bridges based on an image analysis.

In a current study of environmental information, it is important to detect the particular object from environmental images under various conditions for security such as navigation system of ships, bird-watching for the prevention of various damage, invaders check and human rescue in the marine casualty. This paper deals with the image processing techniques for the detection of the rescue target as a particular object from environmental images. When a marine casualty occurs the detection of the rescue target such as life rafts depends on the visual search of human eyes. It is predicted, however, that human eyes sometimes lose their sight and the ability of the detection falls owing to the long flight and a bad weather. For a practical purpose of the prompt rescue of human life, the development of the searching support system in place of human eyes is required. To realize a new type searching system for the detection of the rescue target, the development of new type image processing techniques in real time manner is important. At the first step of our study, we attempt to develop image processing techniques, in order to accurately detect the rescue target under various

A new composite image sensor system is proposed. Two different type image sensors, area and line sensors, are used effectively to make dimension and shape measurements with high accuracy. An area sensor is used to obtain a whole image of the object at an accuracy of 1 mm and to pick up regions which are necessary to measure shapes and position of edges or holes, then a line sensor is used to obtain a high resolution image of picked-up regions by rotating and scanning the line sensor at an accuracy of 0.1 mm. This sensor system has an advantage that the acquisition time of image input is shortened and that the amount of data is much reduced as compared with an image input system using a line sensor only.Experimental results show that this composite image sensor system is available in the field of not only parts'dimension and shape measurement but also three-dimensional geometrical measurement.

We investigated difference of the individual average glandular dose (AGD) as individual patient dose between digital and screen-film mammography systems. The difference was analyzed for every individual breast thickness or breast density. AGD that is average radiation absorbed dose to glandular breast tissue is accepted as an estimation of the patient dose in mammography. If calculating the individual AGD, Glandular Rate, which is the rate of the quantity of glandular tissues to breast, should be known for each patient. Then, we built an estimating system of individual glandular rate from each clinical mammogram by using breast tissue equivalent phantoms that are able to change the mixture ratio of adipose and glandular tissue and the thickness. Two data sets of Cranial- Caudal view mammograms were estimated. The one is 100 images from 50 patients by the screen-film (min-R 2000, kodak), and the another is 49 images from 25 patients by computed radiography (CR, Fuji) system. AGDs of screen-film and CR systems were 1.6mGy and 2.0mGy, respectively. The increase of AGD showed the plus correlation with breast thickness.

To increase the sensitivity of the dosimetry system for high energy x-rays, using Gafchromic EBT film (International Specialty Products) and a flat-bed scanner with transparency unit (ES-2200 or Expression 1680, Epson, Inc.), the optimal sharp cut (SC) filters like optical filters to be added to the scanner were investigated. Gafchromic EBT film (EBT) has the absorption spectrum with a maximum absorption band centered at 636 nm. Consequently the response of EBT will be enhanced by measurement with red light and can be very significantly increased by using a red filter while making scanner measurements. In our research, employed SC filters were SC filter 54, 56, 58, 60 and 62 (product made from Fuji film). SC filters are defined as filters that cut off as much as possible of the wavelength light shorter than a specific wavelength, while transmitting as much of longer wavelength light as possible, within a wavelength range of 350 nm to 800 nm. Generally, it is ideal to use a narrow band pass filter. However we didn’t employ the filter because the transmissivity of the filter of film type is low or the thickness of glass type filter is large, so it is unsuitable for a flat-bed scanner. Film pieces of EBT were exposed to the following doses: 0.3 – 10.0 Gy with the 6 MV x-rays from a linear accelerator. The irradiated film pieces to which the SC filter was added were scanned and the scan data of red channel was converted into optical density. Furthermore the optical density was divided by the absorbed dose and the value was estimated as sensitivity. EBT with SC filter 62 with 50% of transmissivity at 620 nm was about 1.48 times more sensitive on the average than without a SC filter.

Purpose: We developed an infrared flat-bed scanner with a transparency unit (IR 4000, iMeasure, Inc.) in order to increase the sensitivity of measuring density of Gafchromic EBT film (International Specialty Products). Then, the performance of IR 4000 was investigated, comparing with a general-purpose flat-bed scanner. Methods and Materials: Gafchromic EBT film (EBT) has the absorption spectrum with a maximum absorption band centered at 636 nm. Consequently the response of EBT will be enhanced by measurement with red light. Then, the transparency unit of IR 4000 for EBT has light emitting diode (LED) with a peak wavelength of 630 nm, and was produced by replacing LEDs of white light in the transparency unit for the flat-bed scanner (ES 2200, Seiko Epson Corp). Samples of EBT were covered with solid water RMI-457 phantoms and were exposed to the following doses: 0.3 – 10.0 Gy with the 6 MV photon beam from a Siemens MEVATRON77 DX67 accelerator. The irradiated samples were scanned by IR 4000 with 48 bit color mode and 1200 dpi and the scan data of red channel was converted into optical density. And also, the samples to which the yellow filter (International Specialty Products) was added were scanned by ES 2200. Furthermore the optical density by the two scanners was divided by the absorbed dose and the value was estimated as sensitivity. Result and Discussion: EBT using IR 4000 was about 1.60 times more sensitive on the average than using ES 2200 with the yellow filter in the dose range from 0.3 Gy to 10.0 Gy. By using IR 4000, the increase of sensitivity for EBT is easily achieved without using the filter which has the problem of dust, scratch or oxidation. Conclusion: The high capability of all Gafchromic films is utilizable by employing IR 4000 which has the easily exchangeable transparency unit.

Purpose: A balanced steady state free precession (b-SSFP) sequence is very effective in clinical examination for renal MR angiography (MRA) without a contrast medium. Generally, in image evaluation, the increase in spatial resolution causes decrease in SNR. Therefore, we investigated the performance of b-SSFP in the relation between SNR and matrix. Materials and Methods: A phantom consisting of 31 bottles of Gd-DTPA (Schering AG, Germany), from 0 to 50 mmol/L (mM), was used. The protocol of b-TFE was as follows: repetition time (TR) msec/echo time (TE) msec, 8.8/4.4; flip angle, 80 degrees. The combination of matrix and scan percentages were three types, 256 × 256 (100%), 256 × 512 (200%), and 256×1024 (400%). The contrast dynamic range (CDR) was measured from signal intensity of each Gd phantom. Ferucarbotran (Schering AG, Germany) solution of 0.33 mM was injected by 0.2 ml/sec to the inside of simulation vessel of 2 mm in diameter. The profile curve was drawn along the simulation vessel on maximum intensity projection (MIP) images. All images were acquired on 1.0 T MR system (Gyroscan Intera 1.0 T, Philips Medical Systems). Results and Discussions: The signal intensity increased as CDR became narrow. The increase was based on a noise, and was not change of tissue itself. Increasing the matrix did not influence the CDR and sensitivity. In the curve normalized on an inflow point, 1024 matrix of large CDR was higher than 256 matrix of small CDR. We thought that the increase in the noise causes deterioration of the quality of MIP images. Therefore, we consider that it is effective to decrease the noise by optimization of CDR. Conclusion: It is thought that b-SSFP covers the fault in the SNR of the high resolution MRA, and can utilize efficiently the high resolution capability, when the condition of CDR is optimized.

We built an estimation system of individual breast density from digital mammograms by using breast tissue equivalent phantoms that are able to change the mixture ratio of adipose and glandular tissue and the thickness. The method was compared with a visual assessment of breast density by the radiologists as a gold standard. The clinical image data set average of the estimated glandular rate was 35.0%, while varying from 12.0% to 67.0%. This system can apply in clinical practice such as the visual estimation of breast density.

Radiochromic film has been applied to mapping of dose distributions in radiation therapy. However, this particular RCF product suffers from intrinsic nonuniform response. In this study, we have carefully measured the nonuniformity of unirradiated RCF, and the response of irradiated RCF in pixel-to-pixel as function of various absorbed doses, in order to create a method of predicting the error of irradiated, unirradiated RCF.

Wavelet transform has been applied for multi resolution analysis of oceanic feature in ERS1/2 SAR and ASTER image data over the Tsushima Strait. Various wavelet transforms, e.g. Haar, Daubechies, Symlet, Coif, Biorthogonal, and Meyer wavelet, have been tested comparably with different level of synthesized image on horizontal, diagonal, and vertical detail, and approximation to study the internal wave characteristic in image. The results show that higher scale processing of image and higher number of vanishing moments of wavelet function such as Coiflet wavelet, coif5, Daubechies wavelet, db5, Symlet wavelet, sym8, and Discrete Meyer wavelet, are proposed for smoother wave crest. The best result of detection is performed by Discrete Meyer wavelet transform, which adapts FIR filter, orthogonal and biorthogonal basis. Thus, it liable for smoothness of feature, space save coding and avoid dephasing in image.

To improve the cost of performance in manufacturing IC packages, it is required to inspect BGA defects in the online process. The problems of image analysis for the detection of defects are the detection accuracy and image processing time according to a line speed of production. Using the X-ray penetration equipment, we have captured images of an IC package to search an abnormal BGA. To get design data for the development of the inspection system, which can be used easily in the surface mount process, we tried to capture the tomographic images utilizing the latest imaging techniques.

The influence of image quality characteristics on the accuracy of vessel measurement was investigated using simple vascular models, the comparison of the profile curve of the cross section of the vascular models and physical image quality characteristics such as modulation transfer function (MTF), Wiener spectrum (WS) and contrast, with changing scan conditions. The profile curves in the section part that greatly influences the measurement accuracy of the vessel diameter has changed by x-ray tube voltage. The noise has increased evenly in all frequency band regions at the same time though the contrast rose in 2 and 4 mm vascular models. The change was not seen in MTF. It can be said that the decision of the scan parameters that raises the contrast and decreases the noise will improve the measurement accuracy of the vessel diameter.

We built a estimating system of individual breast density from digital mammograms by using breast tissue equivalent phantoms that are able to change the mixture ratio of adipose and glandular tissue and the thickness. The method was compared with a visual assessment of breast density by the radiologists as a gold standard. The clinical image data set that consisted of 49 mammograms of 21 patients were analyzed. The average of the estimated glandular rate was 35.0%, while varying from 12.0% to 67.0%. This system can apply in clinical practice such as the visual estimation of breast density.

Radiochromic film (RCF) has been applied to mapping of dose distributions in radiation therapy. However, this particular RCF product suffers from intrinsic nonuniform response. In this study, we have carefully measured the nonuniformity of unirradiated RCF, and the response of irradiated RCF in pixel-to-pixel as function of various absorbed doses, in order to ascertains for the method of predicting the error of irradiated, unirradiated RCF.

Radiochromic film (RC-film) is of great interest as a film-type dosimeter for radiation oncology applications. We present a two-dimensional image-based evaluation of the measurement accuracy of a commercial RC-film product (Gafchromic MD-55-2 film, ISP TECHNOLOGIES, Inc.) by using a commercial Laser Densitometer (Model 1710, Computerized Medical Systems, Inc.) as an optical density imaging system. The coefficient of variation of the density (pixel-value) in one sample was approximately 3% to 11% at 3 Gy or less, and 3% or less at 4 to 60 Gy. Although the coefficient of variation between three samples at the same dose was about 14% at 1 Gy, it decreased as the dose increased, reaching several percent. In 1 to 6 Gy samples, geometric imaging artifacts [interference (moire) patterns] were observed, and it was found that scan-sampling pitch influenced the accuracy of measurement of the density of the sample. To improve the accuracy of density measurement, sufficient knowledge about characteristic features of the density measuring system is essential.

This paper deals with the developmment of image analysis for the detection of defects at BGA solder joints in PC boards by using X-ray imaging. We can't detect visually defects at BGA solder joints, because they are hidden under the IC package. To improve a cost performance and the reliability of PC boards, an inspection of BGA is required in the surface mount process. Types of defects at BGA solder joints are solder bridge, missing connection, solder voids, open connection and miss-registration of parts. As we can find mostly solder bridge in these defects, we pick up this to detect solder bridge in a production line. The problems of image analysis for the detection of defects at BGA solder joints are the detection accuracy and image processing time according to a line speed of production. To get design data for the development of the inspection system, which can be used easily in the surface mount process, it is important to develop image analysis techniques based on X-ray image data. At the first step of our study, we attempt to measure the shape of BGA based on X-ray imaging.

Our aim was to improve the accuracy of classifying x-ray mammographic breast densities. The histogram approach using the neural network was used for the purpose of constructing a flexible system. In this study the phantom of the synthetic breast-equivalent resin material for the process of the A/D conversion of mammograms was employed. The digital values can offset the difference in characteristics between the mammography system, the unit, etc. Furthermore the features of our system use the neural network, and then tune the neural network by the histogram of the digital values and by the radiologists' and expert mammographers' assessment ability. Although there was an observer's bias, our system was able to classify the breast density automatically according to that observer. This is only possible if the observer has been trained to some extent and is capable of maintaining an objective assessment according to the assessment criteria.

In the surface mount technology, a ball grid array (BGA) has been used in the production of PC boards. This paper deals with the detection of defects at BGA solder joints in PC boards by using X-ray imaging. Types of defects at BGA solder joints are solder bridge, missing connection, solder voids, open connection and misregistration of parts. The problems of image analysis for the detection of defects at BGA solder joints are the detection accuracy and image processing time according to the speed of the production line. To get the design data for the development of the inspection system used in the surface mount process, it is important to develop image analysis techniques based on X-ray image data. At the first step of our study, we attempt to detect the characteristics of the solder bridges based on the image analysis technique.

The authors have developed a sensor-combined measurement system to realize a speedy three-dimensional shape measurement at production lines. The system consists of a camera with both area and line sensors, an image input unit with two LUTs, a computer of the workstation type, and a Gray-type lighting-pattern projector with a liquid-crystal mask. The authors discuss high speed image data processing techniques to obtain the space-encoded images of a measurement space and a shape recognition analysis using a gradient histogram on space-encoded stripes of each segmented object image. Although the measurement space encoding of the area sensor image is carried out in a speedy hardware manner, that of the line sensor binary-image consisting of specified subregion images requires a specific technique to copy pertinent stripe codes. without breaking line sensor image resolution, from those of the area sensor space-encoded image. The adaptive lighting-pattern projection method is used for this purpose. Size measurement accuracy, is improved by the present system to seven times higher than that of conventional area sensor measurement. >

A stereoscopic microscope is widely used in a micromanipulation such as to operate genes and to inspect integration circuits. As in these tasks the micromanipulation is handled and makes too heavy burden to operators, it is desirable to perform the micromanipulation automatically. In this paper, we propose a visual feedback system for micromanipulation with stereoscopic microscope. This system takes less time to control the manipulator by reducing searching area to detect an object

This paper deals with the detection of the rescue target as a particular object from motion images under bad conditions. When shipwrecks occur, the search activities of rescue are done using an airplane. The detection of the rescue target such as a life raft depends on visual search of human eyes. To detect a small rescue target in the wide sea, the searcher man must monitor motion sights in nasty weather. In order to carry out the prompt rescue of human life, development of searching support system using image processing techniques in place of the human eye is required It is very difficult to process the motion images taken by the airplane, because these images are inferior due to shaking of the airplane, sunshine reflections and crest waves under bad conditions. We make a simulation experiment for detection of the rescue target from images. To get basic design data for development of the searching system in the case of a marine casualty, we propose motion image processing techniques using improved S/N ratio and dynamic range of the image data of the rescue target in the motion images under bad conditions

X-ray mammography is the most effective method to produce a high photographic sensitivity in the early detection of breast cancer in which palpation is impossible. Mammography can be adapted to a screening check aimed at discovering abnormal remarks in females without the subjective symptom. However, in order to justify examining a seemingly healthy group in screening mammography, it is necessary to perform with an appreciation of the risk of carcinogenesis that may be associated with the absorbed radiation dose to the breast (patient exposure dose), and to guarantee an optimization of radiological protection. To do so, it is necessary to control the individual patient exposure dose in screening mammography. Moreover, when estimating patient exposure doses, it is important that breast composition is investigated objectively. In this study, a breast dosimetry system that includes digitization of a mammogram for investigating the individual breast composition was built on the basis of the dosimetry method for quality control of screening mammography. It is thought that this system has 10% or less of an error factor including the error determination of breast composition. Patient exposure doses of 124 examples from 76 patients in Japan were actually estimated by this system. Their values were clearly lower than the maximum acceptable dose recommended by the American College of Radiology and Japan.

This paper deals with the detection of the rescue target as a particular object from motion images under bad weather conditions. When shipwrecks occur, the searching activities for rescue is done using an airplane. The detection of the rescue target such as life rafts depends on a visual search by human eyes. To detect a small rescue target in a wide sea area, the searching man must monitor motion sights under nasty weather. The human eyes sometimes loses its sight and the ability of detection fails owing to the long flight and the nasty weather. To support the searching activities in the case of a marine casualty, we propose motion image processing techniques using the improved S/N ratio and dynamic range of the image data of the rescue target in the motion images under bad conditions

The purpose of this paper is to develop a non-contact profile sensor system which will be able to accurately measure 3D free-form machined metal surfaces. The proposed sensor system has many advantages as compared with conventional measuring systems. First, a new detecting system of optical ring images utilized by rotating image detector is developed to measure 3D profiles in the long measuring range with high accuracy. Second, processing time can be shortened within 0.5 second by using the proposed detecting system. Third, the speckle noise is eliminated effectively by the rotating mechanism. Finally, it is concluded that this sensor system makes it possible to measure the profiles within an accuracy of +/- 50 micrometers in the measuring range of 150 mm. In this paper, the measurement principle of the proposed sensor system is analyzed and the performance of the system is experimentally measured and discussed not only for both diffuse reflection surface and specular reflection surface, but also for reduction of the laser speckle noise which has direct influence on the measurement accuracy.

The fuel distribution in diesel spray is a predominant factor to the engine performance. For the purpose of improvement of the combustion characteristics in diesel engine, many studies on the fuel distribution in unevaporated spray have been carried out. It is necessary to reveal the relation between the fuel distribution data obtained in unevaporated spray and the diesel combustion.In this study, the effect of fuel injection pressure on combustion in a diesel engine was investigated and the fuel distribution in unevaporated spray was measured by means of the image analysis. The results are as follows:(1) As the increase in the fuel injection pressure, the combustion duration, the specific fuel consumption and the smoke density decreased.(2) There was a considerable correlation between the fuel distribution measured in unevaporated spray and the rate of heat release in the diesel engine.

In order to clarify a mixing process of injected fuel and air at a combustion chamber of a diesel engine, the behavior of unvaporized spray and its fuel concentration distribution after impinging on a plane wall of the combustion chamber were investigated by the image analysis of shadow photographs of fuel spray. Experimental studies were conducted with variation of the patterns of injection pressure, ambient gas densities and diameters of injection nozzles under the conditions which fuel was injected perpendicularly to an acrylic plate wetted by fuel. The mixing between fuel and ambient gas was promoted by the spray impingement, and the shorter the distance between the injection nozzle and the wall was placed, the bigger the mixing was effected. More quantities of ambient gas entrained into spray was observed by the spray impingement on a wall than one of free spray, especially with bigger diameter of nozzles. The empirical relations for the spray development after impingement on the plane wall were obtained and compared with experimental results. (12 figs, 1 tab, 9 refs)