Kusworo Adi

Unidirectional broadband microstrip antenna for through walls radar application

We present a design and simulation of broadband microstrip antenna with unidirectional radiation pattern for through walls radar application. Design and simulation with computer was confirmed the implementation of the proposed design. The results were validated by measurements in the laboratory. This research is a continuing from our previous research in developing antennas for through walls radar application. In previous research, we have developed a broadband microstrip antenna for through walls radar application in a bidirectional radiation pattern. Based on the laboratory experiment results, bidirectional radiation pattern still have weaknesses on detecting objects movement behind radar antenna. Unidirectional antenna has proven to have greater gain than bidirectional antenna. To obtain unidirectional radiation pattern antenna we added reflector behind the antenna to detect moving objects from in front of the antenna only. The previous bidirectional antenna for through walls application showed 4–5 dBi gain. In this research, the design and simulation of unidirectional antenna for through walls application showed 5.5 to 7.5 dBi gain. The unidirectional antenna will be printed using FR4 material with a dielectric constant and substrate thickness of 4.4 and 1.6 mm respectively. This antenna has a bandwidth of 1.5 GHz (66.67%) with frequency of 1.5 to 3 GHz.

Phase Unwrapping by Markov Chain Monte Carlo Energy Minimization

Phase unwrapping (PU) is a process to obtain the absolute value of a phase field from the wrapped one. Ideally, without phase noise, singularity, and aliasing problems, the phase information can be unwrapped easily. However, the phase data are always contaminated by noise and discontinuities, making the PU process more complicated. Hence, a suitable PU algorithm is required to address the problems appropriately. Markov Chain Monte Carlo (MCMC) Energy Minimization is usually used to partially remove phase noise. In this letter, the MCMC energy minimization that yields the unwrapped phase is shown. Furthermore, the capability of the proposed method to unwrap simulated and actual InSAR phase images is also demonstrated (Fujiyama Mount).

Design of crack detection system for concrete built infrastructure based on fiber optic sensors

This paper presents crack detection in concrete built infrastructure using fiber optic sensors based on the pressure measurement system. Fiber optic sensors in this experimental study were constructed using micro-bending techniques. The bends in optical fiber were evaluated based on pressures given by Compression Testing Machine (CTM). Pressure was performed on the surface of concrete specimens, in which there was a fiber optic in the middle of each specimen. Dimension of concrete specimen was 15 cm x 10 cm, with thickness variations of 2cm, 4cm and 6cm. The results showed that the higher the pressure value, the lower the intensity of signal light in optical fiber. It is mentioned as the attenuation of fiber optic. It depends on the thickness and the maximum strength of the specimen.This paper presents crack detection in concrete built infrastructure using fiber optic sensors based on the pressure measurement system. Fiber optic sensors in this experimental study were constructed using micro-bending techniques. The bends in optical fiber were evaluated based on pressures given by Compression Testing Machine (CTM). Pressure was performed on the surface of concrete specimens, in which there was a fiber optic in the middle of each specimen. Dimension of concrete specimen was 15 cm x 10 cm, with thickness variations of 2cm, 4cm and 6cm. The results showed that the higher the pressure value, the lower the intensity of signal light in optical fiber. It is mentioned as the attenuation of fiber optic. It depends on the thickness and the maximum strength of the specimen.

Detection Lung Cancer Using Gray Level Co-Occurrence Matrix (GLCM) and Back Propagation Neural Network Classification

Lung cancer prevalence is one of the highest of cancers, at 18 %. One of the first steps in lung cancer diagnosis is sampling of lung tissues or biopsy. These tissue samples are then microscopically analyzed. This procedure is taken once imaging tests indicate the presence of cancer cells in the chest. Lung cancer diagnosis using lung tissue sample microscopic analysis has some weakness. One of them is that doctor still relies on subjective visual observation. A medical specialist must do thorough observation and accurate analysis in detecting lung cancer in patients. Hence, there is need for a system that is capable for detecting lung cancer automatically from microscopic images of biopsy. This method will improve the accuracy and efficiency for lung cancer detection. The aim of this research is to design a lung cancer detection system based on analysis of microscopic image of biopsy using digital image processing. Microscopic images of biopsy are feature extracted with the Gray Level Co-Occurrence Matrix (GLCM) method and classified using back propagation neural network. This method is implemented to detection both normal and cancerous lung of biopsy samples. In the stage of training, 20 biopsy image samples were analyzed using back propagation neural network with 95% accuracy. On the other hand, 16 biopsy samples were analyzed during testing, with an accuracy of 81.25%. These results show that microscopic biopsy image processing can be implemented in a system of lung cancer detection.

Geographic Information System Carbon Development Landed on Land of Primary Dry Limits using Method Stock Difference

Land use emission factor is one of the causes of global climate change The Intergovernmental Panel on Climate Change (IPCC) concludes that most of the global average temperature increase is due to human activities through GHG emissions. GHG emissions come from a variety of sources from agriculture, forestry, and other land uses to a land ecosystem, derived from changes in carbon stock and from non-CO2 emissions as well as from various sources. Changes in carbon stocks in a Biomass can be calculated using the Stock-Difference method in which carbon stores in a cage are measured at two different times to assess changes in carbon stores. In finding the value of carbon stocks in primary dryland forest is highly dependent on the availability of data that becomes the supporting variable that is the data of forest for each region, then GIS becomes one of the solutions in obtaining data. Our results conclude that in 2012 and 2016 there was a decrease in carbon savings of an average of 1,842 C tons/ha and an average rate of emission increase of 5,430.73 CO2 tons/ha.

Calculation of Lung Cancer Volume of Target Based on Thorax Computed Tomography Images using Active Contour Segmentation Method for Treatment Planning System

In this research, calculation process of the lung cancer volume of target based on computed tomography (CT) thorax images was done. Volume of the target calculation was done in purpose to treatment planning system in radiotherapy. The calculation of the target volume consists of gross tumor volume (GTV), clinical target volume (CTV), planning target volume (PTV) and organs at risk (OAR). The calculation of the target volume was done by adding the target area on each slices and then multiply the result with the slice thickness. Calculations of area using of digital image processing techniques with active contour segmentation method. This segmentation for contouring to obtain the target volume. The calculation of volume produced on each of the targets is 577.2 cm3 for GTV, 769.9 cm3 for CTV, 877.8 cm3 for PTV, 618.7 cm3 for OAR 1, 1,162 cm3 for OAR 2 right, and 1,597 cm3 for OAR 2 left. These values indicate that the image processing techniques developed can be implemented to calculate the lung cancer target volume based on CT thorax images. This research expected to help doctors and medical physicists in determining and contouring the target volume quickly and precisely.

Beef Quality Identification Using Thresholding Method and Decision Tree Classification Based on Android Smartphone

Beef is one of the animal food products that have high nutrition because it contains carbohydrates, proteins, fats, vitamins, and minerals. Therefore, the quality of beef should be maintained so that consumers get good beef quality. Determination of beef quality is commonly conducted visually by comparing the actual beef and reference pictures of each beef class. This process presents weaknesses, as it is subjective in nature and takes a considerable amount of time. Therefore, an automated system based on image processing that is capable of determining beef quality is required. This research aims to develop an image segmentation method by processing digital images. The system designed consists of image acquisition processes with varied distance, resolution, and angle. Image segmentation is done to separate the images of fat and meat using the Otsu thresholding method. Classification was carried out using the decision tree algorithm and the best accuracies were obtained at 90% for training and 84% for testing. Once developed, this system is then embedded into the android programming. Results show that the image processing technique is capable of proper marbling score identification.

CHEST X-RAY SEGMENTATION to CALCULATE PLEURAL EFFUSION INDEX in PATIENT with DENGUE HEMORRHAGIC FEVER

A study of the calculation of pleural effusion index (PEI) in patient with dengue hemorrhagic fever (DHF) has been conducted. PEI calculation was done through matlab programming language. Some digital image processing methods used in this research were thresholding segmentation, morphology operation, and calculation of pixel number per column in image to get PEI value. PEI values generated from image processing can be an alternative to replace the manual calculations done by physicians and used to demonstrate the gravity level of DHF. Keywords— Dengue hemorrhagic fever, Pleural effusion index, Segmentation.

Detection of the beef quality: Using mobile-based K-mean clustering method

Beef quality is determined by a number of parameters; some of which include size, texture, color feature, or meat smell. Recently, determining the meat quality is done by seeing the color and shape. However this method still has some weaknesses due to, for example, the subjectivity and inconsistency in human assessment. The aim of this research is to make an application to detect the meat quality. The application built was based on mobile using Java Programming Language on the Android integrated with Android SDK, Eclipse, and OpenCV. The method of image processing used pre-processing, k-mean clustering, and the analysis was conducted statistically with mean value and deviation standard. The quality detection meanwhile was performed using the texture and meat texture matching based upon the existing data. The application made could be used to seek the significant k-values and able to detect the level of quality by providing the level of accuracy at 80%.

Beef quality identification using color analysis and k-nearest neighbor classification

Beef is one of the many produce prone to contamination by microorganism. Water and nutrition contents make an ideal medium for the growth and proliferation of microorganism. Contaminated beef will degrade and has less storage duration. Beef is valued by two factors; its price and its quality. The quality itself is measured using four characteristics; marbling, color of meat, color of fat, and meat density. Specifically, marbling is the dominant parameter that determines meats quality. Determination of meat quality is conducted visually by comparing the actual meat and reference pictures of each meat class. This process is very subjective in nature. Therefore, this research aims to develop an automated system to determine meat by adopting the Indonesian National Standard requirement on the quality of carcass and beef (SNI 3932:2008) using the image processing technique. Image segmentation is carried out using the thresholding method and classification is conducted using the k-nearest neighbor algorithm. The features used to differentiate beef quality are marbling score, color of meat, and color of fat. Results indicate that the system developed is able to acquire images and identify beef quality as required in the Indonesian National Standard.