18 research outputs found
Identifikasi Karakteristik Citra Berdasarkan pada Nilai Entropi dan Kontras
Get PDFAbstract
Determining the object boundaries in an image is a necessary process, to identify the boundaries of an object with other objects as well as to define an object in the image. The acquired image is not always in good condition, on the other hand there is a lot of noise and blur. Various edge detection methods have been developed by providing noise parameters to reduce noise, and adding a blur parameter but because these parameters apply to the entire image, but lossing some edges due to these parameters. This study aims to identify the characteristics of the image region, whether the region condition is noise, blurry or otherwise sharp (clear). The step is done by dividing the four regions from the image size, then calculating the entropy value and contrast value of each formed region. The test results show that changes in region size can produce different characteristics, this is indicated by entropy and contrast values of each formed region. Thus it can be concluded that entropy and contrast can be used as a way to identify image characteristics, and dividing the image into regions provides more detailed image characteristics.
Analisis Akurasi Penerapan Algoritma Support Vector Machine Menggunakan Kernel Radial Basis Function pada Penentuan Kelayakan Kredit (Studi Kasus German Kredit Data)
No full textMachine Learning merupakan disiplin ilmu kecerdasan buatan dengan teknik statistika untuk menghasilkan suatu model dari sekumpulan data. Terdapat banyak algoritma machine learning yang biasa digunakan, salah satunya support vector machine (SVM). SVM salah satu metode yang dapat melakukan pengklasifikasi data dengan baik, karena proses yang akan dilakukan bersifat non linear maka akan menggunakan kernel Radial Basis Function (RBF). Fungsi RBF dapat diterapkan dalam mengklasifikasi permohonan kredit. Penelitian ini menggunakan German Credit Dataset. German Credit Dataset dengan 1000 data memiliki 21 variabel terdiri 20 variabel input dan 1 variabel target dengan kelas tidak seimbang. Hasil dari pengujian ditampilkan dalam bentuk, confusion matrix yang akan digunakan untuk perhitungan akurasi dan area under curve untuk perhitungan performa. Dari percobaan yang dilakukan diperoleh nilai akurasi 0.743 (74%) dan nilai performa 0.7689 (77%) termasuk kategori Fair Model. Kata Kunci: Support Vector Machine, Radial Basis Function, German Credit Datase
Analysis and Design of CRC-32 IEEE 802.3 Generator for 8 Bit Data Using VHDL
No full textIn the communication system to achieve better quality data transmission required a method that can detect errors and correct errors. Cyclic Redundancy Check (CRC) is one of the methods used to perform data transmission on data link layer that can detect errors. CRC-32 is used to error-checking on Ethernet or implemented to IEEE 802.3. CRC generator in this research use CRC-32 IEEE 802.3 with 8 bit data width. This research can be implemented in the field of Electrical Engineering, especially in the telecommunications section, namely Ethernet which functions for transfer files and data via a computer network. CRC here has a role to prevent data changes caused by noise during the transmission process. The methods used in this design is modulo-2 division parallel circuit. This design is expected to use a simple schematic circuit, less noise and less resources. Testing is done by matching result of simulation using Xilinx ISE Simulator with implementation on Spartan 3E XC3S500E device with result of count . This research requires a resource of 223 4-input LUTs, 114 Occupied slice, 72 IOB flip flops, 114 bonded IOBs and 1 BUFGMUXs, where this research obtained resources is fewer than with previous researc
Design of birds detector and repellent using frequency based arduino uno with android system
No full textStudi tentang Penangkap Gerakan ( Motion Capture) Manusia
No full textMotion capture merupakan proses perekaman gerakan nyata dalam kehidupandarisubjek sebagai urutan-urutan koordinat Cartesian dalam ruang 3D. Perekam gerakan optis (Optical Motion Capture) menggunakan kamera untuk merekonstruksi postur tubuh pelaku. Salah satu pendekatan dengan menggunakan satu set kamera disinkronisasi untuk menangkap beberapa tanda yang ditempatkan di lokasi strategis pada tubuh. Sistem magnetik menggunakan sensor elektromagnetik sedangkan sistem mekanik menggunakan sensor mekanik untuk mendapatkan data gerakan. Sistem markerless, didasarkan pada visual hull dan kondisi sebenarnya dari subyek. Sebuah sistem motion capture memiliki aplikasi-aplikasi antara lain dalam komputer grafis untuk animasi karakter, dalam virtual reality untuk antamuka kontrol manusia , dan video game untuk simulasi realistis gerakan manusia. Dalam tulisan ini dibahas tentang teori, implementasi dan penelitian-penelitian yang telah dilakukan tekait dengan optical motion capture
Automatic Fire Detector using Arduino with Output Evacuation Route via Short Message Service (SMS) and Warning via Ringtones
Get PDFIn this modern era, the use of electricity is increasing and of course it is very prone to trigger fires. When there is a fire, many people are trapped and difficult to find an evacuation route if it is in a difficult range such as in a tall building. In this final project aims to provide fire warnings and display safe routes as evacuation routes. This fire detection device uses Arduino Uno as a controller and LM35 temperature sensor as a fire detector. GSM Shield is used to send SMS and make calls to the user. The working principle of this tool, by detecting the presence of a fire and the input of the heat temperature is processed when the conversion voltage is 350mV in accordance with the provisions of the fire status. The process is run by the program that has been uploaded to Arduino, after the process is run, then the output appears in the form of fire information and evacuation route with SMS and fire alerts with ringtones. There are 6 temperature sensors used, with 2 sensors installed on each floor. Each sensor that detects a fire, an SMS is sent with the specified status of each sensor and the safe route that must be passed. After the SMS is sent, the phone call comes to alert the user if they do not know of a fire or have not seen the incoming SMS
Concurrently wireless sensor network using microcontroller for home monitoring against fire
Get PDFThis paper proposes a prototype fire monitoring system with Wireless Sensor Network (WSN) in order to detect fires in several places at once and facilitate the placement of the detection. WSN is a technology where several sensors work together to establish communication over a wireless network. This prototype fire monitoring system can be monitored through the website in real time and the detection data is stored on the cloud server. This prototype fire monitoring system uses 4 sensor nodes, each of which is placed in several places to detect fires and sends the detection results to the master node. The master node receives and processes the data sent by the sensor node to generate 16 different conditions concurrently, in the event of a fire it will send a telegram message and the condition data to the cloud server. Several attempts to send data from the sensor node to the master node were completely successful and sending data from the master node to the cloud server as well as sending notification messages have been sent properly
Peer Review : Designing and Implementing System of real Time Face detection and recognition based on RBF
Get PDF
