Muhammad Ihsan Zul

Abstrak – Cuaca ekstrim yang berubah – ubah belakangan ini sering terjadi dan sangat menganggu aktivitas sehari – hari. Untuk itu diperlukan suatu sistem yang dapat mendeteksi kondisi cuaca berdasarkan citra awan. Mencari metode untuk mendeteksi kondisi cuaca dalam satu waktu dengan pengolah citra adalah inovasi baru yang muncul dalam pemodelan cuaca saat ini. Hal ini didorong oleh kebutuhan yang tinggi dari berbagai pihak untuk melakukan penelitian dalam mendeteksi suatu kondisi secara teliti dan tanpa harus mengamatinya secara langsung. Dalam penelitian tugas akhir ini akan dirancang sebuah sistem deteksi dan klasifikasi kondisi cuaca berdasarkan citra awan. Sistem ini menggunakan IP Camera untuk menangkap citra awan. Setelah akuisisi citra, dilakukan ekstraksi ciri dengan metode histogram warna terhadap citra awan yang telah ditangkap oleh IP Camera. Kemudian dilakukan proses klasifikasi menggunakan algoritma k-Nearest Neighbor. Algoritma klasifikasi yang digunakan dalam tugas akhir ini adalah k-Nearest Neighbor, yang diharapkan mampu untuk mengenali citra awan dan mengklasifikasi kondisi cuaca. Berdasarkan pengujian yang dilakukan, sistem menghasilkan tingkat akurasi sebesar 84,21%. Kondisi cuaca yang diprediksi adalah kondisi cerah berawan, berawan, hujan, malam cerah dan malam hujan.

Di Politeknik Caltex Riau (PCR) penentuan dosen pembimbing dan penguji Proyek Akhir (PA) merupakan tugas dari koordinator PA. Penelitian ini bertujuan untuk membantu koordinator PA dalam menentukan dosen pembimbing dan penguji PA. Metode-metode yang digunakan untuk membangun sistem ini adalah Text Mining, k-Nearest Neighbor (k-NN), dan Simple Additive Weighting (SAW). Text mining untuk melakukan preprocessing data, k-NN untuk melakukan pengklasifikasian KBK, sementara SAW untuk melakukan pembobotan kriteria dosen. Penelitian ini dapat merekomendasikan 5 nama dosen yang memiliki bobot tertinggi untuk menjadi pembimbing dan penguji PA berdasarkan KBK judul PA mahasiswa, dan memiliki nilai akurasi 83.3%

IP Camera is a device that used for surveillance system. IP Camera can be used to identify the captured object. This study uses the IP Camera to recognize hand shapes were captured by IP camera. Hand shape recognition process is done by using three feature extraction methods. All of these methods are Edge Detection, Vector Analysis and Pixel to Pixel Distance Analysis. The extraction results will be generated by using k-Nearest Neighbor (k-NN) algorithm to recognize the shape name of captured hand. Pixel to Pixel Distance Analysis method is the best feature extraction method. This method only reaches 11.67% error rate. The error rate was calculated from 60 testing image. This method successfully recognizes 53 hand shapes correctly.

Assessment is the most important part on learning activities. Student achievement can be determined based on achievement of final score on particular course. There are many grading system has been developed to support academic process work effectively and efficiently. The student’s final score can be used to evaluating and predicting future student achievements. This research is conducted to analyzing and comparing two classification algorithm that used to predicting future student final score based on past semester data. This research proposed two classification algorithm such as k-NN and Naïve Bayes. The result of this study states that k-NN (k=15) is the best algorithm. The accuracy of k-NN reach 95%. The best algorithm will be implemented to academic assessment system and used as early warning system to monitor student learning process

"Domotic is a system that, through Information Technology, interactively regulating electrical devices within the building in order to increase comfort, energy efficiency, and security. There are some problems that need to be addressed in current domotic system. They are hardware and software interoperability, and affordability.
This study will design a domotic system that has interoperable characteristic while maintaining cost relatively low. The system is aimed for monitoring and controlling residential houses. A gateway as a communication bridge between heterogeneous protocols is developed. Service Oriented Architecture is utilized to enhance the software interoperability. To reduce retrofitting cost, wireless sensor network is used."

IP Camera is a one of detection devices that most widely used to home, buildings and areas surveillance. IP Camera has been designed to communicate directly over the Internet network. It can be done because the IP Camera has equipped by a web server and web pages as an interface. In general, IP Camera comes with the ability to detect motion. The motion information is delivered using FTP protocol. This research proposed a model of motion detection using a variety of frame differences methods. The HTTP is used as communication protocol. The best algorithm is expected to be used for web-based home surveillance applications.

There are many ways to detect the moving object in term of image motion detection. A common method that is used to detect moving object recognize by comparing two or more sequence images. Comparing image by analyzing all of image pixel is known as frame differences method. Template matching is a method that used to determine the reference image. Reference image which determined dynamically is known a dynamic template matching. This research proposes an algorithm to determine the reference image by using dynamic template matching adaptively. In the system, there are three ways to determine the reference image base on environment condition. This research realizes an algorithm by using web based system and using IP Camera as measures device. This algorithm provide detection accuracy rate 95.5%.

Internet is one of technology that is used for various purposes. Internet is used to support business processes, education, information, social networking, and even for security systems are installed in offices, homes, and buildings. About the security system, camera is a device commonly used for monitoring in the case of security system. IP Camera is one of the devices commonly used to do surveillance with Internet. This research proposed the design of home surveillance system that uses the Internet. The system has been equipped with the ability to detect motion and web service as a means of system interoperability. Design of the system will be started from a review of home networking, design of web based IP Camera management system, motion detection algorithm, and web services. The method utilised for motion detection is frame differences and dynamic template matching. Furthermore, this research does the modification of dynamic template matching method to enhance the system. The method termed as dynamic and adaptive template matching (DATM). This research successfully completed a web-based home surveillance system that is equipped with motion detection system and the web service. The system is called by iHome. IHome works by using several tools including: IP Camera; ADSL; Mini PC; and Switch. Based on algorithm comparison, DATM algorithm produces a better accuracy when compared with dynamic template matching t-1 (DTM t-1) algorithm. The accuracy rate of detection DATM method reached 95.5%, better than DTM t-1 method with 89.8% accuracy rate.