Mubin, Mohammad Nasrul (2021) Sistem Informasi Tempat Parkir Luar Ruangan Berbasis Pemrosesan Citra Dengan Multi Kamera. Masters thesis, Institut Teknologi Sepuluh Nopember.
Text
07111850042002-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (7MB) | Request a copy |
Abstract
Informasi ketersediaan tempat parkir, terutama di area parkir yang luas sangat dibutuhkan masyarakat pengguna parkir. Hal ini dikarenakan informasi tersebut akan menghemat waktu, tenaga, dan bahan bakar (uang). Saat ini, metode deteksi ruang parkir terbagi menjadi deteksi gambar dan non gambar. Deteksi non gambar tidak sepenuhnya baik, bahkan beberapa implementasinya membutuhkan pembongkaran jalan atau bangunan. Selain itu, perawatan yang diperlukan untuk deteksi non gambar dapat dikatakan lebih mahal, sehingga metode deteksi gambar dipilih dalam penelitian ini. Multi kamera akan digunakan untuk menangkap citra area parkir, dimana dengan pemrosesan citra maka akan diketahui tempat parkir yang kosong atau tersedia. Proses pengolahan citra multi kamera dalam penelitian ini akan berlangsung dengan memanfaatkan algoritma multithread. Sistem dirancang dengan terlebih dahulu mengonversi citra-citra yang ditangkap oleh multi kamera ke dalam ruang warna lain, yaitu HSV yang lebih tahan terhadap perubahan iluminasi cahaya. Hasil dari konversi tersebut kemudian diproses dengan contrast limited adaptive histogram equalization (CLAHE). Langkah berikutnya dilakukan transformasi dimensi citra dengan transformasi perspektif. Citra-citra yang telah ditransformasikan tersebut, kemudian dilakukan identifikasi status tiap slot parkir dengan memanfaatkan lingkaran bantu yang ada dalam tiap slot tersebut. Identifikasi dilakukan dengan merubah citra tiap slot tersebut ke dalam citra biner. Dari citra biner ini dapat diidentifikasi status tiap slot parkir dengan algoritma deteksi blob lingkaran. Data hasil deteksi kemudian akan diinformasikan ke pengendara. Penginformasian dilakukan dengan sistem IoT melalui web dengan jaringan lokal. Menurut analisa yang telah dilakukan, sistem memiliki persentase sensitivitas 100% pada semua slot, spesifisitas dan akurasi di atas 90% pada 6 dari 8 slot yang diidentifikasi, serta 100% tingkat kesuksesan dalam penampilan data identifikasi dalam web lokal.
================================================================================================
Information on the availability of parking lots, especially in large parking areas are needed by the public parking users. This is because the information will save time, energy, and fuel (money). Currently, parking space detection methods are divided into image detection and non-images. Non-image detection is not entirely good, even some of its implementations require demolition of roads or buildings. In addition, the treatments required for non-image detection can be said to be more expensive, so the image detection method was chosen in this study. Multi cameras will be used to capture imagery of the parking area, where by image processing it will be known that the parking lot is occupied or available. The multi camera image processing process in this study will take place by utilizing multithreaded algorithms. The system is designed for convert the captured images by multiple cameras into another colour space first, the HSV, which is more resistant to light illumination changes. The result of the conversion is then processed with contrast limited adaptive histogram equalization (CLAHE). The next step is to transform the image dimension with perspective transformation. The transformed images are then identified by identifying the status of each parking slot by utilizing the auxiliary circles in each parking slot. Identification is done by converting the image of each slot into a binary image. From this binary image, the status of each parking slot can be identified using the circle blob detection algorithm. The detection data will then be informed to the driver/user. Information delivery is done with IoT systems over the web with local networks. According to the analysis that has been carried out, the system has a 100% percentage percentage on all slots, above 90% specificity and accuracy on 6 of the 8 identified slots, as well as a 100% success rate in displaying data on the local web.
Item Type: | Thesis (Masters) |
---|---|
Uncontrolled Keywords: | circle blob detection, image processing, internet of things, multi kamera, tempat parkir, circle blob detection, image processing, internet of things, multi camera, lot parking |
Subjects: | T Technology > TA Engineering (General). Civil engineering (General) |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis |
Depositing User: | Mohammad Nasrul Mubin |
Date Deposited: | 15 Aug 2021 21:10 |
Last Modified: | 15 Aug 2021 21:10 |
URI: | http://repository.its.ac.id/id/eprint/86933 |
Actions (login required)
View Item |