Pengembangan Sistem Aplikator Fungisida dan Pendeteksi Penyakit Anthracnose pada Cabai Berbasis Convolutional Neural Network

Abstract

Prototipe sistem aplikator fungisida dan deteksi penyakit anthracnose nir-awak dengan pusat kendali telah berhasil dibuat berdasarkan kebutuhan petani dalam pemantauan dan pengendalian penyakit anthracnose memperhatikan aspek keamanan, kesehatan, efisiensi waktu, dan sumber daya manusia. Sistem terbagi menjadi bagian kendali, aplikator, dan pendeteksi. Metode pengembangan terbagi menjadi perancangan mekanik dan sistem pada pusat kendali terdiri dari tombol sistem gerak, aplikator, dan monitoring, aplikator terdiri dari stasiun fungisida, pompa, dan nozzle sprayer, dan pendeteksi terdiri dari sensor kamera dan layar monitoring. Hasil pengujian bogie berhasil bergerak forward dan reverse pada jalur yang dilengkapi limit switch berpenggerak NEMA 17. Aplikator berhasil mendistribusikan fungisida menggunakan pompa DC 12V 80 PSI. Sensor kamera Xiaovv berhasil terhubung ke layar monitoring. Hasil pengujian aplikator dengan sudut 1080, debit 0,0417 Liter/detik, dan kecepatan penggerak 200 RPM menghasilkan luas pengaplikasian 12,55 m2/menit. Sistem monitoring pendeteksian Anthracnose meghasilkan nilai accuracy 87,5%, precision 88% dan recall 86,8% dengan kecepatan penggerak 100 RPM.

Prototype fungicide applicator system and unmanned anthracnose disease detection with a control center have been successfully made based on the needs of farmers in monitoring and controlling anthracnose disease considering aspects of safety, health, time efficiency, and human resources. The system is divided into control, applicator, and detection parts. The development method is divided into mechanical and system design at the control center consisting of motion system buttons, applicators, and monitoring, applicators consisting of fungicide stations, pumps, and sprayer nozzles, and detectors consisting of camera sensors and monitoring screens. The bogie successfully moves forward and reverse on a path equipped with a NEMA 17 driven limit switch. The applicator successfully distributes fungicides using a 12V 80 PSI DC pump. The Xiaovv camera sensor was successfully connected to the monitoring screen. The test results of the applicator with an angle of 1080, a discharge of 0.0417 liters/second, and a drive speed of 200 RPM resulted in an application area of 12.55 m2/minute. The Anthracnose detection monitoring system produced an accuracy value of 87.5%, precision 88% and recall 86.8% with a drive speed of 100 RPM.