登革熱病媒蚊即時偵測系統之設計與實現

登革熱為一種長期存在於臺灣的傳染性疾病，在臺灣主要的傳播媒介為埃及斑蚊與白線斑蚊。受到登革熱感染的人體，臨床症狀有發燒、身體疼痛等情形，嚴重者將導致呼吸衰竭甚至休克死亡。目前，尚無完善之疫苗可有效抑制登革熱病毒傳播。因此，在登革熱疫情控制方面，衛生單位主要工作重點皆以病媒蚊密度監控為主，藉由登革熱病媒蚊分布情形，實施預先防治工作。近年來，在臺灣針對登革熱病媒蚊密度監控多以人力方式進行，由於傳染病防治重視時效性，此種人力監測方式將無法針對登革熱疫情爆發前做出迅速的反應，使得現今病媒蚊密度監測工作未具有即時防治與通報之能力。隨著感測器技術與資通訊技術的蓬勃發展，探討野生生物之自動化系統在生態環境監控上已有許多應用。本研究利用數位信號處理器為核心，結合紅外線感測器應用電路於蚊類振翅頻率偵測，並以全球行動通訊系統(Global System for Mobile Communications)模組作為感測資料的無線傳輸媒介，藉以發展一套自動化登革熱病媒蚊偵測系統。本研究以MATLAB軟體進行埃及斑蚊與白線斑蚊之辨識演算法開發，藉由快速傅立葉轉換(Fast Fourier Transform, FFT)技術將蚊類振翅信號由時域轉換至頻域進行分析，並導入適應性網路模糊推論系統理論(Adaptive-network-based Fuzzy Inference System, ANFIS)，對登革熱病媒蚊辨識分類研究。本研究提出之偵測系統可有效地對於病媒蚊進行偵測，並對病媒蚊振翅信號進行分析，而實驗結果亦顯示，利用ANFIS之機器學習演算法，對於病媒蚊種類辨識具有良好之結果。

關鍵字

登革熱病媒蚊；自動化偵測系統；快速傅立葉轉換；適應性網路模糊推論系統

並列摘要

Dengue fever has been an infectious disease in Taiwan for a long time. In Taiwan, dengue fever is transmitted mainly by Aedes aegypti and Aedes albopictus. There are some symptoms if getting infected, such as having a high temperature and body aches. In the worst situation, dengue fever may lead to respiratory failure and shock. There is no effective vaccine that is capable of controlling dengue fever so far. Thus, in order to control dengue fever spreading, the main task for authorities in health departments is to monitor the density of dengue mosquitos, so dengue fever can be prevented in advance. In recent years, dengue mosquito monitoring has been done through manual inspection. Because timeliness is very important for dengue fever prevention, manual inspection is generally not able to effectively catch real-time information related to mosquito spreading. Nor is the monitoring method capable of sending out warning messages in a real-time manner. With the thriving development of the sensor and info-communication technology, many auto-monitoring systems have been designed for ecological and environmental monitoring. This study uses a digital signal processor as a kernel in a monitoring system, and a circuit with infrared sensors t is used to detect the frequency generated by the mosquito wings. Also, to achieve real-time monitoring, a GSM module is included in the monitoring system for wireless transmission. Furthermore, a classification algorithm for Aedes aegypti and Aedes albopictus is developed by the MATLAB software in the study. To analyze the wings-waving signal of dengue mosquitos, the fast Fourier transform (FFT) is used to transform the signal from the time domain to the frequency domain. Then, an adaptive-network-based fuzzy inference system (ANFIS) is applied to classify the species of dengue mosquitos. The monitoring system in the study could effectively detect dengue mosquitos and accurately analyze the wings-waving signal. The experiment results also show that the ANFIS algorithm yields a great result in dengue mosquito classification. Keywords: dengue mosquito, automatic detecting system, fast Fourier transform, ANFIS