IJA_2026v16n5

International Journal of Aquaculture, 2026, Vol.16, No.5, 378-396 http://www.aquapublisher.com/index.php/ija 381 start and stop of aerators according to changes in dissolved oxygen, adjust feeding plans based on water temperature, weather, and feeding behavior of cultured organisms, and monitor the operating status of water pumps, valves, power supply equipment, and other facilities, thereby improving the timeliness, precision, and automation level of pond management. In addition, video monitoring, meteorological monitoring, production data management, and intelligent decision-making technologies are gradually being applied in ecological pond aquaculture. Video monitoring can assist in observing the activity status, feeding behavior, and abnormal pond conditions of cultured organisms; meteorological monitoring can provide reference for judging risks such as hypoxia, stress, and abrupt water quality changes; production data management can record seedling stocking, feed use, medication, water quality changes, and product harvest information, providing support for production process traceability and quality and safety management. More advanced smart aquaculture solutions further combine the Internet of Things with fuzzy logic, machine learning, and other methods for autonomous control and predictive analysis of pond environments (Kanwal et al., 2024). Together, these technologies constitute a multi-level application system of digital monitoring in ecological pond aquaculture, enabling aquaculture management to gradually expand from single water quality detection to environmental perception, equipment control, risk warning, and production decision-making. 2.3 Operating mechanism of digital monitoring systems The operating mechanism of digital monitoring systems usually includes data collection, data transmission, data analysis, early-warning feedback, and equipment linkage. From the perspective of technical architecture, digital monitoring systems in ecological ponds generally follow a layered structure from the perception layer to the network layer, platform layer, and application layer. At the perception layer, underwater or in situ sensors deployed in ponds continuously measure water parameters such as pH, water temperature, dissolved oxygen, turbidity, electrical conductivity, salinity, and water level (Flores-Iwasaki et al., 2025). Sensor output signals are usually collected by microcontrollers such as STM32, Arduino, ESP32, or ESP8266, which complete signal conditioning, pre-calibration, in situ calibration, and preliminary processing (Shete et al., 2024). Compared with traditional manual testing, digital acquisition has the advantages of strong continuity, rapid response, and complete data preservation, making it possible to more accurately reflect the dynamic changes of pond environments. At the network and platform layers, the collected data can be transmitted to local gateways, telecom cloud platforms, or dedicated servers through NB-IoT, Wi-Fi, GSM/SMS, LoRaWAN, and other communication methods (Flores-Iwasaki et al., 2025). Subsequently, the data are stored and managed in IoT cloud services such as ThingSpeak, AWS, and Firebase, or in dedicated data platforms, and are visualized through web dashboards, computer terminals, or mobile applications, thus enabling remote and real-time pond supervision (Lin et al., 2021). Aquaculture operators can use mobile phones, computers, and other terminals to view trends in water quality indicators, equipment operating status, and abnormal alarm information at any time. When the system detects conditions such as excessively low dissolved oxygen, large pH fluctuations, increased ammonia nitrogen concentration, or abnormal water level, it can issue early-warning prompts in a timely manner, helping operators take measures such as aeration, water exchange, water conditioning, or reduced feeding as early as possible. In more complete digital monitoring systems, data analysis and equipment control can form a linkage mechanism. Control logic is usually implemented through threshold rules, fuzzy logic, or machine learning models, which compare real-time data with ecological setpoints and trigger actuators such as aerators, water pumps, automatic feeders, or dosing systems to correct environmental deviations (Chen et al., 2022). For example, when dissolved oxygen falls below the preset threshold, the system can automatically start aeration equipment; when pH is unstable, the system can prompt or execute regulation measures; when monitoring parameters exceed the limit, relay-controlled water pumps can be used for water exchange (Gleiser and Moro, 2023). More advanced systems can also integrate large amounts of historical data and intelligent algorithms to predict water quality dynamics and

RkJQdWJsaXNoZXIy MjQ4ODYzNA==