An IoT‐Enabled Growth Management System With Deep Learning‐Based Monitoring for Volvariella volvacea Mushroom Cultivation in Controlled Environments

ABSTRACT

The nutritional profile and medicinal properties of the Volvariella mushrooms make them highly demanded in Malaysia. However, with increasing demand, traditional methods frequently struggle to maintain the ideal environmental conditions necessary for optimal growth, leading to variable yields and quality of the products. To overcome these challenges, we develop a system of an internet of things (IoT) based environmental control system combined with a deep learning‐based growth‐stage monitoring system for Volvariella culture. The suggested IoT system uses the ESP32 microcontroller to communicate with the DHT22 sensor for measuring temperature and humidity, the MQ135 sensor for measuring carbon dioxide (CO ₂ ) and the HW‐038 sensor for detecting the water level. Actuators like foggers, water pumps, heaters, and ventilation fans are used by the system to continuously monitor and automatically control critical environmental parameters. Optimum cultivation conditions were maintained in the range of 28°C–35°C, 75%–95% RH, and CO ₂ level less than 1000 ppm. A YOLOv8 deep learning model was trained to detect the mushroom growth stage, using mushroom images with annotations of four stages: tiny, button, egg, and mature. The experimental results showed that the proposed environmental control system was able to keep the target cultivation parameters within the target range for about 85% of the cultivation period. The overall classification accuracy of the growth‐stage recognition model was 93.12%, macro‐average F1 score was 87.9%, and the mean average precision (mAP@0.5) was 0.78. Real‐time monitoring and control were achieved using the Blynk mobile application, which enabled remote supervision and significantly reduced the need for continuous manual monitoring and intervention. Passive ventilation effects were observed, but the system provided generally good growing conditions and better cultivation reliability in the face of minor fluctuations in the environment. The use of smart environmental control and vision‐based analysis increases yield repeatability, quality, reduces human error and manual operation. The presented results clearly show the capabilities of the proposed system to equip mushroom cultivation with scalability and sustainability, thus improving the productivity of agricultural production and food security.