Automated crop monitoring using image analysis is commonly used in horticulture. Image-processing technologies have been used in several studies to monitor growth, determine harvest time, and estimate yield. However, accurate monitoring of flowers and fruits in addition to tracking their movements is difficult because of their location on an individual plant among a cluster of plants. In this study, an automated clip-type Internet of Things (IoT) camera-based growth monitoring and harvest date prediction system was proposed and designed for tomato cultivation. Multiple clip-type IoT cameras were installed on trusses inside a greenhouse, and the growth of tomato flowers and fruits was monitored using deep learning-based blooming flower and immature fruit detection. In addition, the harvest date was calculated using these data and temperatures inside the greenhouse. Our system was tested over three months. Harvest dates measured using our system were comparable with the data manually recorded. These results suggest that the system could accurately detect anthesis, number of immature fruits, and predict the harvest date within an error range of ±2.03 days in tomato plants. This system can be used to support crop growth management in greenhouses.

Osmotic stress enhances fruit quality, including the dry matter content, in tomatoes (Solanum lycopersicum L.). This study aimed at providing further insight into the precision control of fruit yield and quality on the long-term moderate osmotic stress conditions in tomato fruit production. We compared the growth pattern between fruits of two cultivars, typical Japanese and Dutch cultivars, under two different nutrient concentrations (2.3 and 5.0 dS·m⁻¹) to understand the effect of electrical conductivity (EC) on dry mass and water content of fruits. The experiment was performed with a rockwool bag culture system in a controlled greenhouse. Increasing EC resulted in an approximately 20% decrease in fruit yield and a 0.5–1% increase in fruit dry matter content in both cultivars. This yield reduction was not caused by the fruit number, but by an approximately 25% decrease in individual fresh fruit weight. Non-linear models were used to describe the changes in dry matter content, water content, and dry weight of tomato fruit as a function of cumulative temperature. The decay rate of dry matter content in the fruit decreased with high EC treatments in the Japanese cultivar. The points at which the rates of changes in water and dry weight increased the in fruit were around 585 and 480 °C-days after anthesis, respectively, under the low EC condition. Rates of water increase in the fruit were changed by high EC treatment, while the shifts were opposite with respect to the cultivars. Dry weight increase in the fruit was not affected by EC treatment. Collectively, our findings clarify the effect of EC on the fruit growth characteristics of Japanese and Dutch tomato cultivars, and provide new insights into the yield of high-Brix tomato cultivation.

Localized management based on multipoint measurements of crops and soil is expected to improve agricultural productivity. The difficulties with this process are as follows: it is time-consuming due to the many measurement points; it requires various sensors for various measurements; it can lead to unstable measurements due to sunlight and wind. To solve the above issues, the system we propose has the advantages of efficient measurements performed by a robot, various measurements performed through exchangeable sensor units, and stable measurements through protecting the crop from sunlight and wind. As a pilot study for the system, we conducted an experiment to simultaneously measure the crops and soil in a cabbage field. The robot achieved mobility in the field, >4 h of operation time, and the ability to obtain soil electrical conductivity measurements and crop imaging at 100 points. Furthermore, the stability of the measurement conditions within the sensor unit during the experiment was evaluated. Compared to the case without the covering, the illuminance became 280-fold stabler (standard deviation = 0.4 lx), and the wind-induced crop shaking became 20-fold lower (root mean square error of the image pixels = 0.5%). The potential impacts of this research include high reproducibility because of the shareable sensor unit and the expectation of new discoveries using precise indoor sensors.

To improve the efficiency of nitrogen (N) fertilization, it is necessary to perform rapid direct measurements in the field rather than time-consuming laboratory-based chemical analysis. Herein, crop and soil data from the early stages of cabbage growth were acquired through two fall cultivations. Chlorophyll meter value, height, and projected leaf area were evaluated as crop indicators. A positive correlation was observed between the projected leaf area or its rate of increase 2–3 weeks after transplantation and head fresh weight (FW). After comparing two water-content reflectometers (WCR) and a nitrate sensor, we selected a WCR with a 12 cm-long rod as the soil indicator. The diagnostic method was verified using varying amounts of N basal fertilizer during spring cultivation. The variable rate of N top dressing (25, 50, and 75% total N) based on the electrical conductivity (EC) 14 days after transplantation reduced the subsequent EC variability. No differences in head FW were observed between the treatments. A 25% reduction in N fertilizer was considered possible for half of the plots. The quantity of inorganic N extracted by potassium chloride from the crop soil after cultivation was unaffected by the amount of N fertilizer. Therefore, the diagnostic method proposed herein is suitable for soil N management.

Non-destructive diameter estimation of cabbage heads and yield prediction employing Unmanned Aerial Vehicle (UAV) imagery are superior to conventional approaches, which are labor intensive and time consuming. This approach assesses spatial variability across the field, effective allocation of resources, and supports variable application rates of fertilizer and supply chain management. Here, individual cabbage head diameters were estimated using deep learning-based pose estimation models (YOLOv8s-pose and YOLOv11s-pose) using high spatial resolution RGB images acquired from UAV 6 m during the cabbage-growing season in 2024. With a mean relative error (MRE) of 4.6% and a high mean average precision (mAP) 98.5% at 0.5, YOLOv11s-pose emerged as the best-performing model, verifying its accuracy for pragmatic agricultural use. The approximated diameter was then combined with climatic variables (temperature and rainfall) and canopy reflectance indices (normalized difference vegetation index (NDVI), normalized difference red edge index (NDRE), and green chlorophyll index (CIg)) that were extracted from the multispectral images with 6 m resolution and fed into AI models to develop individual cabbage head fresh weight. Among the machine learning models (MLMs) tested, CatBoost achieved the lowest Mean Squared Error (MSE = 0.025 kg/cabbage), highest R² (0.89), and outperformed other models based on the Diebold–Mariano statistical test (p < 0.05). This finding suggests that an integrated AI-powered framework enhances non-invasive and precise yield estimation in cabbage farming.