Yield monitoring is crucial for the agricultural sector, as it can be used to inform decisions on harvesting, storage, and transportation. Traditionally, several statistical methods and visual inspection techniques are employed to get an early estimate of the final yield of citrus, with the downside of being inaccurate, costly, and time-consuming. In recent years, there have been a lot of advancements in the fields of Artificial Intelligence (AI) and computer vision, providing opportunities to automate plenty of things in different domains, including agriculture. This research proposes a deep learning-based framework that leverages multiple Convolutional Neural Networks (CNN) to efficiently and effectively operate in real-world environments, using field data to provide accurate, improved yield estimates. A high-quality dataset, consisting of citrus tree images, is obtained from orchards at the university research farm Koont and the National Agriculture Research Center (NARC). Afterwards, the CNN-based models are trained thoroughly with various configurations and data augmentation techniques. All the models are rigorously tested and evaluated on the basis of a number of performance metrics. Experiments have shown that YOLOv8m performs with the highest mean average precision, reaching up to 90% with an inference time of a few milliseconds, making it worthy to be deployed for fruit detection, counting, and yield estimation tasks.

Potato is one of the key crops for ensuring food security and can be a potential substitute for cereal crops due to its high yielding nature and nutritional value. Crop nutrient management practices within potato fields are implemented uniformly without considering crop requirements and soil variability, causing uneven and low yield. However, yield can be increased by identifying growth and yield-limiting factors. Geospatial tools are robust and effective in identifying the spatial variations within the field. Proximal sensing allows quick analysis of soil and plant characteristics, decreases the need for laborious and expensive soil and plant sampling, and strengthens precision agriculture techniques. The aim of the study was to quantify the soil spatial variability and identify potato crop growth and yield limiting factors for the optimization of inputs. Two fields were selected in the subtropical region of Pakistan (Koont, Rawalpindi), and each field was cultivated with two different potato varieties. A grid sampling approach was developed to collect soil samples and tuber yields. The soil was tested for nitrogen (N), phosphorus (P), potassium (K), pH, electrical conductivity (E.C), temperature, and moisture content (M.C) by using a soil proximal sensor. Normalized difference vegetation index (NDVI) was recorded using a handheld GreenSeeker, and chlorophyll was estimated using a chlorophyll meter. Descriptive statistics and correlation analysis for soil and crop parameters were performed in Minitab 21, while geostatistical analysis was performed in Arc Map 10.8 to show spatial variability and to generate kriged maps of different soil properties. The coefficient of variation of soil properties and plant parameters showed moderate to high variability within the field, except for pH and temperature. The correlation matrix suggested that N, P, K, E.C., chlorophyll, NDVI, plant height, and leaf area had a significant relationship with potato yield. Most of the soil and plant parameters had a medium to high range of influence (20 to 90 m) and varied greatly within the field. Kriged maps of plant and soil parameters also showed spatial variations and were aligned with descriptive statistics and correlations. Quantification of soil spatial variability within potato fields can assist in measuring yield-limiting soil characteristics to establish management zones for variable rate fertilization for optimum tuber yield and low environmental impact.