Fully automated yield estimation of intact fruits prior to harvesting provides various benefits to farmers. Until now, several studies have been conducted to estimate fruit yield using image-processing technologies. However, most of these techniques require thresholds for features such as color, shape and size. In addition, their performance strongly depends on the thresholds used, although optimal thresholds tend to vary with images. Furthermore, most of these techniques have attempted to detect only mature and immature fruits, although the number of young fruits is more important for the prediction of long-term fluctuations in yield. In this study, we aimed to develop a method to accurately detect individual intact tomato fruits including mature, immature and young fruits on a plant using a conventional RGB digital camera in conjunction with machine learning approaches. The developed method did not require an adjustment of threshold values for fruit detection from each image because image segmentation was conducted based on classification models generated in accordance with the color, shape, texture and size of the images. The results of fruit detection in the test images showed that the developed method achieved a recall of 0.80, while the precision was 0.88. The recall values of mature, immature and young fruits were 1.00, 0.80 and 0.78, respectively.

Flowering (spikelet anthesis) is one of the most important phenotypic characteristics of paddy rice, and researchers expend efforts to observe flowering timing. Observing flowering is very time-consuming and labor-intensive, because it is still visually performed by humans. An image-based method that automatically detects the flowering of paddy rice is highly desirable. However, varying illumination, diversity of appearance of the flowering parts of the panicles, shape deformation, partial occlusion, and complex background make the development of such a method challenging. We developed a method for detecting flowering panicles of rice in RGB images using scale-invariant feature transform descriptors, bag of visual words, and a machine learning method, support vector machine. Applying the method to time-series images, we estimated the number of flowering panicles and the diurnal peak of flowering on each day. The method accurately detected the flowering parts of panicles during the flowering period and quantified the daily and diurnal flowering pattern. A powerful method for automatically detecting flowering panicles of paddy rice in time-series RGB images taken under natural field conditions is described. The method can automatically count flowering panicles. In application to time-series images, the proposed method can well quantify the daily amount and the diurnal changes of flowering during the flowering period and identify daily peaks of flowering.

Accurate estimation of heading date of paddy rice greatly helps the breeders to understand the adaptability of different crop varieties in a given location. The heading date also plays a vital role in determining grain yield for research experiments. Visual examination of the crop is laborious and time consuming. Therefore, quick and precise estimation of heading date of paddy rice is highly essential. In this work, we propose a simple pipeline to detect regions containing flowering panicles from ground level RGB images of paddy rice. Given a fixed region size for an image, the number of regions containing flowering panicles is directly proportional to the number of flowering panicles present. Consequently, we use the flowering panicle region counts to estimate the heading date of the crop. The method is based on image classification using Convolutional Neural Networks. We evaluated the performance of our algorithm on five time series image sequences of three different varieties of rice crops. When compared to the previous work on this dataset, the accuracy and general versatility of the method has been improved and heading date has been estimated with a mean absolute error of less than 1 day. An efficient heading date estimation method has been described for rice crops using time series RGB images of crop under natural field conditions. This study demonstrated that our method can reliably be used as a replacement of manual observation to detect the heading date of rice crops.

Understanding interactions of genotype, environment, and management under field conditions is vital for selecting new cultivars and farming systems. Image analysis is considered a robust technique in high-throughput phenotyping with non-destructive sampling. However, analysis of digital field-derived images remains challenging because of the variety of light intensities, growth environments, and developmental stages. The plant canopy coverage (PCC) ratio is an important index of crop growth and development. Here, we present a tool, EasyPCC, for effective and accurate evaluation of the ground coverage ratio from a large number of images under variable field conditions. The core algorithm of EasyPCC is based on a pixel-based segmentation method using a decision-tree-based segmentation model (DTSM). EasyPCC was developed under the MATLAB^® and R languages; thus, it could be implemented in high-performance computing to handle large numbers of images following just a single model training process. This study used an experimental set of images from a paddy field to demonstrate EasyPCC, and to show the accuracy improvement possible by adjusting key points (e.g., outlier deletion and model retraining). The accuracy (R² = 0.99) of the calculated coverage ratio was validated against a corresponding benchmark dataset. The EasyPCC source code is released under GPL license with benchmark datasets of several different crop types for algorithm development and for evaluating ground coverage ratios.

We use recent explainable artificial intelligence (XAI) based on SHapley Additive exPlanations (SHAP), Local Interpretable Model-agnostic Explanations (LIME), and Light Gradient Boosting Machine (LightGBM) to analyze diverse physical agricultural (agri-) worker datasets. We have developed various promising body-sensing systems to enhance agri-technical advancement, training and worker development, and security. However, existing methods and systems are not sufficient for in-depth analysis of human motion. Thus, we have also developed wearable sensing systems (WS) that can capture real-time three-axis acceleration and angular velocity data related to agri-worker motion by analyzing human dynamics and statistics in different agri-fields, meadows, and gardens. After investigating the obtained time-series data using a novel program written in Python, we discuss our findings and recommendations with real agri-workers and managers. In this study, we use XAI and visualization to analyze diverse data of experienced and inexperienced agri-workers to develop an applied method for agri-directors to train agri-workers.

The measurement of air temperature is strongly influenced by environmental factors such as solar radiation, humidity, wind speed and rainfall. This is problematic in low-cost air temperature sensors, which lack a radiation shield or a forced aspiration system, exposing them to direct sunlight and condensation. In this study, we developed a machine learning-based calibration method for air temperature measurement by a low-cost sensor. An artificial neural network (ANN) was used to balance the effect of multiple environmental factors on the measurements. Data were collected over 305 days, at three different locations in Japan, and used to evaluate the performance of the approach. Data collected at the same location and at different locations were used for training and testing, and the former was also used for k-fold cross-validation, demonstrating an average improvement in mean absolute error (MAE) from 1.62 to 0.67 by applying our method. Some calibration failures were noted, due to abrupt changes in environmental conditions such as solar radiation or rainfall. The MAE was shown to decrease even when the data collected in different nearby locations were used for training and testing. However, the results also showed that negative effects arose when data obtained from widely-separated locations were used, because of the significant environmental differences between them.