The fundamental characteristics of agricultural products are appearance, size, and weight, which affect their market value, consumer preference, and choice. Thus, food and agricultural industries seek rapid, simple, and nondestructive approaches to assess real-time measurements at the post-harvest stage before packaging for the consumer market. While sorting and grading may be performed by humans, it is unreliable, time-consuming, complicated, subjective, onerous, expensive, and easily influenced by surroundings. Therefore, an astute sorting and grading method for tomato fruit is required. We evaluated two tomato configurations on a conveyor belt: single tomatoes (no occlusion) and multi-tomatoes (partially occluded). We used polygon approximation for concave and convex point extraction algorithms to segment the occluded tomatoes. We developed seven models for regression using single-tomato image features. The Bayesian regularization artificial neural network outranked all the trained models in weight estimation with a root-mean-square error (RMSE) of 1.468 g and R² of 0.971. For volume estimation, the RBF SVM had the best performance with R² of 0.982 and RMSE of 1.2683 cm³. It is feasible to implement a proposed system as a noninvasive in-line sorting technique for tomatoes.

The management of fertilizers in a context of climate change and the preservation of the environment is strongly related to the regulation and accumulation of 2-acetyl-1-pyrroline (2AP) in fragrant rice. However, the feasibility of such management strategies in terms of enhancing the accumulation of 2AP has not yet been explored in aromatic cultivars. Here, we investigated the impact of the application of two fertilizers at three depth (surface, 5 cm and 10 cm) levels of placements to improve the aromatic rice quality, including such aspects as the 2AP content-, protein-, amylose- and yield-related traits. For this purpose, two known rice cultivars, Basmati 385 (B-385) and Yunjingyou (YJY), were grown in pots during 2019 and 2020 under fluctuating climates. The deep application of fertilizer at 10 cm significantly affected the 2AP content with such values as 127.53 μg kg⁻¹ and 111.91 μg kg⁻¹ obtained for Fragrant Fertilizer (FF) and Urea in B-385 cultivar, and 126.5 μg kg⁻¹ and 114.24 μg kg⁻¹ being observed for FF and Urea in YJY, respectively, during 2019. In addition, values of 108.41 μg kg⁻¹ and 117.35 μg kg⁻¹ were recorded for FF and Urea in B-385, while 125.91-μg kg⁻¹ and 90.71-μg kg⁻¹ were measured for FF and Urea in YJY, respectively, during 2020. Similarly, B-385 had better 2AP content and yield-related traits, as well as amylose content and cooked rice elongation, as compared to the YJY rice cultivar. The 2AP accumulation and its related biochemical parameters, and their relationships in different plant tissues at different growth stages under FF and Urea treatments, were also improved. Further, the 2AP content and the P5C activity demonstrated strong correlations during the grain filling periods in both fragrant rice cultivars. In conclusion, our findings have the potential to provide useful information to farmers and agriculture extension workers in terms of the saving of fertilizers and the improvement of rice grain quality under fluctuating climate conditions.

Leaf chlorophyll content is an important indicator of photosynthetic capacity and health status and plays a key role in monitoring plant growth. At present, research on chlorophyll content monitoring has mainly focused on establishing a relationship between vegetation index and chlorophyll content, which was composed of several hyperspectral wavelengths, resulting in the loss of effective information contained in certain wavelengths strongly associated with disease. Therefore, it was very important to select spectral wavelengths to reveal the hyperspectral response of chlorophyll content in diseased leaves. Fifty groups of rice leaves infected with bacterial leaf blight were studied, and hyperspectral images and soil plant analysis development (SPAD) values of disease‐stressed rice leaves were collected every 5 days from the first day after inoculation. A method for predicting the SPAD value of diseased rice leaves using random forest (RF) combined with a back‐propagation optimized neural network (BPNN) was proposed. The bird swarm algorithm (BSA) and the immune algorithm (IA) were used to solve the problems of information loss, local minimization, and slow convergence of the BPNN. The BPNN, BSA–BPNN, IA–BPNN, and BSA–IA–BPNN prediction models for the SPAD value of diseased rice leaves are established. The results show that the BSA–IA–BPNN model with characteristic wavelengths selected by RF achieved the best prediction effect, and the prediction sets R², root‐mean‐square error, and relative error are 0.97%, 4.34%, and 3.33% respectively. The shortest prediction modeling time is 1.54 s, and the model exhibits the best stability. This study provides allegorical support and a theoretical basis for achieving rapid, nondestructive detection of the SPAD value of diseased rice leaves, provides an effective reference for estimating the SPAD value of diseased rice leaves in the field, and is of great importance for rice disease monitoring.