Thousand-grain weight is the main parameter for accurately estimating rice yields, and it is an important indicator for variety breeding and cultivation management. The accurate detection and counting of rice grains is an important prerequisite for thousand-grain weight measurements. However, because rice grains are small targets with high overall similarity and different degrees of adhesion, there are still considerable challenges preventing the accurate detection and counting of rice grains during thousand-grain weight measurements. A deep learning model based on a transformer encoder and coordinate attention module was, therefore, designed for detecting and counting rice grains, and named TCLE-YOLO in which YOLOv5 was used as the backbone network. Specifically, to improve the feature representation of the model for small target regions, a coordinate attention (CA) module was introduced into the backbone module of YOLOv5. In addition, another detection head for small targets was designed based on a low-level, high-resolution feature map, and the transformer encoder was applied to the neck module to expand the receptive field of the network and enhance the extraction of key feature of detected targets. This enabled our additional detection head to be more sensitive to rice grains, especially heavily adhesive grains. Finally, EIoU loss was used to further improve accuracy. The experimental results show that, when applied to the self-built rice grain dataset, the precision, recall, and mAP@0.5 of the TCLE–YOLO model were 99.20%, 99.10%, and 99.20%, respectively. Compared with several state-of-the-art models, the proposed TCLE–YOLO model achieves better detection performance. In summary, the rice grain detection method built in this study is suitable for rice grain recognition and counting, and it can provide guidance for accurate thousand-grain weight measurements and the effective evaluation of rice breeding.

Rice is the staple diet for most of the world’s population and is considered a major staple crop in China. Anhui province of China is among the leading provinces for rice production, consumption, and commodities; it is well-known as the “land of fish and rice”. Japonica rice cultivation in Anhui Province is mainly categorized into late-maturing medium, early-maturing medium, and early-maturing late japonica. This review explores the suitable ecological zone distribution and corresponding climate characteristics of the three types of japonica rice in Anhui Province. Data on japonica rice varieties approved in the province over the past twenty years were collected, illustrating the development process of japonica rice varieties in the province and their quality and resistance to rice blast disease. The review shows that the yield is positively and significantly correlated with agronomic traits, such as the number of effective panicles and the total number of grains per panicle, plant height, etc. In addition, it elucidates the major problems faced by Anhui’s japonica rice breeding and cultivation, such as frequent events of high temperatures, rice blast disease, and medium to low soil fertility levels. Considering the existing issues in breeding japonica rice varieties in Anhui Province, this review proposes a strategy for breeding high-yield and disease-resistant japonica rice varieties, particularly varieties adaptable to medium and low fertility soil conditions. In brief, this article provides a theoretical basis and practical recommendations for the sustainable development of japonica rice in the Anhui Province of China.

Nitrogen (N) application rates and planting density are key factors affecting grain yield and quality. In this experiment, four varieties, Huijing 753 (HJ 753), Huijing 855 (HJ 855), Huijing 866 (HJ 866), and Dangjing 8 (DJ 8), were used to investigate their effects on the yield and quality of japonica rice in Anhui, which has a lower soil N supply capacity. The current study set four levels of N application N1, N2, N3, and N4 (120, 180, 240, and 300) kg ha⁻¹ and three planting densities D1, D2, and D3 (high-density: 45 × 10⁴; medium-density: 37.5 × 10⁴; low-density: 32.2 × 10⁴) hills ha⁻¹ for a total of 12 treatments. The effects of N application and planting density on rice yield, composition, processing quality, appearance quality, and cooking quality were analyzed. The results showed that with the increase in N application rates and planting density, the effective panicle number per plant and grain number per panicle of each rice variety increased, while HJ 753, HJ 855, and HJ 866 reached the highest yield under the N3D1 treatment and then gradually declined, and DJ 8 had the highest yield in the N4D1 treatment. In addition, the processing quality of various varieties can be improved by increasing the N application rate and planting density, but the appearance quality will deteriorate, the amylose content will also decrease, and gel consistency will also be lower. Therefore, under medium-planting density conditions, N at a rate of 240 kg ha⁻¹ was the best for HJ 753, HJ 855, and HJ 866, and using 300 kg ha⁻¹ for DJ 8 resulted in high grain yields and superior rice grain quality. In the next step, physiological, biochemical, and genotypic analyses of these three japonica rice varieties were carried out to provide a scientific basis and technical support for accelerating the breeding of high-yield japonica rice varieties.