Real-time detection of fruits in orchard environments is one of crucial techniques for many precision agriculture applications, including yield estimation and automatic harvesting. Due to the complex conditions, such as different growth periods and occlusion among leaves and fruits, detecting fruits in natural environments is a considerable challenge. A rapid citrus recognition method by improving the state-of-the-art You Only Look Once version 4 (YOLOv4) detector is proposed in this paper. Kinect V2 camera was used to collect RGB images of citrus trees. The Canopy algorithm and the K-Means++ algorithm were then used to automatically select the number and size of the prior frames from these RGB images. An improved YOLOv4 network structure was proposed to better detect smaller citrus under complex backgrounds. Finally, the trained network model was used for sparse training, pruning unimportant channels or network layers in the network, and fine-tuning the parameters of the pruned model to restore some of the recognition accuracy. The experimental results show that the improved YOLOv4 detector works well for detecting different growth periods of citrus in a natural environment, with an average increase in accuracy of 3.15% (from 92.89% to 96.04%). This result is superior to the original YOLOv4, YOLOv3, and Faster R-CNN. The average detection time of this model is 0.06 s per frame at 1920 × 1080 resolution. The proposed method is suitable for the rapid detection of the type and location of citrus in natural environments and can be applied to the application of citrus picking and yield evaluation in actual orchards.
In the east of China, low temperature and light energy in winter are the main factors for the decline in cucumber yield, as well as in greenhouses without supplementary light. Optimal utilization of light energy is critical to increase cucumber yield. In this study, experimental measurements were conducted in two scenarios, April to May (Apr-May) and November to December (Nov-Dec) 2015, respectively, to analyze leaf development, dry matter accumulation, and yield of cultivated cucumber. Statistical analysis showed that leaves grew in Nov-Dec had larger leaf area and lower dry matter than leaves grew in Apr-May. This revealed that the dry matter accumulation rate per unit area was lower in winter. To be precise, the yield 0.174 kg/m2 per day in Nov-Dec was 35.3% lower than the yield in Apr-May. Environmental monitoring data showed that there was no significant difference in the average temperature between two scenarios, but the light intensity in Nov-Dec was only 2/3 of that in Apr-May. Three-dimensional (3D) cucumber canopy models were used in this study to quantify the effects of weak light on dry matter production in Nov-Dec. Three 3D canopies of cucumber were reconstructed with 20, 25, and 30 leaves per plant, respectively, by using a parametric modeling method. Light interception of three canopies from 8:00 to 15:00 on 4 November 2015 was simulated by using the radiosity-graphic combined model (RGM) with an hourly time step. CO2 assimilation per unit area was calculated using the FvCB photosynthetic model. As a result, the effects of light intensity and CO2 concentration on the photosynthetic rate were considered. The leaf photosynthesis simulation result showed that during the daytime in winter, the RuBP regeneration-limited assimilation Aj was always less than the Rubisco-limited assimilation Ac. This means that the limiting factor affecting the photosynthesis rate in winter was rather light intensity. As the CO2 concentration in the greenhouse was utmost in the morning, increasing the light intensity and therefrom increasing the canopy light interception at this time will be highly beneficial to increase the yield production. Through a comparative analysis of photosynthetic characteristics in these three virtual 3D canopies, the 25-leaf canopy was the best-performed canopy structure in photosynthetic production in winter. This study provides an insight into the light deficiency for yield production in winter and a solution to make optimal use of light in the greenhouse.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.