Tomato detection method using domain adaptive learning for dense planting environments

This study aimed to address the challenge of accurately and reliably detecting tomatoes in dense planting environments,a critical prerequisite for the automation implementation of robotic harvesting.However,the heavy reliance on extensive manually annotated datasets for training deep learning models still poses significant limitations to their application in real-world agricultural production environments.To overcome these limitations,we employed domain adaptive learning approach combined with the YOLOv5 model to develop a novel tomato detection model called as TDA-YOLO(tomato detection domain adaptation).We designated the normal illumination scenes in dense planting environments as the source domain and utilized various other illumination scenes as the target domain.To construct bridge mechanism between source and target domains,neural preset for color style transfer is introduced to generate a pseudo-dataset,which served to deal with domain discrepancy.Furthermore,this study combines the semi-supervised learning method to enable the model to extract domain-invariant features more fully,and uses knowledge distillation to improve the model's ability to adapt to the target domain.Additionally,for purpose of promoting inference speed and low computational demand,the lightweight FasterNet network was integrated into the YOLOv5's C3 module,creating a modified C3_Faster module.The experimental results demonstrated that the proposed TDA-YOLO model significantly outperformed original YOLOv5s model,achieving a mAP(mean average precision)of 96.80%for tomato detection across diverse scenarios in dense planting environments,increasing by 7.19 percentage points;Compared with the latest YOLOv8 and YOLOv9,it is also 2.17 and 1.19 percentage points higher,respectively.The model's average detection time per image was an impressive 15 milliseconds,with a FLOPs(floating point operations per second)count of 13.8 G.After acceleration processing,the detection accuracy of the TDA-YOLO model on the Jetson Xavier NX development board is 90.95%,the mAP value is 91.35%,and the detection time of each image is 21 ms,which can still meet the requirements of real-time detection of tomatoes in dense planting environment.The experimental results show that the proposed TDA-YOLO model can accurately and quickly detect tomatoes in dense planting environment,and at the same time avoid the use of a large number of annotated data,which provides technical support for the development of automatic harvesting systems for tomatoes and other fruits.

Reduced nitrogen application rate with dense planting improves rice grain yield and nitrogen use efficiency: A case study in east China

Dense planting could be a feasible method for reducing nitrogen(N) application rates without compromising rice grain yield in northeast and central China. It is still unclear whether reduced N application with dense planting(RNDP) can achieve higher rice yield and N use efficiency(NUE) in Jiangsu, east China. Three japonica inbred rice(JI) and three indica hybrid rice(IH) cultivars were grown in a field experiment. Their grain yield, NUE, and related traits were compared under two cultivation treatments:conventional high-yielding practice(CHYP) and RNDP. JI showed similar yields under the two treatments,while IH showed lower yield under RNDP than under CHYP, and the partial factor productivity of N and N use efficiency for grain yield increased(P < 0.05) in both JI and IH under RNDP. Compared with CHYP,RNDP reduced spikelets per panicle but increased panicles per m2 and filled-kernel percentage of JI and IH, and JI's kernel weight was increased(P < 0.05) under RNDP. Shoot biomass weight and nonstructural carbohydrate(NSC) content in the stem at heading and maturity of JI and IH were reduced under RNDP, while harvest index and NSC remobilization reserve were increased(P < 0.05) under RNDP, especially for JI. Our results suggest that RNDP could achieve a higher rice grain yield and NUE, particularly for JI, a dominant rice cultivar type in Jiangsu. For JI, the increased panicles per m2, sink-filling efficiency, harvest index, and NSC remobilization after heading under RNDP contributed to a grain yield similar to that under CHYP.

Reducing environmental risk of nitrogen by popularizing mechanically dense transplanting for rice production in China

The high nitrogen(N)application rates typically used in Chinese cropping systems have led to diminishing returns for yields and have also imposed substantial environmental costs.Here,we estimate that the annual N loss from rice production in China reached approximately 2.6×109 kg from 2011 to 2015,and we demonstrate that adoption of the mechanically dense transplanting technique by producers is an effective method to reduce N loss from rice cropping systems without suffering a yield penalty.

How plant density affects maize spike differentiation, kernel set, and grain yield formation in Northeast China?

A two-year field experiment was conducted to evaluate the effects of plant density on tassel and ear differentiation, anthesissilking interval(ASI), and grain yield formation of two types of modern maize hybrids(Zhongdan 909(ZD909) as tolerant hybrid to crowding stress, Jidan 209(JD209) and Neidan 4(ND4) as intolerant hybrids to crowding stress) in Northeast China. Plant densities of 4.50×104(D1), 6.75×104(D2), 9.00×104(D3), 11.25×104(D4), and 13.50×104(D5) plants ha-1had no significant effects on initial time of tassel and ear differentiation of maize. Instead, higher plant density delayed the tassel and ear development during floret differentiation and sexual organ formation stage, subsequently resulting in ASI increments at the rate of 1.2–2.9 days on average for ZD909 in 2013–2014, 0.7–4.2 days for JD209 in 2013, and 0.5–3.7 days for ND4 in 2014, respectively, under the treatments of D2, D3, D4, and D5 compared to that under the D1 treatment. Total florets, silking florets, and silking rates of ear showed slightly decrease trends with the plant density increasing, whereas the normal kernels seriously decreased at the rate of 11.0–44.9% on average for ZD909 in 2013–2014, 2.0–32.6% for JD209 in 2013, and 9.7–28.3% for ND4 in 2014 with the plant density increased compared to that under the D1 treatment due to increased florets abortive rates. It was also observed that 100-kernel weight of ZD909 showed less decrease trend compared that of JD209 and ND4 along with the plant densities increase. As a consequence, ZD909 gained its highest grain yield by 13.7 t ha-1on average at the plant density of 9.00×104 plants ha-1, whereas JD209 and ND4 reached their highest grain yields by 11.7 and 10.2 t ha-1at the plant density of 6.75×104 plants ha-1, respectively. Our experiment demonstrated that hybrids with lower ASI, higher kernel number potential per ear, and relative constant 100-kernel weight(e.g., ZD909) could achieve higher yield under dense planting in high latitude area(e.g., Northeast China).