The orchards usually have rough terrain, dense tree canopy and weeds. It is hard to use GNSS for autonomous navigation in orchard due to signal occlusion, multipath effect, and radio frequency interference. To achieve autonomous navigation in orchard, a visual navigation method based on multiple images at different shooting angles is proposed in this paper. A dynamic image capturing device is designed for camera installation and multiple images can be shot at different angles. Firstly, the obtained orchard images are classified into sky and soil detection stage. Each image is transformed to HSV space and initially segmented into sky, canopy and soil regions by median filtering and morphological processing. Secondly, the sky and soil regions are extracted by the maximum connected region algorithm, and the region edges are detected and filtered by the Canny operator. Thirdly, the navigation line in the current frame is extracted by fitting the region coordinate points. Then the dynamic weighted filtering algorithm is used to extract the navigation line for the soil and sky detection stage, respectively, and the navigation line for the sky detection stage is mirrored to the soil region. Finally, the Kalman filter algorithm is used to fuse and extract the final navigation path. The test results on 200 images show that the accuracy of visual navigation path fitting is 95.5%, and single frame image processing costs 60 ms, which meets the real-time and robustness requirements of navigation. The visual navigation experiments in Camellia oleifera orchard show that when the driving speed is 0.6 m/s, the maximum tracking offset of visual navigation in weed-free and weedy environments is 0.14 m and 0.24 m, respectively, and the RMSE is 30 mm and 55 mm, respectively.

Apple orchard production management is faced with the double challenge of rising labor costs and population aging. In order to improve operational efficiency, reduce production costs and improve fruit quality, the application of mechanization and intelligent technology in apple orchard production management is becoming more and more important. This paper summarizes the research progress of intelligent mechanization technology in apple orchard production and management under the conditions of wide rows and dense planting and inter-logging and stemming cultivation mode; and the research progress of mechanization technology and equipment and intelligent technology and equipment in key links such as orchard mowing, herbicide application, flower and fruit management, fruit harvest and transportation, and fruit tree branch processing is analyzed. it is found that although the mechanization degree of irrigation and fertilization is relatively high at this stage, pruning, flower thinning, fruit thinning, bagging, picking and other links still rely on manual labor and have a low degree of mechanization; and the application of intelligent technologies such as drones and the Internet of Things in orchard management is not yet widespread, and precision agriculture and automated control need to be strengthened; the existing technologies develop independently and lack effective integration, which affects the improvement of production efficiency and fruit quality; the technology integration is not high and the data collection and analysis capabilities are insufficient. For the current stage of intelligent orchard construction technology model, we need to research and develop orchard machinery and equipment adapted to different terrain conditions, and create a demonstration base for the whole process of mechanization; focus on research and development of mechanized intelligent technology for key production links, and promote easy-to-manipulate intelligent machinery; use multi-source information sensing equipment to realize the digital perception of orchard environmental information; carry out intelligent decision-making and precise operation to improve operational efficiency and quality, and realize mission tasks through the intelligent orchard operation and maintenance system. Intelligent orchard operation and maintenance system realizes the demand for task allocation and scheduling, operation control and monitoring, and provides technical support for building typical application scenarios of mechanization of the whole apple orchard production.

Aiming at the problems of difficult weeding operation, low weeding efficiency and low level of intelligence in orchards in hilly and mountainous areas, a small mountain orchard weeding robot was designed. In order to improve the tracking accuracy of the tracked weeding robot's operation path, the research of tracked weeding robot's path tracking control algorithm was carried out. The ‘inverted triangle swivel structure’ was used to design an imitation floating chassis, which carried a ‘Y-type’ blade assembly for crushing and weeding and powered by an extended-range hybrid power system, and a slope steering control strategy based on model predictive control was proposed to address the problem of large slope steering slippage. The round-trip path planning method was used to plan the full coverage path of the orchard. Combined with BDS positioning and navigation technology and full coverage path planning, the crawler weeding robot was ensured to have high tracking accuracy during operation. The kinematic model of the crawler weeding robot was constructed, and the steering dynamics and control strategy of the crawler weeding robot on the slope were simulated and analyzed in MATLAB software. The simulation results show that the average error of the designed slope steering model is only 0.039 m under the condition of 15°, which demonstrates a good accuracy. Field experiments show that the MPC controller proposed in this paper can effectively improve the path tracking effect under slope steering conditions, and the average error under uphill and downhill conditions is reduced by 51.76%and 63.77%, respectively, compared with the PID controller when the slope angle is 15°. The weeding robot integrated with BeiDou navigation function has an effective weeding rate higher than 97% and can walk normally on a 25° slope. The research results provide a reference for the development of weeding robots in hilly and mountainous areas.

Cable-driven parallel robots have become a hot research topic in the field of robotics in recent years due to the characteristics of simple structure, small motion inertia, reconfigurability, and fast response speed, etc. The cable-driven flexible manipulator provides a certain adaptability and softness for operation due to the flexibility and elasticity of driving cable, and consequently it can achieve good dynamic interaction between the operating objects and the manipulators, and improve the ability to prevent broken loss coupling, which has also attracted widespread attention from researchers in the field of agricultural robotics. Due to the working environment of agricultural picking robots has the characteristics of unstructured and uncertainty, and most of the fruit outer skin is fragile and easy to damage, the requirements for picking robots that come into directly contacted with the fruits are stricter, ensuring stable grasping without damaging the fruits. In order to achieve efficient and precise operations of picking robots, lightweight and flexible end execution agencies that interact with biological friendly interaction are key theoretical technologies that urgently need to be broken through. This article first elaborates on the characteristics and development overview of cable-driven flexible manipulators with commercial application prospects. Then, the research progress of the theory such as the design, modeling and control of cable-driven flexible robotics at home and abroad is reviewed in detail. Further the typical application development of cable-driven flexible robots in the field of physical rehabilitation and dexterous grabbing is outlined. A detailed analysis is conducted on the significance of the application of cable-driven flexible manipulators in the field of agricultural picking, and the current research status of cable-driven flexible manipulators in tomato picking, apple picking, strawberry picking, and blackberry picking. Aiming at the problems of low picking accuracy, poor versatility, high cost, damage and low picking efficiency of cable-driven flexible manipulator in agricultural picking at the present stage, it is pointed out that in the future, the commercialization of cable-driven flexible manipulator needs to make innovations in agricultural machinery and agronomy collaboration, modular and reconfigurable design, improving interaction safety, multi-sensor integration, picking sequence planning and strong robust control in this study and it provides new ideas and methods for the design of mechanized non-destructive picking mechanisms for fruits and vegetables.

Laser, as an emerging artificial light source, is one of the greatest inventions of the 20th century, featuring high power density, excellent directionality, and superb monochromaticity. It has been widely applied in the agricultural sector. This article provides an in-depth analysis of the research progress and development trends of laser technology in agriculture. In plant production, it discusses applications in plant mutation breeding, promoting plant growth and development, enhancing yield and quality, plant protection, plant detection, and phenotyping. In animal management, the focus is on the use of laser technology in animal genetic breeding, growth, medical care, and product testing. For fungal research, it summarizes the use of laser technology in fungal breeding, growth, detection, and identification. The article also addresses the challenges and difficulties faced when integrating laser technology with agricultural production, proposing directions for future development, such as improving precision in plant production, optimizing animal management, and advancing scientific fungal production. Although laser technology has already achieved significant results in agriculture, future research should explore more innovative applications and integrate with artificial intelligence and big data technologies to further advance the deep integration of laser technology with the agricultural industry, leading to more efficient and intelligent agricultural production management and offering new opportunities and breakthroughs for modern agriculture.

Nitrogen is one of the important elements for the growth and development of rice, and accurate estimation of nitrogen concentration is crucial for guiding precise fertilization and assisting in the selection of nitrogen efficient varieties in rice. Traditional field sampling methods make it difficult to obtain real-time nitrogen concentration in rice. With the rapid development of information technology, establishing the relationship between unmanned aerial vehicle hyperspectral data and nitrogen concentration through machine learning methods is currently one of the main technical routes for crop nitrogen nutrition diagnosis. This study constructs an inversion model by using the feature bands of unmanned aerial vehicle canopy hyperspectral data selected by the continuous projection algorithm as input and the measured nitrogen concentration data as output. Extreme learning machine (ELM) has the advantages of fast speed and strong generalization ability compared to similar machine learning methods. However, due to its randomly generated connection weights and neuron thresholds, its training stability is insufficient and it is prone to falling into local optima. The beluga whale optimization (BWO) is a competitive algorithm inspired the behavior of beluga whales to solve single modal and multimodal optimization problems. In this study, the BWO-ELM rice nitrogen concentration unmanned aerial vehicle hyperspectral inversion model was constructed to achieve rapid estimation of rice nitrogen concentration by optimizing the connection weight between the input layer and the hidden layer of the ELM, as well as the initial weight of the hidden layer through the BWO. The research results show that the continuous projection algorithm filters out 10 feature bands, which are 673, 703, 727, 823, 850, 877, 895, 952, 961, and 985 nm, respectively. The training set R² and RMSE of the nitrogen concentration inversion model constructed based on BWO-ELM are 0.742 5 and 0.382 6%, respectively, while the testing set R² and RMSE are 0.702 8 and 0.487 7%, respectively. The predictive ability is superior to that of the nitrogen concentration inversion model constructed based on ELM. In summary, the rice nitrogen concentration unmanned aerial vehicle hyperspectral inversion model based on BWO-ELM can quickly and accurately obtain rice nitrogen concentration, providing a new method for rice nutrition monitoring.

Rice seedlings need sufficient light for normal growth. The increasing temperature demand of early spring seedling has promoted the development of greenhouses, greenhouses and factory facilities. However, due to the reflection and absorption of sunlight by glass and plastic, the lack of light in facility seedlings is prominent. Supplementary lighting can better solve the problem of lack of light in facility seedling, LED as the representative of the supplementary light source has been promoted and applied in facility seedling, but it still has high energy consumption problem. Laser is the only artificial light source with parallel light characteristics of sunlight, with good correlation, good monochromism, good directivity, high brightness, large energy, photoelectric conversion efficiency, energy saving and so on. Studies have shown that the special light quality of laser can effectively improve the photosynthesis efficiency of seedlings. The traditional laser light source represented by He-Ne laser is large in volume, high in cost, and difficult to be popularized. Using semiconductor laser combined with uniform light technology to develop a new laser light source, small size, low cost, can achieve a large area of uniform laser irradiation of rice seedlings, the single lamp irradiation area can reach 60-70 square meters. At the same time, the new laser light source retains the traditional laser correlation, monochromism and directivity, and also has the characteristics of high efficiency and energy saving, and the energy consumption can be only 1/30 of the traditional LED light source. The results showed that reasonable irradiation of new laser light source in the process of rice seedling cultivation can improve the quality of seedlings, fast greening after transplanting, more tillers, early heading, and finally achieve stable and increased rice yield. Since 2021, the test results in many places across the country have shown that the use of a new laser light source to irradiate rice seedlings for about 20 days before transplanting can achieve a yield increase of more than 10%. This result not only provided important technical support and guarantee for improving the quality and yield of rice in China, but also provided a new method and means for other plants to fill light at various stages.

The realization of automatic identification and positioning of tea buds is the basis for the development of high-quality tea intelligent picking equipment. Aiming at the problems of tiny tea buds and the picking environment which is greatly affected by light, this research proposes a tea bud recognition method based on a deep learning network model to carry out the light source design of the recognition system, which can provide technical support for the realization of all-weather and high-efficiency intelligent tea bud picking equipment. First, an aluminum alloy frame was constructed to provide a closed and shaded dark environment; then, three combinations of heights and three combinations of light intensities were created by adjusting the height of the crossbar and the brightness of the light source; finally, the image datasets of tea buds in different combinations were collected, and recognition tests were carried out on one bud and one leaf and one bud and two leaves by using the improved YOLOv5 model. The experimental results show that the overall accuracy of YOLOv5s is 77.13%, and the overall average precision mean is 86.14%, the overall accuracy of the improved recognition model YOLOv5s-SPD is 80.30%, and the overall average precision mean is 87.3%, and the average detection time of a single image is 5.7 ms, which meets the requirement of real-time detection, and is better than the original YOLOv5s with an overall accuracy improvement of 3.17% and an overall average precision mean of 1.16%, which effectively improves the recognition performance of tea buds. Under the condition of height 90 cm and luminance L7(0.164~0.328 μmol/m²), the detection accuracy, recall and AP average of one bud and two leaves were 86.70%, 92.45% and 95.00%, respectively. The method proposed in this paper can effectively and quickly detect tea buds, and the light source design scheme supports the development of all-weather high-quality tea intelligent picking equipment.

In recent years, with the continuous progress of semiconductor laser technology, laser technology applied in the field of agriculture has attracted much attention. The laser light source has the advantages of strong monochromaticity, high electro-optical conversion efficiency and low use cost, which provides new technical support for improving crop quality and increasing yield. In order to study the effects of different light quality laser on the growth, physiological and biochemical indexes, polysaccharide and saponin accumulation of ginseng, biennial ginseng carrier seeds were selected for experiments. From the leaf opening stage to the leaf withering stage: from 7：00 to 11：00, and from 14：00 to 18：00. every day, full-wavelength LED lights, red laser light (R3), blue laser light (B3), hybrid laser light R2B1, R3B2 (optical quantum density of 4.48 μmol/(m²·s), 4.61 μmol/(m²·s)) are used to fill light, respectively. According to the influence of different light quality, this study set up no light as a blank control group. It was found that the net photosynthetic capacity and stomatal conductance of leaves of ginseng were improved. Among them, the photosynthesis capacity of the LED light source was more obvious, and the absorption capacity of the blue light source was also stronger. In particular, the net photosynthetic rate of the LED treatment group was the most prominent, while the blue light treatment group achieved the most significant improvement in transpiration rate and intercellular CO₂ concentration. The stomatal conductance of R3B2 treatment group was the highest. In terms of growth, ginseng performed best under mixed red and blue light treatment. Compared with the control group, blue light could significantly increase plant height and leaf length and width. The root weight and leaf weight of LED full wavelength complement light group were higher than other treatments. In terms of the accumulation of nutrients, the polysaccharide content in blue light treatment was the highest, and the saponin content in red light treatment was generally higher than that in blue light treatment group, indicating that red light promoted the accumulation of saponin content. The content of Rg1 and Rb1 saponin in R2B1 treatment group was the highest, which was much higher than that in other treatments; the content of Re saponin in R3B2 treatment was similar to that in red light treatment group, and higher than that in the other four treatments. In future studies, we will select the most suitable conditions for the growth and nutrient accumulation of ginseng, so as to provide a feasible method for shortening the growth cycle, improving the quality and efficiency of ginseng.

As agricultural production continues to increase its requirements for field management, traditional agricultural machinery and equipment have gradually become difficult to meet the production needs of modern smart agriculture. In this context, soil moisture monitoring technology, as a key means of obtaining soil moisture information in modern agricultural management, is playing an important role in promoting the development of agricultural machinery towards intelligence and automation. Therefore, this study thoroughly reviews and analyzes the current research status of soil moisture monitoring technology at home and abroad, focusing on the research progress of three aspects of soil moisture monitoring methods and principles, model construction algorithms, and signal processing methods. Through comparative analysis, the differences and similarities in monitoring methods, principles, model construction algorithms, and signal processing methods at home and abroad, as well as the problems and challenges in practical applications, are summarized. The future development trends of soil moisture monitoring technology in these three aspects are proposed: in terms of soil moisture monitoring methods, a multi-source soil moisture monitoring information platform is constructed to achieve more comprehensive data collection and analysis; In terms of model construction algorithms, machine learning and deep learning algorithms are adopted to customize model algorithm modules for different soil environments and working scenarios, improving the accuracy and applicability of monitoring equipment; In terms of signal processing, the application of multi-source signal fusion technology is strengthened to reduce the impact of the working environment on monitoring equipment.