Near-infrared (NIR) hyperspectral imaging is a useful method for powdery food quality and safety evaluation. Due to the particularity of the powder, light penetration depth and adulterant particle ...size have an effect on detection of adulterants in wheat flour. In this study, wheat flour of different thickness was placed on talcum powder or benzoyl peroxide (BPO) of different particle sizes to study the light penetration depth. BPO of different particle sizes was also mixed in wheat flour to study the adulterant particle size effects. The results showed that, for talcum, the light penetration depth in wheat flour was 0.6 mm, and for BPO, the light penetration depth in wheat flour was 0.8–1 mm (depending on the particle size of BPO). In terms of the particle size effects, mixtures of wheat flour and BPO with particle sizes of 380–270 μm and 270–250 μm were found to have better adulterant detection results. It indicated that the difference of particle size of subsurface reference powder would affect the light penetration depth. Combined with the ideal “multi-layer effect” hypothesis, it is verified that the penetration depth of light and the particle size of adulterant affect the detection of adulterant in wheat flour.
•The “multi-layer effect” can explain the results for powder adulterant discrimination.•The difference of subsurface reference powder affects the light penetration depth.•BPO particle size close to the camera spatial resolution has a better detection result.
•An algorithm (IDE) was proposed to design a new bionic subsoiling mechanism (BSM).•It is proved that IDE has better solution accuracy and speed.•Proposed a new method for establishing the soil ...model.•It was proved that the BSM has improved tillage performance based on DEM.•The driving parameters combination for the BSM was optimized.
Because of the poor subsoiling effects and high tillage resistances of small subsoiling machines when used on red soil in southern China, a new subsoiling mechanism based on functional bionics and an optimization method that employed an improved differential evolution (IDE) algorithm are proposed in this paper. The tillage performance was improved by causing the subsoiler to mimic the digging motion of a mole forefoot. The path synthesis of the subsoiling mechanism was solved by the IDE, and the tillage performance of the four-bar bionic subsoiling mechanism was evaluated by the discrete element method. First, the digging trajectory of the mole forefoot was mapped as the desired subsoiling trajectory. An optimization model for the path synthesis was established to minimize the position error between the desired subsoiling trajectory and the coupler curve of the mechanism; it was solved by the differential evolution algorithm with improvements in the mutation and crossover operations. It was found that the IDE could more rapidly search for a more accurate solution. The optimization results were used to design a bionic subsoiling mechanism that could mimic the digging trajectory of a mole forefoot. Second, the microscopic contact parameters among the soil particles were calibrated by soil accumulation and direct shear tests. In addition, a discrete element soil model was established that included dual coupling of variable contact models for the tillage layer and the plow pan. The tillage processes of an existing common subsoiling mechanism (CSM), the bionic subsoiling mechanism equipped with a common subsoiler (BSM&CS), and the bionic subsoiling mechanism equipped with a bionic subsoiler (BSM&BS) were analyzed by a joint ADAMS and EDEM simulation. The simulation results were analyzed according to established evaluation indexes for tillage performance. The comprehensive soil disturbance performance of the BSM&CS improved by 257.03% and its drag reduced by 10.87% from the CSM values, while the comprehensive soil disturbance performance of the BSM&BS improved by 312.78% and its drag reduced by 18.40% from the CSM values. Finally, the driving parameter combination for the BSM was optimized using the central composite design method.
•A contracted graph synthesis method based on similarity information is proposed.•A complete synthesis process that can be realized automatically is shown.•Fast enumeration is achieved by tracking ...similarity information.•The data of contracted graphs with up to eight independent loops are given.•Compared with other literatures, the efficiency and accuracy are verified.
Many structural synthesis methods require a complete database of contracted graphs. Therefore, this study proposes a method based on similarity information to efficiently synthesize contracted graphs for planar nonfractionated simple-jointed kinematic chains. First, the degree of all vertices is calculated using the number of independent loops and the degree sequence equation. Then, the method of vertex connection is used to enumerate the contracted graph. In this process, similarity information is used to match the connection type between vertices quickly. The type of similar vertex is determined, and nonisomorphic graphs caused by incompletely equivalent similar vertices are discovered. Additionally, the changes in similarity information are tracked to identify nonisomorphic graphs caused by similarity variation. Then, the characteristic constants representing the similarity of graphs are retained to rule out the isomorphism caused by different parent graphs. Finally, the rationality of contracted graphs is identified. The contracted graph with up to eight independent loops and its solution time are provided. Furthermore, the reliability and efficiency of the method can be demonstrated by comparing it to the database of other literature.
•Localization with the growth direction may enhance the plucking accuracy.•Planning motion sequence improves the efficiency of harvesting tea.•Field experiments validate the effectiveness of the ...proposed harvesting system.
The critical issue of the robotic harvesting high-quality tea is to realize tea shoot detection, plucking point localization, and motion planning. In addition, the accuracy and efficiency of the robotic plucking of high-quality tea in the field are essential. Therefore, a robotic harvesting system is proposed in this paper by combining deep learning, point cloud processing, and spatial path planning. First, the deep learning method and the compressed YOLOv3 network are used to quickly and accurately identify tea shoots. Second, an efficient point cloud processing-based 3D localization algorithm for high-quality tea plucking points was proposed. The genetic algorithm is then used to shorten the end-effector's motion path by optimizing the plucking sequences. Eventually, a harvester robot with a parallel manipulator was developed to conduct field plucking experiments and evaluate the effectiveness of the proposed harvesting system. All experimental results demonstrate that the success rates of detection, localization, and motion plucking are 85.16 %, 78.90 %, and 80.23 %, respectively. Furthermore, the overall process harvesting success rate is 53.91 %, and the average plucking time for a single shoot is 2.233 s. Therefore, the proposed harvesting approach can provide technical support for the precise and rapid harvesting of high-quality tea.
Display omitted
•Rice row paths are fitted by the weighted least square method.•Preview point tracking algorithm is established to solve path tracking.•Integrated navigation control strategy is ...proposed by navigation division line.•Navigation target line is set to avoid the excessive yaw angle.
To achieve autonomous navigation of the crawler tractor in the paddy field, a set of control system was developed by integrating the satellite and visual navigation. The system is mainly composed of a satellite module, a visual module and a controller. The satellite module detects the deviation information between the tractor’s position and the pre-set line AB using GPS Real–time Kinematic (RTK) technology and the navigation is achieved according to the preview point tracking algorithm. The visual module obtains the rice crop images in real time by camera, and the binary image segmentation is completed by ExG(2G-R-B), Otsu and mask method. Then the region search algorithm is used to classify the region of each rice row, and the visual navigation path is fitted by the weighted least square method, and the weight value is based on the lookahead distance. Finally, the preview point tracking algorithm is used to realize the visual navigation. Furthermore, the control division line between the satellite navigation and visual navigation is established, the threshold value of which is determined based on the crop row deviation due to the satellite navigation when rice crops were planted. The crawler tractor navigation tests were conducted to evaluate the performance of the integrated navigation system. When the speed of the tractor was 0.4∼0.6 m/s, 0.7∼0.9 m/s and 1.0∼1.2 m/s, the maximum lateral deviation of the tracking by visual navigation was 0.18 m, 0.15 m and 0.14 m, respectively, and the RMSE was 55 mm, 64 mm and 60 mm, respectively. Under the same speed, when the threshold value was set to 0.15 m, the maximum lateral offset of the tracking by integrated navigation was 0.14 m, 0.13 m and 0.12 m, respectively, and the RMSE was 41 mm, 50 mm and 55 mm, respectively. The experimental results indicate that the satellite navigation module in the integrated navigation system functions well in rice row tracking when some seedlings are missing, and significantly reduces the navigation deviation and the degree of deviation fluctuation.
•An In-situ identification method of shaking frequency for adaptive vibratory fruit harvesting was presented.•The natural frequencies of tree were identified through a frequency sweeping experiment ...using the developed intelligent shaker, and the optimal shaking frequencies were measured by an additional testing experiment.•All experimental results indicate that the identified natural frequencies can be regarded as the optimal shaking frequencies for vibratory fruit harvesting.
The improvement of the fruit removal efficiency is important in vibratory fruit harvesting. Most relevant research has focused on the detection of dynamic parameters of the tree system by conducting experiments or simulations. The present study proposes a method of identifying multiple natural frequencies of a fruit tree in-situ during vibratory harvesting. An intelligent shaker system consisting of an eccentric shaker with adjustable speed and an acceleration signal acquisition analyzer is developed. A frequency sweeping experiment using the intelligent shaker system is carried out to measure natural frequencies by searching for extreme values of the acceleration signals. Additionally, a testing experiment for measuring acceleration responses at different shaking frequencies is performed to obtain the optimal shaking frequencies. The experiments show that the natural frequencies could be obtained effectively by the frequency sweeping experiment in-situ, and the identified natural frequencies of different branches are found to be different. All results indicate that the identified natural frequencies using the proposed method can be regarded as the optimal shaking frequencies for adaptive vibratory fruit harvesting, and that it is necessary to detect the natural frequency of each large branch of a tree during vibratory fruit harvesting.
Smart materials and structures have attracted a significant amount of attention for their vibration control potential in engineering applications. Compared to the traditional active technique, shunt ...damping utilizes an external circuit across the terminals of smart structure based transducers to realize vibration control. Transducers can simultaneously serve as an actuator and a sensor. Such unique advantage offers a great potential for designing sensorless devices to be used in structural vibration control and reduction engineering. The present literature combines piezoelectric shunt damping (PSD) and electromagnetic shunt damping (EMSD), establishes a unified governing equation of PSD and EMSD, and reports the unique vibration control performance of these shunts. The schematic of shunt circuits is given and demonstrated, and some common control principles and equations of these shunts are summarized. Finally, challenges and perspective of the shunt damping technology are discussed, and suggestions made based on the knowledge and experience of the authors.
Purpose
The purpose of this paper is to study the preload range of tendon-driven manipulator and the relationship between preload and damping. The flexible joint manipulator (FJM) with joint ...flexibility is safer than traditional rigid manipulators. A FJM having an elastic tendon is called an elastic tendon-driven manipulator (ETDM) and has the advantages of being driven by a cable and having a more flexible joint. However, the elastic tendon introduces greater residual vibration, which makes the control of the manipulator more difficult. Accurate dynamic modeling is effective in solving this problem.
Design/methodology/approach
The present paper derives the relationship between the preload of the ETDM and the friction moment through the analysis of the forces of cables and pulleys. A dynamic model dominated by Coulomb damping is established.
Findings
The linear relationship between a decrease in the damping moment of the system and an increase in the ETDM preload is verified by mechanics analysis and experiment, and a curve of the relationship is obtained. This study provides a reference for the selection of ETDM preload.
Originality/value
The method to identify ETDM damping by vibration attenuation experiments is proposed, which is helpful to obtain a more accurate dynamic model of the system and to achieve accurate control and residual vibration suppression of ETDM.
•Two types of nonlinear electromagnetic shunt damping (N-EMSD) are proposed.•The nonlinear electromagnetic coupling coefficients are modelled and discussed.•The theoretical models of the two types of ...N-EMSD are established.•N-EMSD can realize nonlinear damping to improve the isolation performance.
Traditional linear electromagnetic shunt damping (L-EMSD) has been studied deeply, however, the vibration isolation performance becomes worse with the increase of linear damping in the isolation region. To overcome the limitation of L-EMSD, this paper proposes two types of nonlinear electromagnetic shunt damping (N-EMSD) to develop nonlinear damping for the broadband vibration isolation of linear vibration isolators (LVIs). Both the two types of N-EMSD consist of two permanent magnets (PMs) and coils, however, the ways to connect to the external shunt circuits and the configurations between the PMs and coils are different. Nonlinear electromagnetic coupling coefficients are analyzed theoretically. The theoretical transmissibility of the LVI with N-EMSD is derived via the harmonic balance method (HBM). Numerical simulations and experiments are conducted to verify nonlinear damping performance of N-EMSD. The results demonstrate that compared to L-EMSD the proposed N-EMSD have similar vibration suppression performance in the resonance region, but are more effective in the isolation region.
Display omitted
•An improved RGB-D multimodal tea shoot detection model based on YOLOv7 is proposed.•It created a parallel lightweight backbone network for depth images.•It designed a new cross-module spatial ...attention fusion module.•Experiments were conducted on normal and extreme environment datasets respectively to verify the advanced and robustness of the model.
Due to the increasing scarcity of tea pickers, the implementation of intelligent harvesting for premium tea is a crucial prerequisite for the sustainable development of the premium tea industry. The initial step towards achieving intelligent and precise harvesting is the accurate detection of tender shoots, which consist of one bud and one leaf. However, accurately identifying tea shoots poses a challenging visual task due to their small size, variable shapes, as well as similar colors and backgrounds. The existing model, based on RGB images, can only detect partial targets. To address this issue and further enhance the detection of tea buds, this study proposes the utilization of multi-modal features encompassing red, green, blue, and depth (RGB-D) for identification. In addition, a unidirectional complementary multi-modal fusion method is introduced to minimize the adverse effects caused by low-quality depth information. Firstly, an RGB-D dataset comprising high-quality tea leaves is constructed, and the samples are carefully calibrated. Subsequently, an enhanced end-to-end RGB-D multi-modal object detection network, referred to as YOLO-RGBDtea, is developed based on You Only Look Once version 7 (YOLOv7). This model incorporates a parallel lightweight depth image feature extraction backbone network and incorporates a self-attention mechanism to prioritize contextual information. Lastly, a cross-modal spatial attention fusion module (CSFM) is devised to collaboratively integrate depth features with RGB features in a unidirectional manner. The experimental results reveal that YOLO-RGBDtea achieves an AP50 of 91.12% when confronting complex outdoor tea shoots, exhibiting significant performance improvements compared to YOLOv7, especially in scenarios involving small targets, overlapping target groups, and highly overexposed images. Notably, the parameter increment in YOLO-RGBDtea compared to the original YOLOv7 model is merely 17.8%, and the additional components can be seamlessly transferred to other models. Overall, this study introduces a straightforward yet effective multi-modal fusion method that bears theoretical and practical significance in advancing the detection of high-quality tea shoots in complex outdoor environments.
PDF
