Search

The image segmentation technique is a vital component of vision-based fruit harvesting robots because the accuracy of this technique greatly affects the recognition and identification by robots. However, the uncertainty and ambiguity of natural scenes in orchards make image segmentation a challenging task. Here, we developed a new algorithm to detect and remove the shadows generated under intense illumination and direct sunlight conditions. Group pixels and edge probability maps were fused in our algorithm to generate superpixel blocks with precise boundaries. We applied an affinity matrix to obtain an ultrametric contour map to detect shadows and then used a relighting method to remove the detected shadows. Additionally, the shadow detection and removal and image segmentation procedures were evaluated. Our shadow detection results showed that the root mean square error decreased from 7.9% to 6.4% when an edge probability map was applied. Using the new shadow removal algorithm, the precision, balanced accuracy, specificity, and modified segmentation accuracy were improved by 10, 11, 4.5, and 10.1%, respectively. The average segmentation processing time was 0.59 s, which meets the requirements of real-time applications (

With large-scale production and the need for high-quality tomatoes to meet consumer and market standards criteria, have led to the need for an inline, accurate, reliable grading system during the post-harvest process. This study introduced a tomato grading machine vision system based on RGB images. The proposed system performed calyx and stalk scar detection at an average accuracy of 0.9515 for both defected and healthy tomatoes by histogram thresholding based on the mean g-r value of these regions of interest. Defected regions were detected by an RBF-SVM classifier using the LAB color-space pixel values. The model achieved an overall accuracy of 0.989 upon validation. Four grading categories recognition models were developed based on color and texture features. The RBF-SVM outperformed all the explored models with the highest accuracy of 0.9709 for healthy and defected category. However, the grading accuracy decreased as the number of grading categories increased. A combination of color and texture features achieved the highest accuracy in all the grading categories in image features evaluation. This proposed system can be used as an inline tomato sorting tool to ensure that quality standards are adhered to and maintained. Keywords: Grading, Calyx, Defected, Recognition models, Machine vision

Prediction of a precise subsoiling using an analytical model (AM) and Discrete Element Method (DEM) was conducted to explain cutting forces and the soil profile induced changes by a subsoiler. Although sensors, AMs and DEM exist, there are still cases of soil structure deformation during deep tillage. Therefore, this study aimed to provide a clear understanding of the deep tillage using prediction models. Experimental data obtained in the soil bin trolley with force sensors were used for verification of the models. Experiments were designed using Taguchi method. In the AM, the modified-McKyes and Willat and Willis equations were used to determine cutting forces and soil furrow profile respectively. Calculations were done using MATLAB software. The elastoplastic behavior of soil was incorporated into the DEM. The DEM predicted results with the best regression of 0.984 $$R^{2}$$ R 2 at a $$NRMSE$$ NRMSE of 1.936 while the AM had the lowest $$R^{2}$$ R 2 of 0.957, at a $$NRMSE$$ NRMSE of 6.008. All regression results were obtained at p

Proper straw cover information is one of the most important inputs for agroecosystem and environmental modeling, but the availability of accurate information remains limited. However, several remote-sensing (RS)-based studies have provided a residue cover estimation and provided spatial distribution mapping of paddy rice areas in a constant field condition. Despite this, the performance of rice crops with straw applications has received little attention. Furthermore, there are no methods currently available to quantify the wheat straw cover (WSC) percentage and its effect on rice crops in the rice-wheat cropping region on a large scale and a continuous basis. The novel approach proposed in this study demonstrates that the Landsat satellite data and seven RS-based indices, e.g., (i) normalized difference vegetation index (NDVI), (ii) Normalized difference senescent vegetation index (NDSVI), (iii) Normalized difference index 5 (NDI5), (iv) Normalized difference index 7 (NDI7), (v) Simple tillage index (STI), (vi) Normalized difference tillage index (NDTI), and (vii) Shortwave red normalized difference index (SRNDI), can be used to estimate the WSC percentage and determine the performance of rice crops over the study area in Changshu county, China. The regression model shows that the NDTI index performed better in differentiating the WSC at sampling points with a coefficient of determination (R2 = 0.80) and root mean squared difference (RMSD = 8.46%) compared to that of other indices, whereas the overall accuracy for mapping WSC was observed to be 84.61% and the kappa coefficient was &kappa, = 0.76. Moreover, the rice yield model was established by correlating between the peak NDVI values and rice grain yield collected from ground census data, with R2 = 0.85. The finding also revealed that the highest estimated yield (8439.67 kg/ha) was recorded with 68% WCS in the study region. This study confirmed that the NDVI and NDTI algorithms are very effective and robust indicators. Also, it can be strongly concluded that multispectral Landsat satellite imagery is capable of measuring the WSC percentage and successively determines the impact of different WSC percentages on rice crop yield within fields or across large regions through remote sensing (RS) and geographical information system (GIS) techniques for the long-term planning of agriculture sustainability in rice-wheat cropping systems.

A Discrete Element Method (DEM), a computerized testing rig and a soil bin were used to explain soil profile deformation caused by mouldboard plough. By using DEM, the motion of particles is given by collisions between particles. Soil-tillage tool interaction simulation was established by a 3D geometry of a mouldboard plough created using PTC Creo Parametric 3D Modelling software and soil bed modeled by discrete element particles using EDEM software. A bond element was introduced between particles and cohesion of soil was formed to create a simulation of the actual cohesive soil. The horizontal and vertical forces and soil disturbance in terms of bulk density at different speeds and depths were investigated. However, the simulation experiments followed the field experiments which were conducted using a computerized testing rig. Further, reaction forces on the plough and the resulting bulk densities were estimated from the simulated results and compared with the actual results from the field. Also, the accuracy of the developed model to predict soil displacement was measured and compared with the soil bin data with the average relative error of 1.317%. The simulated results of horizontal, vertical forces and bulk density matched well with the results obtained in the field experiment with the relative error of 4.435, 3.029 and 3.119% respectively. The vertical force provided the best regression results of 0.9948 R 2 with RMSE of 0.5964, followed by 0.9884 R 2 with RMSE of 0.4089 and 0.8391 R 2 with RMSE of 0.5441 for horizontal force and bulk density respectively. All regression results were obtained at p 0.05 . The ANOVA test showed that the p-values for the horizontal force, vertical force and bulk density were 0.753, 0.695 and 0.025 respectively. There was no significant difference between the simulation and experiment results for cutting forces while there was a significant difference for the bulk density at p 0.05 . This model can be applied as an accurate, consistent and fast method of effective prediction of the final soil condition and forces needed for tillage operation.

Plenty of straw is left in field after harvesting, so soil-tillage tool interaction has changed into straw-soil-tillage tool interaction. To obtain the detailed movement of soil and straw and identify interaction between straw, soil and tillage tool, soil bin experiments were conducted by tillage with a rotary blade. Lengths of wheat straw were 130–230 mm with increments of 20 mm. One IT225 rotary blade was operated at rotary speed of 77r/min and forward speed of 0.222 m/s, and three rotavator kinematic parameters (6.8, 8.8 and 10.8) with working depth of 100 mm. Tin alloy cubes (2 cm 3 ) and colored straw were used for measuring soil and straw movement. Moreover straw forward movement was observed greater than soil movement. The unburied percent of 150 mm straw decreased from 98% to 91% when rotavator kinematic parameters increased from 6.8 to 10.8. Further results indicated that straw mixtures might not fit for studying soil displacement, especially for lateral soil displacement; higher rotavator kinematic parameter resulted in soil and straw being spread in a larger area and more straw being buried. The average displacement of lateral and longitudinal straw tracers is recommended for forward straw displacement, and longitudinal straw tracer is used for side one.

Rice is an important cereal crop and different types of wastes are produced from it such as; rice straw as a crop residue, rice husk and rice bran as agro-industrial waste and rice waste as a food waste, which are needed to be tested for bio-hydrogen production under similar set of experimental conditions. The bio-hydrogen production potential of these wastes was compared by co-digestion with heat shocked sludge under mesophilic thermophilic temperatures by using an anaerobic bio reactor under no pH control conditions. It was observed that the bio-hydrogen production potential increased with an increase in temperature from 37 °C to 55 °C for all wastes, except for rice waste. Although, the highest experimental yield of 40.04 mL/VS rem was produced from thermophilic rice straw, but the maximum bio-hydrogen production rate of 97.08 mL/h was observed from rice bran after 29.03 h under thermophilic conditions. The bio-hydrogen production period of rice bran was 33–38% smaller than rice straw due to which the highest yield was observed from rice straw. The lignin content of rice husk was 19.34%, which was double than rice straw and hence the rice husk produced the least experimental yield of 23.05 mL/VS rem under mesophilic conditions. Increasing the temperature from mesophilic to thermophilic conditions changed the final pH by 1.79%, −8.33%, 5.17% and 5.41% for rice straw, rice bran, rice husk and rice waste, respectively. The optimum pH range of bio-hydrogen production from rice straw, bran and husk was observed from pH 7 to pH 6. The average yield of wastes produced from rice crop was 30.36 and 33.16 mL/VS rem under mesophilic and thermophilic conditions, respectively. The quadratic modeling found useful in order to explain more than 90% variation caused by incubation time and temperature on cumulative bio-hydrogen production, drop in pH and VFA production. The 3D and 2D plots developed on the basis of quadratic modeling explained the process parameter is in better way as compared to conventional representation of experimental results. As a whole, increasing temperature not only increased bio-hydrogen yield, but also decrease the volatile fatty acid production.

The effect of temperature on bio-hydrogen production by co-digestion of sewerage sludge with food waste and its two derivatives, i.e. noodle waste and rice waste, was investigated by statistical modelling. Experimental results showed that increasing temperature from mesophilic (37 °C) to thermophilic (55 °C) was an effective mean for increasing bio-hydrogen production from food waste and noodle waste, but it caused a negative impact on bio-hydrogen production from rice waste. The maximum cumulative bio-hydrogen production of 650 mL was obtained from noodle waste under thermophilic temperature condition. Most of the production was observed during the first 48 h of incubation, which continued until 72 h of incubation. The decline in pH during this interval was 4.3 and 4.4 from a starting value of 7 under mesophilic and thermophilic conditions, respectively. Most of the glucose consumption was also observed during 72 h of incubation and the maximum consumption was observed during the first 24 h, which was the same duration where the maximum pH drop occurred. The maximum hydrogen yields of 82.47 mL VS−1, 131.38 mL COD−1, and 44.90 mL glucose−1 were obtained from thermophilic food waste, thermophilic noodle waste and mesophilic rice waste, respectively. The production of volatile fatty acids increased with an increase in time and temperature in food waste and noodle waste reactors whereas they decreased with temperature in rice waste reactors. The statistical modelling returned good results with high values of coefficient of determination (R2) for each waste type and 3-D response surface plots developed by using models developed. These plots developed a better understanding regarding the impact of temperature and incubation time on bio-hydrogen production trend, glucose consumption during incubation and volatile fatty acids production.

In this study, turbulent flow fields in a baffled vessel stirred by counter-axial flow impeller have been investigated in comparison to the Rushton turbine. The resultant turbulence was numerically predicted using computational fluid dynamics (CFD). Turbulence models were developed in ANSYS Fluent 18.1 solver using the Navier-Stokes equation with the standard k-ε turbulence model. The Multiple Reference Frame (MRF) approach was used to simulate the impeller action in the vertical and horizontal planes of the stirred fluid volume. Velocity profiles generated from the simulations were used to predict and compare the performance of the two designs. To validate the CFD model, the simulation results were compared with experimental results from existing work and a satisfactory agreement was established. It was concluded that the counter-axial flow impeller could provide better turbulence characteristics that would improve the quality of mixing systems.

Design and operation of mixing systems using agitated vessels is a difficult task due to the challenge of obtaining accurate information on impeller-induced turbulence. The use of Computational Fluid Dynamics (CFD) can provide detailed understanding of such systems. In this study, experimental tests and computational fluid dynamics simulations were performed to examine the flow characteristics of four impeller designs (anchor, saw-tooth, counter-flow and Rushton turbine), in achieving solution homogeneity. The impellers were used to mix potassium sulfate granules, from which values of electrical conductivity of the solution were measured and used to estimate the distribution pattern of dissolved solid concentrations within the vessel. CFD models were developed for similar mixing arrangement using commercial software, ANSYS Fluent 18.1 solver and the standard k-epsilon (ε) turbulence model. The Multiple Reference Frame (MRF) approach was used to simulate the impeller rotation. Velocity profiles generated from the simulations were in good agreement with the experimental predictions, as well as with results from previous studies. It was concluded that, through CFD analysis, detailed information can be obtained for optimal design of mixing apparatus. These findings are relevant in choosing the best mixing equipment and provides a basis for scaling up mixing operations in larger systems.

Soil management practices are used to enhance soil organic carbon, fertility, and crop productivity around the world. However, accurate information about the appropriate amount of straw incorporation is not available, because it is generally believed that at least 30% of the soil surface should be covered by straw, which is not implemented in all field environments. Therefore, a two-year (2016–2017) field experiment was conducted to investigate the impact of different percentages of straw incorporation and tillage methods, i.e., reduced tillage (RT) and conventional tillage (CT), on crop yield, soil organic carbon (SOC), total nitrogen (TN), and soil carbon storage (SCS) in rice–wheat cropping systems, under eight treatments. The experimental results showed that the greatest reduction in soil dry bulk density ( ρ b ) was found under CT with 100% straw coverage (9.79%), whereas the least reduction occurred under CT with no straw (1.31%). The mean TN concentration, soil organic matter (SOM), and soil carbon storage (SCS) were significantly higher by 0.98 g/kg, 17.07%, and 14.20%, respectively, under reduced tillage with 60% straw incorporation (RTsi60) compared with all other treatments. Our findings demonstrate that the incorporated wheat residues resulted in the highest rice production (7.95–8.63 t/ha) under RTsi60. We recommend the adoption of reduced tillage with 60% straw incorporation to increase rice yield, improve soil structure, and enhance TN, SOM, and SCS in paddy soil under rice-wheat rotation fields for agricultural sustainability.

Spatial distribution of soil forces on the surface of plough is an important aspect that can help engineers for improving efficiency of tillage implement. It was analyzed at eleven different points of the moldboard plough with the help of sensors accompanied with the virtual instrument developed in LabView software with the aid of other supporting instruments. It was observed that soil forces increased with an increase in speed and depth. Depth changed soil forces more at upper parts than lower parts whereas speed affected rear parts more than the front part of the plough. Draft forces followed almost similar trend and least value of 308.17 N experimental draft force was found at 1 m/s speed and 5 cm depth under 33% moisture content. Cumulative soil forces found too smaller than the draft as they represented the force spatial distribution of specific parts of plough. It was observed that sensor technology provided real time picture of force variation during tillage process that could save time and effort.

Soil failure patterns play an important role in obtaining a better understanding of the mechanical behavior of soils. Despite the large number of studies over the past few decades, a better understanding of soil failure patterns and its relation to soil and tool parameters for particular soils such as dry land and paddy has not been developed. This study investigated soil failure patterns and related draft at sticky, plastic and liquid consistency limits and the sticky point of dry land and paddy soils. A soil cutting test rig was developed to perform soil cutting at three consistency limits (sticky limit, plastic limit and liquid limit) and the sticky point of soil, three rake angles (15°, 30° and 45°), and three operating depths (30 mm, 50 mm and 70 mm). A flat triangular shaped tool operating at a constant speed of 10 mm s−1 was used in all experiments. Soil failure patterns were observed and recorded using a digital camera, and draft per unit displacement was measured by load cells attached to the soil bin. A direct relationship between soil failure patterns or draft and the consistency limits of soil was found. Brittle failure was obtained at the sticky limit, chip forming failure was observed at 15° rake angle and 30 mm depth, and bending failure with little strains of elements at 30° and 45° rake angles and 50 mm and 70 mm depths at plastic limit, while flow failure was linked to the liquid limit of the soil. At the sticky point, flow failure was observed at an operating depth of 30 mm and 15° rake angle, while flow with considerable bending and no strains of elements occurred at 50 mm and 70 mm operating depths and 30° and 45° rake angles. However, bending was more prominent at 70 mm depth and 45° rake angle. The draft at the sticky limit, plastic limit and sticky point was cyclic in nature, whereas at the liquid limit it was comparatively diverse and fading. The highest draft was found at the plastic limit, and the lowest at the liquid limit. Since the soil failure patterns may change with moisture content, soil type and particle size distribution within the same textural class, consistency limits can provide clearer and more accurate definitions of soil failure patterns than moisture content levels alone.

In this study Burger’s four element model was used to analyze rheological parameter of paddy soil of t he Southern China. Two types of tool shape include cir cular shape (with an area of 12.56 cm 2 ) and rectangular shape (with an area of 12.56 cm 2 ) were use to measure the rheological parameter of paddy soil under different loading pressure (7.65, 10.20 and 12.75 k g) and soil water content (27 and 30%), the test wa s carried out an indoor soil bin. The results show th at the Burger model possesses an excellent prospect ive for proper representation of the time dependent behavio r of a stress-strain time graph curve can well imit ate the tool structural change during the soil deformation under different soil water content. Overall results show that, the soil water content has great impact on soil def ormation under different loading pressure, where pr oper tool shapes can control the soil deformation under diffe rent pressure. Soil water content on rheological pa rameter (Em, Ek, λm and λk) were predominantly significant, while loading on rheological parameter (Em, Ek, λm and λk) were not significant. The difference in the mean values among the different levels of M.C and Load has significant difference (p =

Three common pretreatments (mechanical, steam explosion and chemical) used to enhance the biodegradability of rice straw were compared on the basis of bio-hydrogen production potential while co-digesting rice straw with sludge under mesophilic (37 °C) and thermophilic (55 °C) temperatures. The results showed that the solid state NaOH pretreatment returned the highest experimental reduction of LCH (lignin, cellulose and hemi-cellulose) content and bio-hydrogen production from rice straw. The increase in incubation temperature from 37 °C to 55 °C increased the bio-hydrogen yield, and the highest experimental yield of 60.6 mL/g VSremoved was obtained under chemical pretreatment at 55 °C. The time required for maximum bio-hydrogen production was found on the basis of kinetic parameters as 36 h–47 h of incubation, which can be used as a hydraulic retention time for continuous bio-hydrogen production from rice straw. The optimum pH range of bio-hydrogen production was observed to be 6.7 ± 0.1–5.8 ± 0.1 and 7.1 ± 0.1–5.8 ± 0.1 under mesophilic and thermophilic conditions, respectively. The increase in temperature was found useful for controlling the volatile fatty acids (VFA) under mechanical and steam explosion pretreatments. The comparison of pretreatment methods under the same set of experimental conditions in the present study provided a baseline for future research in order to select an appropriate pretreatment method.

This study investigates the comparison of mechanical properties of wheat and rice straw such as shear strength, specific shearing energy and cutting forces. The experiments were conducted at three loading rate of 15, 20 and 25 mm min-1 and three internode positions 70 (N1), 130 (N2) and 190 (N3) mm down from the ear. Results show that by increasing the loading rate, strength of wheat and rice straw changed from 8.12 to 22.94 and 6.06 to 14.33 MPa and specific shear energy was varied from 12.10 to 18.64 and 10.40 to 16.17 mJ mm-2, respectively. Moreover, the values of cutting forces of wheat and rice straw were within the ranges 13.23 to 19.50 and 9.40 to 16.70 N. Whereas the shear strength, specific shearing energy and cutting force were higher at higher loading rate at the third internode of both straw internode positions. The shear strength, specific shearing energy and cutting force of rice straw were significantly higher (p

A laboratory rig was developed for a scale- and energy-controlled tender soil fragmentation test, which required multiple impacts and step-by-step decomposition of the soil structure on a macroscale. Soil sample size classes were scaled down by a factor of two, giving ranges of 128, 64, 32, 16, 8, 4, 2, 1, and 0.5 mm, respectively. These size classes were also used as control scales for respective fragmentation stages. Fragmentation tests were conducted to illustrate the possible influences from shape effects, falling weights, displacements, and fragmentation stages, and to investigate the relations among specific fragmentation energy, number of strokes per fragmentation stage, and fragmentation fractal dimension. Specific fragmentation energy was influenced by both the size and the shape of the soil samples. The cylindrical- and ball-shaped soil samples behaved similarly. Thus ball samples were used for a benchmarking test on soil structural stabilities. Subsequent fragmentation on 60- and 80-g ball samples with changed displacements from 5 to 45 cm yielded a decreased number of strokes from around 150 to 10. Whereas, their respective fragmentation fractal dimensions varied insignificantly, ranging from 1.52 to 1.64 for samples of 60-g mass and 1.55 to 1.72 for 80 g, values much smaller than previously reported. Fragmentation fractal dimension appeared to be a nominal index for remolded soils under laboratory impacting actions. There was a slight increase in specific fragmentation energy as the number of impacts increased. But this increase was aggravated for a single or only few impacts. Thus 20 to 40 strokes was the appropriate stroke number to finish one fragmentation stage. A five-stage fragmentation on a sample in the 32- to 64-mm size range yielded fragmentation fractal dimensions of 1.29, 1.52, 1.45 and 1.61, for the first four stages, respectively, whereas their specific fragmentation energies increased linearly.

Food waste along with its two individual components, noodle waste and rice waste, were tested for bio-hydrogen production by using sludge as a source of mix consortia of Clostridium under different physical conditions (pH 5, 6, and 7; temperature 37 °C and 55 °C). The increase in pH increased the bio-hydrogen yield for all tested wastes, whereas an increase in temperature increased the bio-hydrogen yield just for food waste. The highest experimental yield of 115.76 ml/VSremoved was produced in the mesophilic noodle waste reactor at pH 7. The drop in pH from 7 to 4.8 ± 0.2 was found optimum for bio-hydrogen production for all tested wastes under mesophilic as well as thermophilic conditions. Most of the hydrogen production was observed within 72 h of incubation, which can be used as the optimum bio-hydrogen production period for food waste. The bio-hydrogen yield, final volatile fatty acids (VFA), and glucose consumption at 72 h were analyzed with the help of the response surface methodology. The resultant plots represented an increase in glucose consumption with the increase in pH from 5 till pH 6 ± 0.5, after which glucose consumption started to decrease up to pH 7. The final VFA represented a similar trend as that observed for glucose except that the change in VFA production was observed due to the temperature and transition was observed at 47.5 ± 1.5 °C for food waste as well as for noodle waste.

A route detecting algorithm was proposed for the cucumber picking robot navigation in greenhouse environment. Possible navigation route was determined from the additive value of the column scanning results. Color difference between cucumber plants and the mid row soil was analyzed through combining the RGB elements segregations with extra-green (ExG) value and extra-red (ExR) value. The obtained gray image was segmented with OTSU threshold method, which made the gray-color abruptly changed pixels along the two sides of the central line identified. Mean value of these selected pixels in each line was calculated to provide the discrete points along the navigation route. Finally least square method was used to fit these discrete points to provide the navigation lines, from which, the route for current navigation scheme can be determined. The proposed algorithm was applied to multiple images for its high speed and anti-noise verification.© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

Part 1: Decision Support Systems, Intelligent Systems and Artificial Intelligence Applications; International audience; Object recognition is one of the hottest issues in the field of vision system for harvesting robot. How efficiently and accurately to remove the background and get the object in image is the key research. The attention mechanisms of human visual system (HVS) can be segmented an image into the region of interesting (ROI) which is considered important and the background which is less important, and recognized the object from ROI using the local information. In this paper, an algorithm based on the characteristic of HVS is proposed. In algorithm, the image was partitioned into many blocks of equal size. ROI was got through calculating the factor of weight of each sub-block image, and the object was extracted by segmenting the ROI. Experiment results show that the algorithm can be recognized the object efficiently and accurately. A new method for vision system of harvesting robot is provided.

The Internet of things (IOT) is a new integrated system from various information techniques, till now, it is becoming a hot spot of present research and application. Furthermore, IOT has a good prospect of application and extension in the agricultural engineering field. This paper introduces common and key technologies, and proposes several applications of smart farming based on IOT— the plant factory automation systems, monitoring and control system of agricultural pest, supervision and management system of quality and safety for agricultural products.

The pulse-coupled neural network (PCNN) algorithm is an efficient method widely used in image segmentation. Parameters adjusting is usually difficult in a classic model of PCNN. In this study the pulse-coupled neural network model was simplified for optimal segmentation by reducing the number of parameters of PCNN. In addition, the local standard deviation was utilized for adjusting the connection strength coefficient adaptively. The simplified PCNN was used for separating the cucumber from complex background in a cucumber image effectively. To evaluate the performance of this algorithm, a simple evaluation method was designed for evaluating the segmentation image. The experimental results show that the average rate of correct segmentation reaches up to 82.4%.