Ernest W. Tollner

APPLE CLASSIFICATION BASED ON SURFACE BRUISES USING IMAGE PROCESSING AND NEURAL NETWORKS

Maintaining prime fruit quality is the key to success in the fresh fruit business. Quality defects such as bruises in apples adversely affect their market value. Line–scan x–ray imaging has shown potential for detecting these quality defects. Quality assessment of apples with computer vision techniques is possible; however, two basic issues must be addressed before an automatic sorting system can be developed: (1) which image features best correlate with the fruit quality, and (2) which classifier should be used for optimal classification. These issues are discussed in this article. Red delicious (RD) and golden delicious (GD) apples were line–scanned for bruise damage. Spatial and transform features were evaluated for their discriminating contributions to fruit classification based on bruise defects. Stepwise discriminant analysis was used for selecting the salient features. Spatial edge features detected using Robert’s edge detector, combined with the selected discrete cosine transform (DCT) coefficients proved to be good indicators of old (one month) bruises. Separate artificial neural network (ANN) classifiers were developed for old (one month) and new (24 hour) bruises. When an ANN classifier was used to sort apples based on old bruises, it achieved an accuracy of 90% for RD apples and 83% (93% after threshold adjustment) for GD apples. For new bruises, the accuracy was approximately 60% for both RD and GD apples. New bruises were not adequately separated using this methodology.

RELATING X-RAY ABSORPTION TO DENSITY AND WATER CONTENT IN APPLES

X-ray computed tomography (CT) was used to image interior regions of ‘Red Delicious’ apples under varying moisture and, to a limited extent, density states. Images were actually maps of x-ray absorption of fruit cross-sections. X-ray absorption properties of ‘Red Delicious’ apples were evaluated using normal apples alternately scanned and sequentially freeze dried, fruit affected by watercore disorder, and normal apples freeze-dried to varying levels. The studies were designed to allow quantification of the x-ray absorption coefficient associated with the dry solids portion of the fruit and the x-ray absorption coefficient associated with moisture. The coefficients associated with moisture were in the vicinity of 0.0191 mm–1 and 0.0182 mm–1, the expected value for water and ice, respectively. The coefficient associated with the dry solids was not significant from zero, due in part to scanner resolution limits, limited dynamic range in density values, and to variation in the physical density measurements. The results of this study suggest that internal differences in x-ray absorption within scans of fruit cross-sections are largely associated with differences in volumetric water content.

Network environ theory, simulation, and EcoNet® 2.0

We introduce and codify the mathematics of Ecological Network Analysis (ENA) in general and Network Environ Analysis (NEA) in particular used by the web-based simulation software EcoNet? 2.0. Where ecosystem complexity continues to drive an increasingly vast environmental modeling effort, ENA and NEA represent maturation, in part, of the compartment modeling approach. Compartment modeling mathematically represents compartment storages with both internal-connecting and external-environmental flows as ordinary differential equations. ENA and NEA expand these mathematics into complex systems analysis and corresponding network theory. EcoNet was developed to facilitate the mathematical modeling, to enhance the overall presentation, and to improve the subsequent long-term progress of ENA and NEA systems analysis. Thus, as a continuing enhancement to the overall understanding, but more importantly, to the future growth of environmental modeling associated with ENA and NEA, we derive and summarize the canonical mathematics of ENA, NEA, and EcoNet, which facilitates their future use.

Runoff Curve Numbers for 10 Small Forested Watersheds in the Mountains of the Eastern United States

AbstractEngineers and hydrologists use the curve number method to estimate runoff from rainfall for different land use and soil conditions; however, large uncertainties occur for estimates from forested watersheds. This investigation evaluates the accuracy and consistency of the method using rainfall-runoff series from 10 small forested-mountainous watersheds in the eastern United States, eight annual maximum series from New Hampshire, West Virginia, and North Carolina, and two partial duration series from Georgia. These series are the basis to compare tabulated curve numbers with values estimated using five methods. For nine of 10 watersheds, tabulated curve numbers do not accurately estimate runoff. One source of the large uncertainty is a consistent decrease in storm-event curve numbers with increasing rainfall. A calibrated constant curve number is suitable for only two of 10 watersheds; the others require a variable watershed curve number associated with different magnitude rainfalls or probabilities...

CLASSIFICATION OF SWEET ONIONS BASED ON INTERNAL DEFECTS USING IMAGE PROCESSING AND NEURAL NETWORK TECHNIQUES

Maintaining product quality is the key to success in the fresh fruit and vegetable market. Quality assessment with computer vision techniques is possible; however, two basic issues need to be addressed before an automatic sorting system can be developed: (1) which image features best correlate with the product quality, and (2) which classifier should be used for optimal classification. To address these issues, sweet onions were line–scanned for internal defects using x–ray imaging. Spatial and transform features were evaluated for their contributions to product classification based on internal defects. The Bayesian method was used for selecting the salient features. Spatial edge features combined with selected discrete cosine transform (DCT) coefficients proved to be good indicators of internal defects. A neural classifier performed better than the Bayesian classifier for sorting onions into two classes (good or defective) by achieving an overall accuracy of 90%. Losses and false positives were limited to 6% and 10%, respectively. The accuracy, losses, and false positives for the Bayesian classifier were 80%, 16%, and 17%, respectively.

Automated Planning and Optimization of Lumber Production Using Machine Vision and Computed Tomography

An automated system for planning and optimization of lumber production using Machine Vision and Computed Tomography (CT) is proposed. Cross-sectional CT images of hardwood logs are analyzed using machine vision algorithms. Internal defects in the hardwood logs pockets are identified and localized. A virtual in silico 3-D reconstruction of the hardwood log and its internal defects is generated using Kalman filter-based tracking algorithms. Various sawing operations are simulated on the virtual 3-D reconstruction of the log and the resulting virtual lumber products automatically graded using rules stipulated by the National Hardwood Lumber Association (NHLA). Knowledge of the internal log defects is suitably exploited to formulate sawing strategies that optimize the value yield recovery of the resulting lumber products. A prototype implementation shows significant gains in value yield recovery when compared with lumber processing strategies that use only the information derived from the external log structure.

Physical and mechanical properties of fresh and stored sweet onions

Some physical and mechanical properties of a sample of ‘Granex-Grano’ type sweet onions were examined in an attempt to more deeply understand the conditions under which those onions may most appropriately be grown, harvested, stored, processed, shipped, and marketed. Those onions evaluated had a mean mass, surface area, volume, and density of 98 g, 111 cm2, 95 cm3, and 1 100 kg/m3, respectively. The overall mean equatorial diameter was 62 mm and the mean polar diameter was 42 mm. Of the mature onions examined, 59% of the onions were oblate and 41% were prolate. Crushing load and puncture force are indications of mechanical strength of the onions to withstand mechanical harvesting and postharvest handling. Onions evaluated from this study had a mean crushing load of 26.4 N and puncture resistance of 25.0 N.

EXPERIENCES WITH A FOOD PRODUCT X-RAY INSPECTION SYSTEM FOR CLASSIFYING ONIONS

Maintaining product quality is critical for success in fresh fruit and vegetable marketing. Some onion packinghouses are considering the addition of x-ray inspection systems to their existing optical inspection systems. X-ray systems enable detection of voids that are likely to be associated with the presence of various bacterial or fungal rots in onions. A series of tests were conducted at the University of Georgia Vegetable and Vidalia Onion research and education center in Toombs County, Georgia, with a commercially available x-ray inspection machine. In 2001, two 100-onion batches of medium-sized onions and 100 jumbo onions were machine-inspected, and then halved for a visual internal evaluation. In each series of tests, the accuracy rate was greater than 93% and the false positives were less than 6%. In 2002, two 100-onion batches were run on a similar machine as in 2001. Additionally, in 2002 and 2004, multiple onions with slight to severe defects were each passed through the inspection machine 50 times, respectively, with orientation not controlled to ascertain consistency in defect detection. The machine passed onions with no to slight defect presence (based on subsequent internal visual evaluation of onion halves) nearly 100% of the time. Onions with severe defects were rejected 100% of the time. In 2004, a center rot disease (caused by Pantoea ananztis) study showed that 80% of bulbs that had passed a routine surface inspection and had been deemed to be diseased by the machine in fact exhibited the disorder on halving. False positives were in the 10% to 15% range. In the 2002 and 2004 studies, the machine detected bulbs with disease that passed human visual inspection (HVI) and (in 2004 only) individual tactile grading. These accuracy and false positive rates are very close to the 90% and 10% levels generally accepted for these respective statistics. With appropriate addition of multiple lanes, commercially viable throughputs are possible.

Comparison Of Synthetic and Natural Bulking Agents In Food Waste Composting

Synthetic (plastic) bulking agent was compared with bark bulking agent for the composting of fresh vegetable wastes. Two composting trials of ten weeks duration each were conducted with different quantities of bulking agents. Composting matrix temperature, and oxygen, dimethyl sulfide, and hydrogen sulfide concentrations were measured weekly. Addition of 19% bark or 7% synthetic (dry weight basis) bulking agent in the composting mixes was found to be sufficient to provide thermophilic temperatures and high matrix oxygen concentrations. The lowest oxygen concentrations within the matrix during the thermophilic period of composting were 12.7 and 5.3% for synthetic and bark amended mixes, respectively. The final products from the synthetic bulking agent mixes were screened through a 6.3 mm (0.25 inch) screen to recover bulking agents. Average recovery by screening was 98.9 ± 0.7% by weight. Concentrations of dimethyl sulfide and hydrogen sulfide measured during the experiments were characterized by large variability. The final products in all cases were biologically stable with stability indices ranging from 0.172 to 0.314 mgO2/gVS/h. Maturity indices ranged between 87.4 and 100% indicating a quality product. The product from synthetic amended mixes contained 0.29% by weight of plastic residues.

Nuclear Magnetic Resonance for Moisture Analysis of Meals and Soils

ABSTRACT WITH the advancements in microelectronics for more sophisticated controls and rapid data processing, nuclear magnetic resonance (NMR) can now be used for making certain characterizations of bulk materials. A literature review of limited NMR applications in agricultural systems is presented with anticipation of using NMR on-line in processing and in the field. Moisture contents of two soybean meals and four soils were characterized using one particular type of NMR analysis (the Hahn spin echo T2 analysis) which appears most appropriate for field use. Several established NMR parameters (FID peak, Echo peak) as well as proposed NMR variables (I22 intercept, Echo ratio) are evaluated as to efficacy in quantifying moisture content. The FID peak and Echo peak measurements appear most useful in instances where much prior knowledge about the material is available. The proposed intercept and ratio variable takes more time to compute but may be more useful in cases where considerable variation in bulk material is anticipated.