September 17, 2014 TO: FROM: SUBJECT: USHBC Mechanical Harvesting and Handling Subcommittee Fumiomi Takeda, Research Horticulturist USDA-ARS, Appalachian Fruit Research Station, Kearneysville, WV Progress Report for project titled Improving blueberry mechanical harvest efficiency: Quantifying with blueberry impact recording device (BIRD) and developing information to assist in reducing soft berries in machine harvested blueberries Objectives for 2014: 1. Evaluate packing lines with and without foam padding to reduce bruising using the instrumented sphere (Blueberry Impact Recording Device) sensor 2. Evaluate blueberries processed through padded and non-padded packing lines by assessing fruit bruise 3. Measure impact and compression forces on blueberries during transit from the field to packing facility Project Director: Principal Investigator: Co-Principal Investor: Dr. Fumiomi Takeda, Research Horticulturist USDA-ARS, Kearneysville, WV 25430 (304) 725-3451, Fumi.Takeda@ars.usda.gov Dr. Changying (Charlie) Li, Associate Professor University of Georgia, Athens, GA (706) 542-4697, cyli@uga.edu Dr. Gerard Krewer, Professor Emeritus University of Georgia Krewer Consulting, Woodbine, GA (229) 392-1388, gkrewer@uga.edu Mr. William Cline, Extension Plant Pathologist North Carolina State University, Castle Haynes, NC (910) 675-2314, bcline@ncsu.edu
Summary Currently, there are two methods to evaluate a fruit packing line. One method is to evaluate the bruising rate of fruits after running large number of samples through the packing line, which is time consuming and cannot reveal how the fruit bruised within the packing line. A more widely used method is to record the impacts during the packing process using an Instrumented Sphere (IS). We have termed our IS as a Blueberry Impact Recording Device (BIRD) and it is designed to relate the impacts with fruit bruising rate. In essence, the BIRD is a miniaturized data logger with size and weight of a real blueberry fruit that can mimic the mechanical behavior of a fruit when subjected to the same mechanical stress as a fruit. Therefore, the BIRD can provide quantitative information of the impacts created by the packing line, helping researchers and packing house operators understand the internal mechanical damage to fruits. In 2014, our research team evaluated 5 packing lines in Florida and North Carolina. The analysis included the measurements of drop heights at transition points within the packing line ad measurements of impacts on BIRD sensors going through these transitions points with and without padding. Also, Li and Takeda, along with collaborators from FL, GA, NC, PA, MI, MS, OR, CA, and WA and with financial pledges and commitments from USHBC, and blueberry growers and equipment vendors/manufacturers in FL, MS, GA, NC, MI, CA, and WA, submitted a research proposal to the USDA National Institute of Food and Agriculture Specialty Crop Research Initiative program in June 2014. The project titled Scale-Neutral Harvest-Aid System and Sensor Technologies to Improve Harvest Efficiency and Handling of Fresh-Market Highbush Blueberries, received a high mark from the review panel. I am happy to report to the subcommittee that Dr. Li has received a phone call from NIFA SCRI National Program Leader that this project will be funded for $2,500,000. Our project will run for 4 years with the following objectives: Please note that some of the objectives were in our original proposal submission to this subcommittee. 1) Develop highthroughput phenotyping technologies to aid the selection of southern and northern highbush blueberry genotypes suitable for semi-mechanical harvest, 2) Design a new semi-mechanical harvest-aid system for efficient mechanical fruit harvesting for small- and medium-size blueberry farms, 3) Develop the next-generation berry impact recording sensor (BIRD NXT) and use it to improve harvest and postharvest operations through a critical understanding of mechanical impacts, 4) Determine microbial contamination and critical control points along the harvest and postharvest chain with the new harvest system, and 5) Conduct ergonomic and socio-economic analyses of the developed technologies and determine the implications to the rural community through outreach and technology transfer. In December 2014, Dr. Li, Mr. Rui Xi, a MS student who worked on the BIRD II research and development, and I will be travelling to Chile s (Talca, Linares, Challan) blueberry production areas to evaluate packing houses. Local arrangements are being coordinated by Dr. Jorge Retamales at the University of Talca and Mr. Andris Armstrong of the Chilean Blueberry Committee. We also plan to perform a preliminary field work related to the work described under Objective 2 of the SCRI project. We will have four working BIRD sensors and detailed operating manuals for USHBC in December 2014 and have the final report to the committee in early 2015. Finally, you will find a research paper from our team that was published in 2014.
RESEARCH UPDATE Objectives for 2014: 1. Evaluate packing lines with and without foam padding to reduce bruising using the instrumented sphere (Blueberry Impact Recording Device) sensor Status: MET 2. Evaluate blueberries processed through padded and non-padded packing lines by assessing fruit bruise Status: MET 3. Measure impact and compression forces on blueberries during transit from the field to packing facility STATUS: PARTIALLY MET Blueberry packing line In 2013 and 2014, 11 commercial blueberry packing lines have been evaluated using the BIRD I and II sensors (Table 1). Four packing lines highlighted in BOLD (Line No. 7 11) were evaluated in 2014. Some lines(hoyt and Jack Tanner in GA) were evaluated with blueberries, while others were evaluated as empty packing lines. The packing line in Florida was tested using the third generation sensor (BIRD II) in April 2014. Of particular note is that in some packing lines the evaluations were performed before and after adding padding on some critical transfer points. Table 1 Summary of packing lines. The packing lines evaluated in 2014 are presented in bold letters. Line Number Farm Name Location Time Sensors Number of Replicates Number of Transfers 1 Hoyt GA June, 2013 BIRD I 5 6 W 2 Jack Tanner GA June, 2013 BIRD I 5 7 W 3 Bird Plantation MI August, 2013 BIRD I 4 4 O 4 Harris Blueberries MI August, 2013 BIRD I 6 4 O 5 Jubilee Farms MI August, 2013 BIRD I 6 4 O 6 New Day Farm MI August, 2013 BIRD I 5 5 O 7 Lewis Farm NC May, 2014 BIRD I 6 7 O 8 Lewis Farm NC May, 2014 BIRD I 6 6 O 9 Sweet Berry NC May, 2014 BIRD I 6 6 O 10 Sweet Berry NC May, 2014 BIRD I 6 7 O 11 Straughn Farm FL April, 2014 BIRD II 6 7 O W/O Fruit A typical commercial blueberry packing line has four main types of line components: blower, color sorter, soft sorter, and conveyor belts that connect each component (Fig. 1). Different
combination and alignment of these components forms different line with different number of transfers. The specific structure of the tested lines is shown in. Typically, blueberries are dumped at the beginning of the packing line and then conveyed to the wind through a lifting conveyor. The wind blower can blow out light blueberries, leaves, and other foreign trash. The convey belt for the wind blower is usually made of steel with holes that allow let small foreign trash go through. Because of the wind from the blower, the blueberries could bounce several times and impacts will happen during this section. However, the BIRD I sensor may not record any impacts because of it large weight keeps it stay on the conveyor. After the blower, the blueberries will go through the color sorter and soft sorter, to remove immature and soft fruit. A line can have multiple color sorters or soft sorters and the color sorter and softer can use same conveyor. After the color sort or soft sorter, the blueberries will be inspected by workers at the manual sorting conveyor belt. The blueberries will be transferred through one or multiple conveyors before they transfer to hopper are packed into clam shells. Fig. 1. Schematic of a typical blueberry packing line
Fig. 2. Schematic for the ten packing lines evaluated by BIRD I sensor. The drop height for certain transfer points are missing due to the unmeasurable position. For the line 1 to 10, the BIRD I sensor was placed gently at the beginning of each line and allowed to go through the line to the clamshell or collecting tray in the end, recording the impacts at each transfer (point of drop or roll form one operation to the next). Multiple impacts may be recorded at each transition point due to the bounces of the BIRD I sensor. Based on preliminary test, the BIRD I sensor was set to record impacts larger than 25 g at a frequency of 2 KHz. Each packing line was tested at least 4 times in order to get a reasonable estimation of the impact damage. For the line 11, the BIRD II sensor was manually dropped on the packing line and it went
through the empty packing line 6 times. Then the line was modified by adding padding on certain transfer points. The same BIRD II sensor was used to test the modified line for 6 times. The BIRD II sensor was set to record impacts larger than 25 g at 2 KHz (e.g. 500 times per second). Line Modification: The line 11 (Fig. 3) was first tested using BIRD II sensor and after checking the data recorded by the sensor on the empty line, transitions 1, 5, 6, 7 and 8 were identified as critical points which produced the largest impacts. Water-proof padding sheet (Poron No- Bruise, A&B Packing Equipment Inc.) were padded on these transition points to reduce the impacts. For transition 6, the padding worked like a ramp which reduced the speed of the sensor when it dropped from the manual sorting conveyor belt to the following conveyor belt. Only one side of the collecting hopper was covered by the padding. After modification, the packing line was tested 6 times using the same sensor with the same procedure described above. Using the 6 replication, the average impact level (average, maximum and cumulated peakg) was calculated for each transition point. Fig. 3. A schematic drawing line 11 (Straughn Farm, Windsor, FL). The top pictures show the transition point with padding (circled). Bottom row of photos show the transition points where padding was added e.g. 1: dump to the conveyor belt. 2: from conveyor belt to wind blower. 3: from wind blower to color sorter. 4: form color sorter to soft sorter. 5: from soft sorter to manual sorting belt. 6: from manual sorting belt to a conveyor belt. 7 from a conveyor belt to another belt. 8: from a conveyor belt to collecting hopper. Data processing and analyzing: The raw data recorded by the BIRD I and BIRD II sensor were processed using Matlab. Each replicate of each line was first separated into single sections which only contain the impacts at each transfer point. Then the average impact level (average, maximum and accumulated peakg) at each transfer point were calculated and summarized
using the all the replicates. The average impact level for each packing line was also calculated. The maximum peakg relates to the maximum damage to the fruit and the accumulated peakg reflects the accumulated damage by several impacts. The velocity change reflected the hardness of the contact material. In general, soft material (e.g. cushion) has larger VC and smaller peakg comparing to hard material (e.g. steel) if the sensor was dropped from the same height onto them. In order to accurately synchronize the video with the impacts recorded by the sensor, a LabView based program was designed to align the video display with the impact data. There are two panels to in the program: one is used to display the video; the other one graphs the impact data and a cursor was used to indicate the current time of impact data. The user first manually aligns the cursor with the first impact in the data file and the first impact occurred in the video, and then the program will automatically move the cursor as the video plays so the use can associate the impact with the location of the sensor. Results and Discussion BIRD I sensor reading. Although the amplitude recorded by the BIRD I and BIRD II are not comparable because of the different surface property of the casting material, the impacts distribution over time recorded by the two sensor are identical when both record impacts above the threshold. Video record showed a few impacts occurred when the sensor hit on the sidewall of the conveyor belts. However, these impacts were excluded from the data analysis since they are insignificant and rarely happened in normal operation with fruits. At each transfer point, the sensor recorded a big impact at the initial drop which is followed by several small impacts due to the bounce of the sensor. The largest number of impacts were recorded when sensor was transferred to the hopper when the sensor bounced against the steel sides. The amplitude of the impacts mainly depended on the drop height of the transfer point, the surface properties of the contact conveyor and the speed of the conveyor belt. Note that most lines had design features at one or more transition points where drop heights were greater than 30 cm (12 in.) (Figs. 2 and 3). For example, in Line 11, the BIRD I sensor recorded the biggest impact when transferred from color sorter to soft-fruit sorter, due to a ~ 16-in drop. The speed of the conveyor belt for color sorter and softer sorter are normally fast than other conveyor belts with one packing line, which is reflected from the time interval between two successive big impacts. Therefore, big impacts usually occurred at the transfer points when fruits come out from color sorter and soft-fruit sorter.
Fig. 4. Real time response of the BIRD I sensor reading for a single run (a) with the schematic of the measured packing line (a).
There are multiple transfer points at each line and most of impact surface are steel, plastic and rubber (Fig. 5). An average peakg of 60 to 111g, maximum peakg of 129 to 402g, cumulative peakg of 879 to 3196g and number of impacts of 15 to 38 were recorded on the 10 lines. Among all the 10 packing lines, line 1 and 2 has the smallest impact level due to the cushion of the fruit on some transfer points. Among 8 empty lines, line 3 has the smallest average and cumulative peakg and number of impacts and line 4 has the smallest maximum peakg, since line 3 and 4 has only 4 transfer points which are the smallest among all the lines. Line 5 also has 4 transfer points, however, due to the large drop height in the last transfer point, line 5 has the largest maximum peakg. The largest average and cumulative peakg was found in Line 9. Fig. 5. Overall impact level for the ten packing lines that were tested with BIRD I. The white bars indicate the lines that were tested with blueberries.
The impact level on each transfer points are plotted in Fig.. Among all the transfer points of the 8 empty lines, the last transfer point (clam shell filler) produced the largest impact level for most of the lines, which is expected because the sensor dropped into the hopper or tray from a relative large height (15 to 28 cm) and bounce several times along the sides. Although the blueberries in the hopper can reduce the impacts during normal operation, it is still recommended to cushion the side of the hopper to minimize the impact damage to the fruits. Fig. 6. Bubble plot of the impact level for the transfer point when transferred onto indicated component. The PeakG value is proportional to the area of the circle. The blue, red and black circle presents the average, maximum and cumulative PeakG value respectively. The two packing line under the dash line were tested with blueberries. The right plot shows the reference size with the indicated value.
Effect of line modification with padding material. The BIRD II sensor recorded impacts occurred on line 11 before and after line modification (Fig. 7). The impact histogram of the entire line (Fig. 7a) showed that the number of small impacts (below 25 g) was increased after padding the line due to the cushioning of the padding. The largest impacts at transition 1 (Fig. 7b) were reduced to less than 125 g after padding the line while the impacts could be as large as 300 g before the padding. Large impacts above 100 g in transition 5 (Fig. 7c) and 6 (Fig. 7d) were reduced to less than 100g and the small impacts below 25 g were substantially increased after modification. At transition 7, not only the impacts become smaller but also the number of impacts was reduced. As mentioned before, the padding at transition 7 (Fig. 7e) was not rigidly fixed and worked like a ramp. Because of the loose fixation, the displacement of the padding prevented the BIRD II sensor bouncing when it hit on the padding. Therefore, the number of impacts was fewer. However, the reduction of number of impacts was not observed at other padded transition points because the padding at these transition points were rigidly fixed and the BIRD II sensor could bounce several times as it hit on the padding. The difference in number of impacts suggests that using loose fixation for the padding can prevent bounces from the blueberries, thus reducing the number of impacts. Fig. 7. Impact histogram of the entire line and at each transition point. g = 9.8 m/s2. a) Impact histogram of the entire line. b) Impact histogram at transition 1. c) Impact histogram at transition 5. d) Impact histogram at transition 6. e) Impact histogram at transition 7. f) Impact histogram at transition 8. Count Count 140 120 100 80 60 40 20 0 16 14 12 10 8 6 4 2 0 a d Unpadded line Padded line 0 50 100 150 200 250 300 PeakG (g) Unpadded line Padded line 0 50 100 150 200 250 300 PeakG (g) Count Count 40 30 20 10 0 25 20 15 10 5 0 b Unpadded line Padded line 0 50 100 150 200 250 300 e PeakG (g) Unpadded line Padded line 0 50 100 150 200 250 300 PeakG (g) Count Count 25 20 15 10 5 0 50 40 30 20 10 0 c Unpadded line Padded line 0 50 100 150 200 250 300 f PeakG (g) Unpadded line Padded line 0 50 100 150 200 250 300 PeakG (g)
Bruise Assessment. A system approach is needed to advance the mechanical harvest of highbush blueberries for fresh market. Breeding for firmer blueberries, cultural practices for machine harvesting, engineering to improve harvesters and reducing mechanical impacts and bruising can deliver higher quality fruit and increase harvest efficiency. In August 2013 and June 2014, we evaluated nine cultivars in MI (Elliott, Jersey, Draper, Aurora, Nelson, Legacy, Brigetta, Liberty, and Bluecrop) and four in NC (O Neal, Star, Reveille, and Farthing). In Michigan, hand-harvested samples were dropped either from 24 or 48 inch height onto an inclined catch plate. Dropped berries were held for 24 hours at room temperature, after which berries were checked for firmness (Table 3) and sliced to be evaluated for bruise (Figs. 8 and 9). Table 3. Effect of drop height on fruit firmness. Drop height (in) Cultivar Control (not dropped) 24-in 48-in Aurora 184 160 141 (43) Bluecrop 178 164 140 (38) Brigetta 212 173 148 (64) Draper 237 213 183 (54) Elliott 178 155 137 (41) Jersey 208 163 149 (59) Legacy 209 193 172 (39) Liberty 200 163 154 (46) Nelson 199 182 155 (34) Fig. 8. Effect of blueberries falling 24-inches or 48 inches onto a catch plate. Aurora (left) and Draper (right)
Fig. 9 Effect of Legacy blueberries falling 24- or 48-inches onto a catch plate. To simulate what could be happening to blueberry fruit in the packing line, we conducted a study in which Bluecrop blueberries were dropped once onto a hard surface from 16-inch height or 5 times from only 4-ich height (Fig. 9). Fig. 9. Effect of single and multiple drops onto a hard surface.
Impact (g) Fig. 10. The relationship between impact measured by the BIRD sensor and drop height. The BIRD sensor was dropped from a height of 6, 12, 18, and 24 inches onto an inclined or flat BEI catch plate. 800 700 600 500 400 300 200 100 0 0 6 12 18 24 30 Drop Height (inch) Plastic incline Plastic flat Loss of fruit firmness and development of bruise damage was evident after only 24 hours of drop tests. This was clearly evident in O Neal and Star samples that were dropped from a height of 24 inches. Mature highbush blueberry plants are generally 5-feet-tall. When they are machine harvested, the majority of detached fruit fall more than 36 inches onto a hard catch plate surface. In the case of our study, the drop heights ranged from 6 to 24 inches and fruit bruising in O Neal, Star, Reveille, and Farthing was not as extensive as we had seen with northern highbush varieties that have been dropped from >40 inch heights. We believe the bruising will become more evident as fruit are stored longer. During the trip to Chile in December, the team will collaborate with Chilean colleagues to evaluate different drop heights and impact material or surfaces. Fruit samples will be segregated by firmness (soft, medium, or firm at harvest), and size. Biological and physical tests of drop fruit will be performed as many as 30 to 45 days of cold storage. This study and our other studies confirmed that there is a large variation in bruise susceptibility among highbush blueberry genotypes. Selection FL 05-528 was identified as the most promising machine harvestable genotype among the four FL genotypes compared previously. Among the cultivars evaluated in MI and NC a strong relationship was established between the impact data recorded by the BIRD and the bruising rates. The bruising data on padded and hard surfaces for these genotypes indicated significant benefits from padding. Future work with the BIRD sensor will delineate different padding types. The BIRD sensor and the interpretation method developed to relate fruit damage to BIRD data has enabled
researchers to evaluate various padding material and machine designs in terms of the potential bruise damages they may create in the blueberry fruit. Such device will be made available to USHBC later this year. FIELD STUDIES CONDUCTED IN 2014 According to Brown et al. (1996), 78% of the blueberry harvested by commercial mechanical harvesters had severe bruise damage compared with 23% for hand harvested fruit. The percentage of bruised fruit can be decreased significantly when the novel, crisp-textured blueberry genotypes are machine-harvested (Mehra 2013, Takeda et al., 2013), as documented in our previous SCRI project. Still, more technological improvements are needed to further reduce bruising, limit percent ground loss of fruit, and reduce detachment of immature (green or red) fruit. In addition to these key technical limitations, the Over-The-Row mechanical harvesters (~$100-200 k) are generally not affordable to small- and medium-sized growers. In many blueberry producing states such as Oregon (W. Yang, pers. comm.) and Florida (J. Williamson, pers. comm.), most of the blueberry farms are under 50 acres. As an alternative to hand harvesting, portable hand-held shaking devices (e.g. Campagnola pneumatic system) have been used to harvest fruit onto a small, manually positioned catch frame. However, the hand-held shaker-harvest technology has two major problems that limit its adoption: low efficiency and potential worker fatigue and injury. The handheld shaker is powered by electro-mechanical and pneumatic systems and is operated by 1 or 2 workers. The workers must position the catch frame under the plant, move the air compressor, shake the fruit off the plant, and then handle the fruit collection box or have other workers handling the catch frame and collection boxes, and move the compressor. Field observations indicate that the workers are actively engaged in an assigned task of fruit harvesting less than half of the time. Thus, the pace of fruit harvesting is slow and tedious. In addition, the workers can easily get fatigued by operating the handheld shaker because of vibration and the weight of the shaker (7lbs) and other laborious operations such as constantly moving the catch frame. It is well documented that any repetitive job having potentially excessive frequencies, high forces and pure limb postures could result in work-related musculoskeletal disorders (WRMSDs) (National Research Council 2001). In agricultural work, WRMSDs can account for as much as 28% of the ~200,000 lost-time injuries on U.S. farms, whereby hand/wrist problems were a significant proportion of these injuries (Meyers 2001, Gomez 2003). Due to these limitations, the handheld shaker devices have not been widely adopted among blueberry growers. For instance, our survey results showed that only 10% of surveyed growers tried to use handheld shakers and the majority of the feedback of using the device was not favorable. To overcome the key limitations of the current handheld shaking devices and provide an alternative to the traditional ORT mechanical harvesters, a platform-based harvest-aid system proposed in this project could offer a promising solution. This harvest-aid can accommodate at least four workers each operating a hand-held shaking device on a self-propelled platform. Compared with existing harvesting methods (OTR harvesters, hand-harvest, and the handheldshaker-harvest), the semi-mechanical harvesting system offers the following four benefits:
1) It can significantly improve harvest efficiency compared with hand-harvest. The workers do not need to position the catch frame under the plant, move the air compressor, shake the fruit off the plant, and then handle the fruit collection box. Because of the elimination of these steps, efficiency is expected to be improved 10-20 times compared with hand-harvest. 2) The harvest aid system will significantly reduce fatigue because the workers do not need to hold the shaker all the time (the shakers are tethered and counter-weighted by the harvest aid). Injury risk is lower because shaker vibration will be mitigated. A comprehensive ergonomics study will be conducted to assess and improve workers stress and injury risk. 3) Ground losses are expected to be greatly reduced due to a new design of the conveyance system. A new catch frame design at the base of the harvester will use a deceleration device powered by upward air flow. This would reduce bruising and increase fruit quality substantially. 4) The system will be scale neutral: its performance will be appealing to large and small growers alike, whereas its cost will be substantially lower than a regular OTR harvester and be affordable to small- and medium-sized growers A preliminary study was conducted in 2014 in which O Neal blueberry was harvested by hand-held devices with fruit falling into a netted catch frame, and with a traditional over-the row (OTR) harvester. The internal bruising data showed that fruit harvested with hand-held devices resulted in much less internal bruise damage than those harvested by the OTR (Fig. 2). Also, our study showed that showed fruit softening or loss of firmness and extent of internal bruising were closely linked (Table 2). Fig. 2. Effects of harvest methods (hand, two handheld shakers, and Little Blue over-the-row harvester) on fruit bruise. Table 2. Effect of harvesting methods on fruit firmness and bruise incidence in O Neal blueberry. Harvest method Hand Model H-1 Campagnola Little Blue Firmness (g/mm) 248 ± 43 191 ± 38 222 ± 42 148 ± 39 % Bruise 4 ± 9 9 ± 11 16 ± 11 39 ± 4
Conclusion The packing house evaluations that were performed in 2013 and 2014 using BIRD I and BIRD II sensors quantitatively measured the mechanical impacts created during the packing line process. One packing line was modified by adding padding on critical transfer points. There were many transfer points that created high impact level and most of them were constructed of hard-surface material. The highest impact level were caused by collecting hoppers. The presence of fruits could lower the reading of the sensor on some transfer points due to the cushioning from the fruit, therefore, impact level may be overestimated for the packing lines tested without any fruit. When the critical transfer points were cushioned impact levels were significantly decreased. Sensor readings clearly showed that cushioning is an inexpensive and efficient way to reduce the impact damage. Laboratory test must be complete to determine the relationship between the impact level and the bruise rate of the fruit so we can accurately estimate the bruise potential for each packing line. However, the potential for bruise damage seems high at some transfer points. Manuscript in preparation: Evaluation of blueberry packing lines using a miniaturized instrumented sphere (authors are Rui Xi, Changying Li, Fumiomi Takeda, and Gerard Krewer). Proposed journal: Applied Engineering in Agriculture, a publication of American Society of Agricultural and Biological Engineering. Research article published in peer-reviewed journal in 2014: Yu, P., C. Li, F. Takeda, and G. Krewer. 2014. Visual bruise assessment and analysis of mechanical impact measurements in southern highbush blueberries. Applied Engineering in Agriculture 30:29-37. Yu, P., C. Li, F. Takeda, G. Krewer, G. Rains, and T. Hamrita. 2014. Measurement of mechanical impacts created by rotary, slapper, and sway blueberry mechanical harvesters. Computers and Electronics in Agriculture 101:84-92. In 2014, as an outreach effort, the following presentations on blueberry harvest mechanization: At the Gulf-South Blueberry Growers Association Blueberry Education Workshop (February 2014 in Hattiesburg, MS) At the 2014 University of Florida SCRI blueberry in-service agent training workshop (March 2014 in Gainesville, FL) At the XXIX International Horticultural Congress (August 2014, Brisbane, Australia)