The invention relates to apparatus and methods for processing shrimp.
Originally introduced because of the high labor costs associated with peeling small shrimp by hand, shrimp-peeling machines are now widely used in the shrimp processing industry. Roller-type shrimp-peeling machines, in particular, dominate the bulk peeling industry. U.S. Pat. Nos. 2,778,055 and 2,537,355, both to Fernand S., James M., and Emile M. Lapeyre, describe the basic structure and principles of operation of roller-type shrimp peelers.
Many factors affect the throughput, quality, and yield of peeled shrimp. Some factors related to the shrimp themselves include the species, size, uniformity, and freshness of the shrimp. Factors related to the peeling equipment, include the feed rate of shrimp to the peeler, water flow to the peeler, and finger-frame pressure. Other factors relate to other shrimp-processing equipment, such as cleaners, shrimp feed systems, roller separators, air separators, and graders. The equipment-related factors are generally manually adjustable to improve peeling quality and yield for a given batch of shrimp or to compensate for peeling-roller wear. Because the quality and yield of the peeled shrimp directly affect their production cost and the price they can command, proper adjustment of the peeling equipment is important. But proper manual adjustment requires diligent monitoring of the output quality and yield and experience in selecting the adjustments that should be made.
One version of a method embodying features of the invention for processing peeled shrimps includes: (a) distributing deheaded, peeled shrimps onto a support surface; (b) creating a digital image of the shrimps on the support surface; (c) determining the number of tail segments present in each of the shrimps from the digital image; and (d) classifying each of the shrimps into one of a plurality of classes according to the number of tail segments present in each of the shrimps.
A shrimp-processing system embodying features of the invention comprises a shrimp peeling machine removing the heads and shells from shrimp to produce peeled shrimps. A conveyor conveys the peeled shrimps from the shrimp peeling machine to downstream processing. A vision system captures a digital image of peeled shrimps on the conveyor. A processor determining the number of tail segments present in each of the shrimps from the digital image and classifies each of the shrimps into one of a plurality of classes according to the number of tail segments present in each of the shrimps.
Another shrimp-processing system embodying features of the invention comprises a peeler for removing the shells and heads from unpeeled shrimps to produce peeled shrimps. A feed system feeds unpeeled shrimps to the peeler. A conveyor system conveys peeled shrimps from the peeler downstream along a process path. One or more shrimp-processing machines downstream of the peeler in the process path act on the peeled shrimps. One or more sensors disposed along the process path detect one or more operational variables of the shrimp-processing machines or of the peeler or physical characteristics of the shrimp and produce sensor signals indicative of the one or more operational variables. A processor receives the sensor signals and derives control signals from the sensor signals. The control signals are sent to the peeler or the one or more shrimp-processing machines to adjust one or more operational settings
These features and aspects of the invention are described in more detail in the following description, appended claims, and accompanying drawings, in which:
One way to determine the quality of the peel is by counting the number of contiguous tail segments of the peeled shrimps exiting a peeler and classifying each peeled shrimp as High Quality, Medium Quality, or Low Quality. For example, each shrimp having six full tail segments S1-S6 (with or without all of its telson) could be classified as High Quality (
An automated peeling system embodying features of the invention is shown in
A vision system 20 including one or more cameras 22 captures a frame image of the shrimps on a portion of the conveyor. The vision system 20 produces digital images of the shrimps 14 on the conveying surface 18. The shrimps generally rest side-down on the conveying surface 18, which may be a darker surface than the shrimp meat to provide contrast for better imaging. The digital image of the frame is sent to a processor 24, which processes the image. Imaging algorithms detect physical characteristics, or features, of the shrimp, such as, for example, outer upper and lower edges showing the indentations (I, J;
One exemplary version of a control scheme usable with the system of
The process described with respect to the flowchart of
Other vision-system algorithms can alternatively be used to determine each shrimp's volume. In a one-camera, two-dimensional image capture, the shrimp's projected area, its perimeter, its arc length, or other dimensional attributes, which may be used to determine the segment count, could also be used to estimate the shrimp's volume. With knowledge of the meat density (weight/volume) of the species of shrimp being measured, the weight of each shrimp piece is determined by multiplying the volume by the density. An exemplary weight-based algorithm for calculating yield and throughput is depicted in the flowchart of
Some of the missing segments exit the process upstream of the camera position. For example, some of the missing segments are pulled through the peeler rollers and discarded. But other of the segments missing from the shrimps are conveyed to the vision system and imaged. That's why those missing segments, or bits, if they are not going to be sold, are not counted in calculating yield or throughput. The calculated weight of each non-bit shrimp piece is added to the accumulated weights for its class 90. The six-segment weights are summed 92 as well to compute a running six-segment total weight; i.e., what the accumulated weight of the peeled shrimp in the sample would be if all the shrimp had their six segments intact. The yield is computed 94 for each class by dividing the accumulated weights for each class by the accumulated six-segment weight of the sample. The throughput of the sample of peeled shrimp by class and overall is derived 96 by dividing the accumulated weights for each class and the six-segment weight by the time interval represented by the sample. The yield and throughput computations do not have to be performed at the same rate as the weight-summing, which is performed as each shrimp is analyzed. For example, the yield and throughput calculations could be performed only once per frame and could be filtered with previous values as described with reference to
Thus, the vision system just described is based on estimated shrimp weights rather than on shrimp counts. The vision system identifies which segments are missing from individual shrimps to determine the quality level of each shrimp (from its number of intact segments). Then one or more segment-weight algorithms (using empirically determined curves as in
The accuracy of the weight-based methods can be improved by the addition of an in-line, real-time weight measurement provided by a weighing device 28 (
The visioning system with its tail-segment-counting, throughput, yield, and other algorithms can be integrated into a larger automated shrimp-processing system as shown in
A global control processor 70 is used to monitor and control the entire shrimp-processing system 50. The global control processor can be realized as a single central processor or a network of distributed processors. The global control processor receives image data over input lines 71 from the peeler-output visioning systems 20 and other visioning systems 72 positioned at various points along the shrimp's process path through the system. The processor 70 can receive sensor signals from other sensors measuring other system variables, such as temperatures, weights, and speeds. For example, shrimp entering the peeler 12 can be imaged to determine if the throughput is too high. The output of the cleaners 60 can be monitored to determine the quality of the cleaning process. Likewise, the qualities of the roller-separation and air-separation processes can be determined by monitoring the outputs of the roller separators 62 and the air separators 64. The inspection belts 66 can be monitored to check on the infeed rate of shrimp to the graders 68. Besides the Model A peeler, Laitram Machinery, Inc. manufactures and sells other shrimp-processing equipment, such as the Laitram® Automated Fee System, the Model RTFS Rock Tank and Feed System, the Model C Cleaner, the Model S Roller Separator, the Model AS Air Separator, the Model IB Inspections Belt, and the Model G-8 Grader. Equipment such as the Laitram Machinery equipment mentioned is outfitted with actuators that can adjust various operational parameters of the equipment in response to control signals. The processing-equipment stations downstream of the peeler are linked by a conveyor system that may include conveyor belts, elevators, fluid conduits, or other transport apparatus transporting shrimps along the process path. The global control processor 70 runs algorithms and routines that develop shrimp images from the vision data and compute throughput, quality, and yield results at various points in the shrimp-processing system. The results can be displayed and used to derive control signals to automatically control the operation of the system over processor control output lines 74 to improve quality and yield. For example, the processor 70 can control the rate of delivery of shrimp to the rock tank 54 by controlling the speed of the input pump 52 or conveyor. The processor can also control the rate of delivery of shrimp from the rock-tank system's receiving tank to the feed tank 56 by controlling the speed of the receiving tank's pump. Both of these adjustments may depend on the volume of shrimp in the feed system 58 being distributed to the peelers 12. The volume of shrimp or their feed rate, or throughput, is determined from the vision system or from weigh scales. If the throughput is too high, the infeed pumps can be slowed and the speed of conveyor belts in the feed system can be slowed. If the throughput is too low, it can be increased by increasing the speeds of the pumps and the conveyor belts. Just like the processor 24 in
Number | Name | Date | Kind |
---|---|---|---|
2964181 | Demarest et al. | Dec 1960 | A |
4079416 | Faani et al. | Mar 1978 | A |
4790439 | McIntyre | Dec 1988 | A |
4818380 | Azegami | Apr 1989 | A |
4819176 | Ahmed | Apr 1989 | A |
4847954 | Lapeyre et al. | Jul 1989 | A |
4934028 | Stipe | Jun 1990 | A |
4963035 | McCarthy et al. | Oct 1990 | A |
5020675 | Cowlin | Jun 1991 | A |
5064400 | Stipe | Nov 1991 | A |
5165219 | Sekiguchi | Nov 1992 | A |
5184732 | Ditchburn | Feb 1993 | A |
5195921 | Ledet | Mar 1993 | A |
5229840 | Arnarson et al. | Jul 1993 | A |
5246118 | Mosher | Sep 1993 | A |
5346424 | Chiu et al. | Sep 1994 | A |
5352153 | Burch et al. | Oct 1994 | A |
6200209 | Shelton | Mar 2001 | B1 |
6808448 | Kanaya | Oct 2004 | B1 |
8189901 | Modiano | May 2012 | B2 |
20120085686 | Radema | Apr 2012 | A1 |
20140168411 | Ledet et al. | Jun 2014 | A1 |
Number | Date | Country |
---|---|---|
1-202241 | Aug 1989 | JP |
04058373 | Feb 1992 | JP |
03-013745 | Feb 2003 | WO |
Entry |
---|
“USDC/NOAA/Seafood Inspection Program, United States Standards for Grades of Fresh and Frozen Shrimp,” Jan. 14, 2009, XP055109806, retrieved from the Internet: URL: http://www.seafood.nmfs.noaa.gov/ShrimpFreshandFrozen.pdf [retrieved on Mar. 25, 2014] p. 4, paragraph iii (a). |
Murat O. Balaban, Sencer Yeralan, Ymir Bergmann, “Determination of Count and Uniformity Ratio of Shrimp by Machine Vision,” Journal of Aquatic Food Product Technology, vol. 3(3) 1994, pp. 43-58, The Haworth Press, Inc. |
Pan et al., “Prediction of shelled shrimp weight by machine vision,” Journal of Zheijiang University Science B, 2009 10(8): 589-594. |
Dah-Jye Lee et al., “An Efficient Shape Analysis Method for Shrimp Quality Evaluation,” 2012 12th International Conference on Control, Automation, Robotics & Vision, Guangzhou, China Dec. 5-7, 2012. |
Number | Date | Country | |
---|---|---|---|
20140168411 A1 | Jun 2014 | US |
Number | Date | Country | |
---|---|---|---|
61739457 | Dec 2012 | US |