Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: SORTING AGRICULTURAL MATERIALS FOR DEFECTS USING IMAGING AND PHYSICAL METHODS

Location: Foodborne Toxin Detection and Prevention

Title: Real-time methods for non-destructive detection of pits in fresh cherries

Authors
item Jackson, Eric
item Haff, Ronald
item Gomez, Joseph

Submitted to: Proceedings of the American Society of Agricultural and Biological Engineers International (ASABE)
Publication Type: Proceedings
Publication Acceptance Date: April 28, 2009
Publication Date: N/A

Interpretive Summary: The presence of pits in processed cherries is a concern for both processors and consumers, in many cases causing injury and potential lawsuits. While machines used for pitting cherries are extremely efficient, if one or more plungers in a pitting head become misaligned, a large number of pits may pass before corrective action is taken, and consequently the overall pit count is generally much higher. X-ray imaging has been used to detect pits, but the traditional equipment is both expensive and bulky, and implementation on the processing line is cumbersome. A simpler, faster, and more economical x-ray inspection system has been developed by modifying a traditional x-ray detector. This has reduced the data collection process from a two dimensional image to a much simpler one dimensional signal, which results in faster and simpler processing and classification of the resulting signal than conventional imaging and processing. This simplification reduces the size and complexity of the x-ray detection system significantly, leading to lower cost and greater ease of implementation. An algorithm designed to differentiate unpitted from pitted cherries yielded recognition rates of 97.3% for the pitted and 94% for the unpitted cherries, with a total error rate of 3.5%. When the algorithm was adjusted to maximize removal of pitted fruit, 100% of pitted cherries were detected with a total error rate of 8.5 percent. If orientation could be controlled after pitting, total error could be as low as 1%.

Technical Abstract: The presence of pits in processed cherries is a concern for both processors and consumers, in many cases causing injury and potential lawsuits. While machines used for pitting cherries are extremely efficient, if one or more plungers in a pitting head become misaligned, a large number of pits may pass before corrective action is taken, and consequently the overall pit count is generally much higher. X-ray imaging has been used to detect pits, but the traditional equipment is both expensive and bulky, and implementation on the processing line is cumbersome. A simpler, faster, and more economical x-ray inspection system has been developed using an array of photodiode based x-ray detectors in a linescan configuration whose outputs are combined. This reduces the data collection process from a two dimensional image to a much simpler one dimensional signal, which results in faster and simpler processing and classification of the resulting signal than conventional imaging and processing. This simplification reduces the size and complexity of the x-ray detection system significantly, leading to lower cost and greater ease of implementation. An algorithm designed to differentiate unpitted from pitted cherries yielded recognition rates of 97.3% for the pitted and 94% for the unpitted cherries, with a total error rate of 3.5%. When the algorithm was adjusted to maximize removal of pitted fruit, 100% of pitted cherries were detected with a total error rate of 8.5 percent. If orientation could be controlled after pitting, total error could be as low as 1%.

Last Modified: 7/22/2014
Footer Content Back to Top of Page