Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 12/22/2002
Publication Date: 4/1/2003
Citation: CHENG, X., TAO, Y., CHEN, Y.R. 2003. NIR/MIR DUAL-SENSOR MACHINE VISION SYSTEM FOR ON-LINE APPLE STEM-END/CALYX RECOGNITION. TRANSACTIONS OF THE ASAE. 46(2):551-558. Interpretive Summary: On-line apple inspection for defects, including contamination, is very inportant to the apple industry. The traditional visual apple-by-apple inspection is labor intensive and prone to human error. A machine vision system for automatic on-line defect inspection is needed to speed up the inspection procedure. Imaging techniques have been widely studied for on-line detection of fruit quality and safety because of their quick, noninvasive, and reliable inspection properties. One of the major obstacles in implementation of automatic apple defect detection is correctly identifying the apple stem-end/calyx. The possibility of misclassifying stem-ends/calyxes as defects is high and unacceptable. This research studied the feasibility of applying a dual-wavelength method and developing image-processing algorithms for on-line apple defect inspection. Algorithms were developed to expand the dual wavelength method along with the dual image registration and synthesis strategies so that the on-line defect identification accuracy can be improved. These research results are very important to the researchers who are developing on-line apple contamination and defect detection systems. Application of this technique to on-line detection systems will benefit the apple production and processing industry.
Technical Abstract: A near-infrared (NIR) and mid-infrared (MIR) dual-camera imaging approach for on-line apple stem-end/calyx detection is presented in this paper. How to distinguish the stem-end/calyx from true defects is a persistent problem in apple defect sorting systems. In a single NIR camera approach, the stem-end/calyx of an apple is usually confused with true defects and, consequently, it is often sorted incorrectly. In order to solve this problem, a dual-camera NIR/MIR imaging method was developed. Based on experimental results, the MIR camera can identify only the stem-end/calyx parts of the fruit, while the NIR can identify both the stem-end/calyx portions and the true defects on the apple. A fast algorithm has been developed to process the NIR and MIR images. The dual-camera imaging system has great potential for reliable on-line sorting of apples for defects. Test results of this method are also presented.