Potato External Defect Detection Method Based On Image Processing

At present, the domestic research on external defect detection based on static analysis mostly potatoes, unable to meet the needs of the actual production line testing. The higher cost of imported potato line testing equipment to our existing conditions can not be generalizable. Given the above, we design the potato dynamic image data acquisition system to study the use of image processing techniques to detect external defects potatoes. The main contents are as follows:1. A transfer unit designedDesign has a height difference of the two gear transmission with potato, potato achieve motion image capture; and key components of the mechanical analysis to verify the reliability of the device.2. Potatoes dynamic image acquisition system automaticallyPotato dynamic image acquisition system design, the system consists of single chip microcomputer, infrared sensors, CCD cameras, transmission and other components. Height difference between the two positions installed in a pair of belt drive camcorders, potato whereabouts during the single chip microcomputer control infrared sensor, CCD camera, the image is automatically collected.3. Image processing based on external potato defect detection methodFor defective part that is dark potato bites, decay, fissure, a larger cross-section of mechanical damage, dry rot and other proposed method interception fixed gray threshold dividing its defects and reduce the detection error combined with morphological method.For the Green Paper and sprouting potato tuber defects and normal skin color paper proposes the use of Euclidean distance threshold for defect detection color settings.For potato mechanical damage defects slender paper proposes the use of Hough transform parameters accumulator with mechanical damage to parts of the quick detection.All three detection methods, the potato has no defect detection accuracy rate of97.5percent, respectively,90%, and can basically meet the requirements of the actual test.4. Software design and testMatlab7.0use image processing software for potatoes, the realization of its external defect detection. In an experimental basis, to further optimize the mechanical transmission system, improve the accuracy of the potato defect detection.This design defect detection system potato, potato sets transmission, automatic image capture, image processing in one. Can be applied to the corresponding image feature detection methods for different potato defect detection accuracy basically meet online testing requirements.