Method for identifying number and size of wood veneer knots based on image processing

Abstract

A method for identifying the number and size of wood veneer knots based on image processing comprises the steps that (a) images are collected; (b) the images are converted to gray images; (c) a gray valve corresponding to the one-dimensional maximum entropy is obtained according to an entropy function definition formula and is set to be a threshold value; (d) the images are subjected to binarization processing according to the threshold value, only the black color and white color are displayed, N primarily-selected knots are separated from a background, and N is larger than or equal to zero; (e) a morphological closed operation is used for eliminating insolated points outside the knots and peripheral interference features; (f) all primarily-selected knots are processed respectively, wherein the step (f) further comprises the steps that (1) contour points are searched for and judged if having communication areas, if not, the contours are filtered out, and if yes, a fitting ellipse is obtained and the next step can be executed; (2) the lengths of a major axis and a minor axis of the fitting ellipse and the ratio of the lengths are calculated, if the ratio is smaller than 0.5 or the ratio is larger than 2.0, the contour point is filtered out, and if the ratio is larger than 0.5 and smaller than 2.0, the next step is executed; (3) interference factors are filtered out according to the matching degree of all points of the communication area and the obtained fitting ellipse; (g) the number of the knots is counted, and the area of the largest knob is calculated.

Description

technical field [0001] The invention relates to an image processing-based recognition method for the number and size of wood veneer knots, belonging to the field of wood detection science and technology. Background technique [0002] In the grading of wood veneers, knots are an important evaluation index. The number of knots on the board surface and the area of ​​the largest knot determine its grade to a certain extent. In the past, the knot identification of wood veneer mainly relied on the naked eye observation method, and the efficiency was low. [0003] In the patent application "Wood Knot Detection Method, Device and Program (Patent Publication No.: CN101180536)", it is proposed to set the part of the captured image cut according to the specified threshold as the knot candidate, and the larger circularity as the knot detection standard, but failed to propose a circularity consideration method and the removal of other interference features, and the accuracy needs to be ...

AI Technical Summary

Problems solved by technology

[0003] In the patent application "Wood Knot Detection Method, Device and Program (Patent Publication No.: CN101180536)", it is proposed to set the part of the captured image cut according to the specified threshold as the knot candidate, and the larger circularity as the knot detection standard, but failed to propose a circularity consideration method and the removal of other interference features, and the accuracy needs to be improved; the article "Research on the recognition method of solid wood floor knots with the fusion of gradient operator and gray threshold value" ("Forestry Science and Technology" Issue 01, 2012), proposed the segmentation, extraction and recognition of knots on the surface of solid wood flooring, but failed to filter out other influencing factors such as the shape of knots, and may detect non-knot features, and the accuracy needs to be improved

Images