Multi dimension energy detection method and apparatus for fruit shape

Abstract

The invention discloses a fruit-form multiple-dimensioned energy measuring method and device, which will get 256 edge points by interpolating cubic spline to the fruit edge curve, compute the normalization distances from the points to the center and ascertain the corresponding random radius function, then compute the maximum value of the F(k)'s mathematical expectation and assure the origin of the curve. Using coiflets quadrature filtering device to process the F(k) and get the energy distribution vector E={Ed(1), Ed(2), Ed(3), Ed(4), Ed(5), Ed(6), Ea}, detecting the vector E's input sheet, implicit sheet and the nerval net of the output sheet whose number is 7, 8, and 3. The measuring device comprises an approach switch, two carrier wheels, chain conveying appliances, a frenquency divider, an illuminating bag, a camera, an image collecting card, a computer and a fruit-form detecting software.

Description

technical field [0001] The invention relates to a multi-scale energy detection method and device for fruit shape. Background technique [0002] Fruit shape is an important indicator for fruit automatic grading. The description of fruit shape is the basis for fruit shape classification. A good shape descriptor should have the following three properties: [0003] ① Invariance: For two identical shapes, how to zoom in and out, how to translate and rotate in the field of view should have the same description; [0004] ② Uniqueness: For two different shapes, there should be different descriptions; [0005] ③Stability: A small change in the same shape should only correspond to a small change in the description. [0006] The current fruit shape description methods can be divided into two categories: [0007] ① The fruit shape description method based on the aspect ratio: the fruit shape is described by the ratio of the fruit longitudinal diameter to the transverse diameter. This...

AI Technical Summary

Problems solved by technology

[0010] The purpose of the present invention is to provide a multi-scale energy detection method and device for fruit shape, which can describe the fruit edge curve in the form of multi-scale energy distribution, satisfy the three properties of the shape, and solve the limitations of the traditional starting point selection method. Resampling reduces the influence of noise, and the multi-scale method is used to describe the local features of the shape well, which improves the accuracy of shape classification

Images