A method for counting light-emitting diodes using local extremum clustering

Abstract

The invention belongs to the technical field of machine vision, and particularly relates to a method of using local extremum clustering to count light emitting diodes. The method includes the steps: subjecting LED chip images shot by an industrial digital camera to mean value filtering and gauss filtering to remove image noise; solving a local minimum value point to determine an approximate area of light emitting diodes according to grayscale difference between a light emitting diode area and background color; setting up a data sector for each local extreme points and surrounding gray values thereof, and performing k-means mean value clustering according to Euler distance of the vectors to extract a local extreme cluster of the light emitting diodes; and finally, as for monopole light emitting diode chips, defining the number of light emitting diode local extreme points as the number of the light emitting diodes, and as for rectangular or parallelogram double-pole light emitting diode chips, determining a light emitting diode rule according to two pole matching to determine the number of the light emitting diodes. According to tests, the method is accurate in counting, low in time consumption and high in robustness.

Description

technical field [0001] The invention belongs to the technical field of machine vision, and in particular relates to a method for counting light-emitting diodes. Background technique [0002] Light-emitting diodes (LEDs) are widely used in indication, display and other fields due to their advantages of long life and low energy consumption. The breakthrough in white LED technology and the continuous improvement of the luminous intensity of a single LED make it possible for LED to be used in the field of lighting [1]. In the finished product testing process of LED chips (collectively called square chips in the market at present), the problem of measuring the number of LEDs is involved. However, due to the slow counting, low precision, and high labor intensity of manual counting, and the high cost of sensor detection and counting equipment, it is not easy to maintain. This kind of LED counting method is a difficult problem that the industry needs to solve urgently. [0003] T...

AI Technical Summary

Problems solved by technology

Although the existing image LED counting method has achieved certain results, the recognition effect of the LED square slice top view taken by an industrial digital camera is unstable, and there are large connected areas or deformed areas in the densely stacked LED images, which is the key to the identification and counting of LEDs. It brings difficulties, which leads to the low efficiency of the current image LED counting method algorithm, the processing time is too long, and the processing of poor quality images is not ideal. In addition, in practical applications, it is very dependent on the light source, and the algorithm is greatly affected by the light. Therefore, there is an urgent need for an LED counting system with fast response speed, accurate counting and high robustness.

Images