| Summary: | 碩士 === 逢甲大學 === 自動控制工程學系 === 107 === In this study an innovative ensemble learning method in dynamic imaging system is presented. It is based on the cascade classifier, further designs the feature comparison classifier, and then uses the machine vision correlation algorithm to analyze the target type information. First, the system collects the target image and background to establish the samples database, and then uses the Local Binary Patterns (LBP) feature extraction algorithm to extract the feature values for classification. When the first stage classification is completed, the classification results are target features, and edge feature comparisons. The feasibility of this system is tested in two environments.
The first test environment is the crack detection of the retaining wall in the climbing area or the mountain road. At this time, the crack is the target to be tested, and the retaining wall is patrolled through the drone flight path setting, and then the image is instantly returned by using the wireless transmission of the system. The innovative ensemble learning classifier is used to analyze the image and determine the location of the crack for risk assessment.
Another test environment is the detection of vegetable worms on the leaf, using the innovative ensemble learning classifier to find the location of the worm. The system monitors the images of the cabbage leaf of the cabbage, find the worm, records its position, and analyzes the physiological dynamics of the worm, In order to understand the effect of different residual pesticides on the vegetable leaf surface for the vegetable worm.
|
|---|
