Dead pine tree identification method and system based on high-resolution image

Abstract

The invention discloses a dead pine tree identification method and system based on a high-resolution image. The method comprises the steps of firstly, obtaining a pine forest visible light image and anear-infrared image to serve as original image data, and preprocessing the original image data to form to-be-classified image samples; secondly, performing ground object color feature extraction on verified disease and pest damaged tree wave spectrum information, classifying color features of a ground object according to different stage states of dead pine trees, and generating training samples and a plurality of parameter rules; thirdly, performing the ground object color feature extraction on all pixel points in the to-be-classified image samples by using the training samples and the plurality of the parameter rules; and finally, performing denoising screening classification on sample pixel points in the to-be-classified image samples subjected to the ground object color feature extraction, thereby generating a state classification result of the dead pine trees. Therefore, high-detection-precision identification of the dead pine trees with different disease areas and different disease degrees under the condition of complex terrains is achieved; and the processing speed is high and the stability is extremely high.

Description

technical field [0001] The invention relates to a method for identifying diseased and dead pine trees, in particular to a method and system for identifying diseased and dead pine trees based on high-resolution images. Background technique [0002] The spectral feature method is the most common method for identifying pine trees with pests and diseases. It is based on the fact that when the plant is infested by pests and diseases, the difference in physiological changes will be reflected in the spectral characteristics, especially the difference in spectral characteristics between the red region and the near-infrared region. Use imaging equipment such as spectrometers to obtain spectral data information of reflection and radiation, and on this basis, discover the changing rules of reflection spectra in different bands. [0003] Compared with the traditional manual field survey method, obtaining the spectral feature information of objects in remote sensing images to identify di...

AI Technical Summary

Problems solved by technology

[0005] The support vector data description multi-classification method is based on image pixels for classification and recognition. One of the disadvantages is that a large number of samples need to be trained, so the execution efficiency is slow; moreover, this method cannot realize the identification of multi-level diseased pine trees, such as early infection, mid-term Infection, late infection and dead pine trees; in addition, it has the disadvantages of high misjudgment rate and relatively low recognition accuracy when solving multi-classification recognition problems

Images