Hyperspectral image detection method for mutton quality indexes

The invention discloses a hyperspectral image detection method for mutton quality indexes. The hyperspectral image detection method comprises the following steps: S1, preparing a chilled sample; s2, acquiring sample multi-dimensional data; s3, spectrum extraction and separation; s4, establishing and verifying a prediction model; s5, processing and analyzing spectral data; and S6, outputting a detection result. Through spectrum extraction and separation, the effective wavelength can be separated from the invalid wavelength, so that the redundancy of spectrum information can be reduced, the processing speed of spectrum data is improved, and the parameters of spectrum characteristics can be set by utilizing spectrum data processing and analysis; mutton quality hyperspectral imaging is fused with indexes such as moisture content, color, TVB-N content and texture, so that the detailed degree of mutton quality division is improved, samples can be quickly classified after spectral data features are extracted, abnorm