Study on detection method of lettuce leaf disease based on hyperspectral image

The invasion of plants by pathogens is a long-term process, during which the epidermal oil, cell wall and cellulose of leaves are changed, and the innate immune system of plants also causes the changes of disease-resistant enzymes, proteins, plant protection hormones and other molecular organic matter in the leaves. At present, the sensitivity of hyperspectral technology has reached the level of single molecule, so these changes at the micro level can affect the spectral characteristics of hyperspectral. At the same time, the disease infection has a strong stage, and the plant disease characteristics have significant changes in the invasion stage and the symbiotic stage when the pathogen becomes the host. Therefore, analyzing the influence of these disease characteristics on the spectral characteristics and finding the correlation between them is conducive to establishing the characterization relationship between the spectrum and the disease, so as to realize the accurate detection of various diseases and disease stages.

experimental materials

Lettuce, commonly known as leaf lettuce, is a genus of lettuce in the Compositae family, which was first introduced into China from the Mediterranean region of Europe. Because of its medicinal value and edible value, it has been rapidly and widely cultivated in China. Spring and autumn are the high incidence seasons of lettuce infection. Pathogens enter the damaged weak parts of leaves through soil and fertilization after overwintering, and gradually destroy the healthy tissues inside the leaves after establishing a parasitic relationship.

The lettuce samples used in this experiment were collected from a fruit and vegetable planting base. Leaves with a size of 8cm×12cm were selected for picking, and the picked diseased leaves were classified according to disease category and disease cycle. The following picture shows some experimental samples of lettuce collected.

experimental equipment

Spectrograph is an imaging spectrometer with transmission grating as the spectral component. By attaching CMOS camera to the front of the imaging spectrometer, the image and continuous spectral information of the object can be obtained through spatial scanning. The Color Spectrum FigSpec® series imaging hyperspectral camera uses a transmission grating spectral module with high diffraction efficiency and a high sensitivity surface array camera, combined with built-in scanning imaging and auxiliary camera technology, to solve the traditional hyperspectral camera requires an external push-scan imaging mechanism and complex focusing and other difficult problems. It can be directly integrated with the standard C interface imaging lens or microscope to achieve fast acquisition of spectral images.

Microstructure of lettuce under disease stress

The epidermal tissue of lettuce leaves mainly consists of stomata and epidermal cells. The spores of infected viruses enter the internal cell tissue of leaves mainly through stomata. With the evolution of the degree of invasion, the resistance of plants to disease leads to corresponding changes in the cell wall (the cell wall is thickened by precipitation such as lignin to resist virus invasion). Therefore, the damage degree of stomata on the leaf surface and the changes in the cell wall have a strong correlation with the development of diseases. Two characteristic indicators of stomata and cell wall were mainly studied in the experiment. The following figure shows the microstructures of epidermal cells and bubbles of healthy, early and middle anthrax lettuce leaves respectively.

Analysis of disease detection mechanism of lettuce leaves

When the lettuce leaves are irradiated by the external light source, the groups and macromolecules inside them absorb rays of certain frequencies, and the net change of dipole moment is caused by vibration or rotational motion. The molecules transition from the ground state to the excited state driven by vibration energy. However, the excited state is an unstable state, and eventually the molecules will return to the ground state, and the absorbed energy will be released in the regression process. The energy of green leaf plants is partly released by heat and light, and partly dissipated by other forms of energy, so the energy of reflected light is weaker than that of excitation light, and the wavelength is longer than that of excitation light. Changes in the microstructure of the internal tissue of the leaves directly lead to changes in the structure of the molecular energy bond, which affects the absorption characteristics of the irradiated light. The influence on the spectral characteristics can be described by the spectral curve. The following figure is the spectral intensity curve of leaf reflection obtained after hyperspectral image acquisition experiments on healthy leaves, early and middle leaves of anthrax disease by hyperspectral image technology.

It can be seen from the above figure that the spectral reflection intensity curves of each disease stage are morphological consistency, which indicates that although the defense system leads to the increase of disease-resistant organic matter, these organic matter does not change the structure of group molecular bonds in the leaves