Multispectral Characterization of Pesticide Residues in Strawberries

The current article presents the implementation of a methodological approach to collect and process multispectral images, addressing all essential aspects to obtain high-quality data for identifying pesticides on strawberry surfaces. In the first step, various types of pesticides were compared to select the most suitable one for deliberately contaminating the study’s strawberries. Seeking the best way to apply the pesticide for controlled contamination, pesticide parts per million (ppm) within international limits were evaluated. Subsequently, information was gathered on variables such as lighting and focal distance that could affect the quality of multispectral images used to detect chemical residues on strawberries. To ensure precise data capture, a controlled environment was established considering object distance, lighting source, and background contrast. The MicaSense RedEdge camera captured multispectral images and recorded data across multiple spectral bands of contaminated strawberries. These images underwent preprocessing processes like segmentation and size reduction. Techniques for combining and manipulating spectral bands were also applied to generate vegetation indices such as NDVI and CGI. The findings demonstrate that the proposed methodology allows for the acquisition of multispectral images with comprehensive information in different spectral bands, enabling the generation of combinations through the use of vegetation indices, thus allowing the effectiveness of the suggested method to be validated in the future.