Abstract: This study aims to provide general technicians who manage pests in production with a convenient way to recognize insects. A novel method to classify insects by analyzing color histogram and dual tree complex wavelet transform (DTCWT) of wing images was developed. The wing image of lepidopteran insect is preprocessed to get the region of interest (ROI). First, the color image is converted from red-green-blue (RGB) to hue-saturation-value (HSV) space, and the 1D color histogram of ROI is generated from hue and saturation distribution, respectively. Then, the color image is converted to grayscale image, and rotated and translated to a standard position to extract the DTCWT features. Matching is first undergone by computing the correlation of the histogram vector between testing and template images. If the correlation is higher than a threshold, then their DTCWT features are further matched. The DTCWT feature matching are realized by computing their Canberra distance, and the nearest neighbour is considered as the most matched species. The method was tested at the insect database with images of 100 lepidopteran species, the recognition rate was as high as 76%, and the recognition rate for the subset of forewing images was as high as 92%. An ideal time performance was also achieved. The test results proved the efficiency of the proposed method.

Key words: Insects, Lepidoptera, image identification, image processing, color histogram, dual tree complex wavelet transform (DTCWT)