Yellow spot silkworm cocoon identification method combining deep learning and image processing

Abstract

The invention discloses a yellow-spot silkworm cocoon identification method combining deep learning and image processing. The method comprises the following steps: establishing a silkworm cocoon picture data set containing upper cocoons and yellow-spot cocoons; the SE-ResNet network is trained, and a to-be-tested picture is sent to the trained network for prediction; directly outputting the network identification result if the confidence coefficient of the network identification result is greater than or equal to 75%, and carrying out image processing secondary identification if the confidence coefficient is less than 75%; converting the image into an HSV format, carrying out image channel classification, carrying out threshold segmentation on a silkworm cocoon region on an S (saturation) single-channel image, counting the area of a macular region, and counting the average pixel value of the macular region on the S-channel image, namely the color saturation of the macular; and setting a yellow spot area double threshold and a yellow spot color saturation threshold, firstly performing area double threshold judgment, then performing yellow saturation threshold judgment, and finally outputting an image processing secondary identification result. The yellow-spot silkworm cocoon identification method combining deep learning and image processing is used for identifying the yellow-spot silkworm cocoon, can save labor cost, and is a quantitative and qualitative standardized detection method which does not transfer experience of a detector.

Description

technical field [0001] The invention relates to the fields of deep learning and digital image processing, in particular to a macular cocoon recognition method combined with deep learning and image processing. Background technique [0002] The silk industry in my country has always been a very fast-growing industry with a very large demand, and the market has a great demand for silkworm cocoons, which are raw materials for silk products; therefore, the automatic detection and identification of silkworm cocoon quality affects the production efficiency and production efficiency of the entire silk product industry. product quality. The silk reeling industry calls silkworm cocoons with qualified quality as upper car cocoons. During the cocooning process of silkworm breeding, due to the contamination of silkworm cocoons by silkworm excrement, yellow spots will be formed on clean and white silkworm cocoons. When the area of ​​yellow spots is too large or yellow If the cocoon is too...

AI Technical Summary

Problems solved by technology

[0003] The identification method of manual observation is inefficient, time-consuming, and greatly affected by personal experience, which is not conducive to the unified standard qualitative and quantitative cocoon quality inspection in the industry

Images