Image Enhancement Based Fusion Feature Target Detection Method

In order to effectively solve the problems of image texture blurring and contrast degradation caused by the influence of fog, raindrops and other environmental factors on the images of transmission line components acquired by aerial photography in the natural environment, this paper proposes a fused feature target detection method based on image enhancement. Firstly, image enhancement is performed on the original aerial transmission component image based on dark channel defogging and Retinex image enhancement algorithm, then, image features are extracted based on HOG method, secondly, feature fusion is performed based on spatial attention mechanism, and finally, classification and regression tasks are accomplished based on SVM. This paper analyzes the gold fixture as a typical case of transmission line components, and constructs seven types of gold fixture datasets for experiments, and through the experimental results, it can be concluded that the overall detection effect of the target detection method proposed in this paper is better than that of the traditional detection model, and it can effectively solve the problem of the decline in the detection rate of the target due to the influence of the environment on the aerial images of the transmission components.