A Visual Feature Defog and Image Stabilization Detection System

Abstract

The invention relates to a visual characteristic dehazing and image stabilizing detection system, which includes the following modules and functions: (1) After the image data is input to the haze characteristic analysis and enhancement module, the enhanced image and correction parameters 1 are obtained through calculation output; (2) The enhanced image is input to the target detection system module based on multi-level parameter feedback control. This module calculates the position of the target to be detected through detection and recognition, and outputs the target offset and correction parameters 2; (3) The image stabilization offset calculation module will stabilize the image. The image offset is output to the image stabilization and retrace integrated control platform; (4) The image stabilization and retrace integrated control platform completes the image stabilization function according to the image stabilization offset, and adjusts the sensor integral according to correction parameter 1 and correction parameter 2 time. This system implements functions such as dehazing enhancement, photoelectric-mechanical integrated image stabilization, detection and tracking based on visual characteristic analysis. It has well achieved the goal of clear, stable and accurate detection systems, and can be embedded into different types of target detection systems. middle.

Description

technical field [0001] The invention relates to a visual characteristic defogging and image stabilization detection system, which belongs to the interdisciplinary field of automatic control and electronic information. Background technique [0002] Target tracking refers to the task of giving the position of the target in a certain frame of the video, and then the algorithm continuously calculates the target position in the subsequent frames. At present, one of the mainstream target tracking algorithms is based on the target detection method. The commonly used detection methods include inter-frame difference, background difference and motion segmentation. The inter-frame difference uses the change between frames to realize the detection, and the common method is two-frame or three-frame difference method. The background difference method uses the background image or reconstructs the background through the model, and obtains the moving target area by difference with the curre...

AI Technical Summary

Problems solved by technology

For the inter-frame difference method, when the displacement of the moving target is small, the inter-frame difference cannot segment the target

The inter-frame difference method cannot deal with the problem of simultaneous motion of the background, and cannot isolate the background from the target

The background difference method cannot deal with the background motion problem. When the background continues to move, the background difference method cannot successfully construct the background model, making the difference impossible.

The defect of the motion segmentation method is that the calculation of optical flow will contain more noise, and it is difficult to accurately distinguish the noise from the target when performing optical flow segmentation, resulting in a high false alarm rate of the algorithm

None of these methods have made special processing for the light transmission characteristics of haze, and the processed image has more or less distortion, which affects further enhancement processing.

Images