DEHAZING RESEARCH ON BRIGHTNESS EQUALIZATIONMODELOF DRONE IMAGE

Due to the rapid development of drone technology, aerial imagery of drones is increasingly used in various fields. However, the aerial image of the drone is highly susceptible to weather conditions during the imaging process. Most aerial images are inevitably affected by fog when they are acquired. Due to the scattering and absorption of the atmosphere, the aerial image of the drone in foggy days has the characteristics of low contrast and unclear scenery. Due to the scattering and absorption of the atmosphere, the aerial image of drone acquired in the foggy environment has the characteristics of low contrast and unclear scenery. The Defogging technology for aerial image of drone can obtain a large amount of useful information in a pictures with low information amount through a certain image processing method, and convert the image with low information amount into a useful image. Therefore, the image processing research carried out for such image degradation caused by natural phenomena has universal practical significance. Aiming at the problem that the aerial image of drone is often affected by haze and the image is blurred and the image quality is degraded, this paper proposes a new model for defogging aerial image of drone. The brightness equalization model is used to improve the degraded image with fog defects. The brightness equalization model obtains the brightness channel of the original image based on the HSI transform. The brightness equalization filter is used to dynamically adjust the brightness to the appropriate interval to achieve the purpose of defogging and then further optimizes the result image by using Gaussian blur and color reshaping. Two images with fog problems were compared, using the brightness equalization model of this paper. And the quality evaluation parameters are selected to evaluate the processing results of the dehazing model. The average value of the images processed by the model is more suitable and the main quality evaluation parameters such as standard deviation and entropy are better than those of the original image.The experimental results show that the brightness equalization model can effectively remove the influence of fog in the aerial image of the drone and improve the visual effect of the image.


INTRODUCTION
In recent years, the emerging drones remote sensing technology has made up for the lack of flexibility and adaptability of satellite remote sensing and aerial remote sensing, and has become a powerful complement to the latter two [1] , playing an important role in resource management and emergency rescue. However, the quality of aerial imagery of drones fluctuates greatly in foggy environments [2] . In the foggy environment, atmospheric particles have scattering and refraction effects on light [3] , resulting in drone image contrast reduction, loss of detail, and lack of visual vividness [4] .The degradation of image quality has greatly restricted the validity and reliability of images [5] . Therefore, how to remove the influence of fog and improve the visual effect of images is a subject that has received extensive attention.
At present, the defogging processing methods for images are mainly divided into two categories [6] [7] : image restoration The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W10, 2020 International Conference on Geomatics in the Big Data Era (ICGBD), 15-17 November 2019, Guilin, Guangxi, China methods based on physical models and image enhancement methods. The image restoration methods mainly considers the root cause of the fog image degradation on the physical field, and analyzes the whole process of atmospheric scattering and its final influence on the image, and then inverts the process to obtain the color recovery image. For example, dark channel prior image defogging algorithm [8] [9] and so on. The main disadvantage of such methods is that additional parameters are needed to establish foggy image degradation models, such as depth of field information, etc. However, in practical applications, users generally cannot obtain these information.
The image enhancement method is to enhance the contrast and highlight the feature information to achieve the purpose of defogging. For example, homomorphic filtering image dehazing method [10] , Retinex theory [11] [12] and so on.
In order to remove the interference of fog and make the image clearer and more realistic, this paper based on HSI color transform [13] according to the spatial positional relationship of brightness. Using local adaptive filtering to reconstruct the brightness and nonlinear dynamic adjustment to achieve the purpose of defogging. After the HSI inverse transform, the Gaussian blur is used to preserve the image detail information and the color reshaping algorithm is used to maintain the vividness of the image color. It has been proved by experiments that the model of this paper can obviously remove the influence of foggy environment and maintain the visual effect of image detail texture, and the color of image is basically free from distortion.

HSI COLOR SPACE
In image processing, there are usually two kinds of color space: RGB color space composed of three primary colors of red, green, and blue, and the other is HSI color space.The HSI color space consists of three variables: hue (category of color), saturation (purity of color), and Intensity (lightness and darkness perceived by the human eye).
The HSI color space represents images in terms of hue, saturation and brightness [14]. Since the human eye is very sensitive to changes in brightness, the use of HSI color space is in line with human visual characteristics.

GAUSSIAN BLUR
Gaussian blur is an image processing algorithm that combines Gaussian distribution and convolution filtering, as shown in the following figure: (a) Before Gaussian Blur (b) After Gaussian Blur The closely spaced pixels in the image are more closely related. Therefore, it is reasonable to use a weighted average whose weight decreases with increasing distance: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W10, 2020 International Conference on Geomatics in the Big Data Era (ICGBD), 15-17 November 2019, Guilin, Guangxi, China This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-3-W10-1289-2020 | © Authors 2020. CC BY 4.0 License.

COLOR RESHAPING
The color reshaping process emphasizes the feature information of the dark area by re-adjusting the brightness value distribution of the image, and then repairs the color defects caused by the inverse HSI of the image.
The expression of the adjustment factor is as follows: where  = nonlinear strength of the adjustment factor ( , ) i L x y = Gaussian low-pass filter result.
The vividness of the color of the image after processing is improved. After the image is reshaped, the image is more natural and realistic under the human visual perception.

BRIGHTNESS EQUALIZATION MODEL
The model is mainly divided into three major steps. The   100; 255 The separated brightness map is subjected to local filtering 1, x T The distance measure function is the Manhattan distance formula: The second step, dynamic expansion, re-plans the brightness map and dynamically adjusts to the appropriate size [15] : The new brightness map and the original hue and In the third step, a Gaussian blur with a template size of

COMPARISON AND ANALYSIS OF EXPERIMENTS
In order to measure the dehazing effect of the model in this paper, the image quality evaluation criteria such as mean,  The   (2)In this paper, gaussian blurring processing and color reshaping processing are introduced in the brightness equalization model, which not only retains certain details but also enhances the fresh activity of color.
(3)The brightness equalization model of this paper has excellent defogging effect. And can improve the main parameters of the image, such as image brightness mean, standard deviation and entropy.The brightness equalization model of this paper has certain guiding significance and use value in the dehazing processing of images.