ES2537783A1 - Procedure for obtaining a remote sensed image from a photograph (Machine-translation by Google Translate, not legally binding) - Google Patents

Abstract

The invention consists of a procedure to obtain a remote-sensed image from a photograph, framed mainly in the technical sector of agriculture in its broadest sense, although it can also be encompassed in the chemical and military sectors. The invention brings as a novelty a procedure, based on the treatment of images, thanks to which we can transform a photographic image into a remote sensed image. Among the uses of the invention are the decision-making, in any agricultural and/or forestry areas, when installing drainage networks, treatments against weeds, treatments against pests, fertilizing a crop or providing nutrients, as it allows to know if a river or reservoir is contaminated or not, as well as to know the dynamics of the pollutant in the water. (Machine-translation by Google Translate, not legally binding)

Description

DESCRIPTION

Procedure for obtaining a remote sensed image from photography.

Object of the invention 5

The object of the invention is a new method for obtaining a remote sensed image from a photograph, framed mainly in the technical sector of agriculture in its broadest sense, although it can also be encompassed in the chemical and military sectors. The invention provides as a novelty a procedure, based on image processing, thanks to which we can transform a photographic image into a remote sensed image. Among the uses of the invention are both decision making, in any agricultural and / or forestry areas, when installing drainage networks, weed treatments, pest treatments, fertilizing a crop or providing nutrients, as allowed knowing if a river or reservoir is contaminated or not, as well as knowing the dynamics of the contaminant in the water.

The invention is mainly framed in the technical sector of agriculture in its broadest sense, although it can also be encompassed in the chemical and military sectors. With regard to the first, we can include it within the agricultural production 20 activity, while the second focuses on the dynamics of pollutants in all types of water courses. The third is contained in the activity sector related to defense.

State of the art 25

According to Ramírez Juidías, E. et al (2009), the analysis of satellite images has traditionally been divided into two distinct phases, one visual and one digital. In this regard, a raw satellite image has a dull aspect, so that the visual analysis is not easy, which if we add the acquisition cost, results in the set of techniques developed to highlight certain aspects of said image with the purpose of facilitating its analysis, sometimes it is difficult to use when you want to use it in the development of low or very low cost projects.

Although the electromagnetic spectrum encompasses a large number of regions and the human eye has a high discrimination capacity of these colors, we can break down any color into three components, that is, blue, red and green, which coincidentally correspond to both regions of the visible.

Currently, the image display devices combine different levels 40 of intensity of red, green and blue in order to form their images. Each pixel on the screen is represented by three points of light, called RGB composition. Similarly, a satellite image has several bands, some of which correspond to regions of the visible and others with regions outside the visible spectrum.

 Four. Five

The sensors used in remote sensing are calibrated to receive very high radiation values without becoming saturated, so it is normal for all received values to be well below the maximum possible. This results in the images being very little contracted, being possible to solve it by linear expansion, restricted linear expansion or, an equalization of the histogram. fifty

When it is interesting to detect some specific aspect of the earth's surface, it is convenient to use indices that use some of the bands, the most commonly used being the vegetation index, the tasseled cap transformation and, the analysis of main components, each of which suppose A transformation of the bands. 5

According to Chivieco Salinero, E. (2008) on page 59 of his book, the growing development that spectral resolution sensors are having in recent years has stimulated the creation of spectral libraries. These are collections of spectral signatures, taken with laboratory radiometers under controlled conditions, which serve as a reference to know the type behavior of a given cover. These libraries are being used to calibrate the reflectivity values estimated from the image, facilitate automatic training in the interpretation of covers and serve as input for the extraction of spectral mixture models.

 fifteen

As far as the existing patents are concerned, and in relation to remote sensing, it is proposed that in the ES2303476 patent a procedure for the automatic obtaining of agronomic and environmental indicators of tree plantations by remote sensing was proposed. For its part in WO2006024686, a procedure was proposed for the discrimination and quantification of olive groves with vegetal covers by means of remote sensing with high spatial resolution sensors.

Patent ES2122935 proposes a procedure for detecting and monitoring satellite fires, while patent ES2311322 formulates another for discrimination and mapping of grass weed stands in 25 cereal crops by remote sensing.

Patent document ES2297612 performs a refinement of the spatial resolution of multispectral remote sensing data, while ES2332567 postulates an automatic procedure for sectioning remote images and characterizing agronomic and environmental indicators therein.

With regard to international patents we have that US7068816 consists of a method to obtain agricultural information from remote sensed images obtained directly from a sensor or platform, but which are not obtained after photographing the area in question. The patent WO2008133790 postulates a method to obtain information, also from images already remote sensed directly thanks to space sensors or platforms, georeferenced of the image, such as the health of the vegetation.

 40

The patent document US8135178, which could be thought to be related to the new procedure that is intended to be patented, has nothing to do with it, since it is a method of processing and processing images thanks to which it is intended to find certain patterns of development in the field of cultivation in order to focus on the possible intervention to it, but does not perform any treatment to said images necessary to obtain an image remote sensed from color photography. Similarly we can talk about the patent EP1791090, only in this case it is based on GIS.

Other international patents that directly use images already remote sensed, or 50 filtered, or with near-infrared incorporated to obtain different results are

WO9919824, CN101839979, WO2009086158, JP2006085517, US6567537, WO0008590 and WO2008051207, however none of them performs any treatment to said images necessary to obtain a remote sensed image from color photography, which lacks near infrared.

 5

Finally, there is still reference to patent JP2002117402, which performs a digital treatment of the image based on a totally different procedure than the one proposed in this new procedure.

As can be seen in the documentation cited, so far no procedure has been proposed capable of obtaining remote sensed images from color photography (from a normal and current camera, without the use of any type of filter incorporated into the camera itself, and with the absence of near infrared}, so that the procedure proposed is completely new in that regard.

 fifteen

Bibliography:

Anderson John (05/20/2003). US6567537.

Asner Gregory, P. (05/02/2008). WO2008051207. twenty

Borg, E., Fichtelmann, B., Richter, R., Bachmann, M. (05/01/2008). ES2297612.

Chuvieco Salinero, E. (2008). Environmental remote sensing: the observation of the earth from space. Editorial Ariel S.A. 25

Clare John, J., Russell David, P., Wolfe Chistopher, W. (06/11/2008). WO2008133790.

Dickson Monte, A., Hendrickson Larry, L. (04/22/1999). WO9919824.

 30

Dorado Gutiérrez, JM., Martín Rico, C., Gonzalo de Grado, J. (12/16/1998). ES2122935.

García Torres, L., Peña Barragán, JM., López Granados, F., Jurado Expósito, M. (08/01/2008). ES2303476.

 35

García Torres, L., Peña Barragán, JM., Caballero Novelda, JJ., Jury Expósito,

M., López Granados, F. (02/08/2010). ES2332567.

Hendrickson Larry, L., Clark Rendell, B. (03/13/2012). US8135178. 40

Hendrickson Larry, L., Dickson Monte, A., Han Shufeng. (02/17/2000). WO0008590.

Kelle Olavi, Macom Eric, Pliszka Robert, Mathawan Neeraj. (07/09/2009). WO2009086158. Four. Five

Knoblauch Gregory, E., Dinville Andrew, P., Paris Jack, F. (06/27/2006). US7068816.

López Granados, F., García Torres, L., García Pulido, J., Peña Barragán, JM., Jury Expósito, M., Jarcía-Ferrer Porras, A., Sánchez de la Orden, M. (09/03 / 2006). 50 WO2006024686.

López Granados, F., García Torres, L., Peña Barragán, JM., Jurado Expósito, M. (02/01/2009). ES2311322.

Nakayama Osamu (04/19/2002). JP2002117402.

 5

Park Gwang, W., Kwon Young, H., Choi Kyoung, Oh Seung, H., Shin Sang, H., Kang Dong, J. (11/23/2006). EP1791090.

Ramfrez Juidías, E .; Infante Perea, MMª .; Esteve Pardal, R. (2009). Remote sensing and photogrammetry. Publications Service of the University of Seville. 10

Xiaolei Deng, Minzan Li, Xiuhua Li, Hong Sun, Yane Zhang, Ze Zhang, Ruijiao Zhao, Lihua Zheng. (09/22/2010). CN101839979.

Yumiba Norio. (03/30/2006). JP2006085517. fifteen

Description of the invention

With the procedure proposed, it is possible to obtain a remote sensed image from a simple photograph taken with any camera, without the need for it to be coupled with any type of external or internal device capable of capturing the radiation emitted or reflected by each body, object or fluid. To do this and, after taking the necessary photograph (s) in color, an image treatment is carried out in order to pass it to panchromatic, but correcting the hue, saturation and brightness of the image, resulting in an image apparently distorted where it is impossible to differentiate and identify the objects in the photograph. Subsequently, and on the distorted image, a final digital treatment is performed consisting of identifying the exact coloration of each pixel on the panchromatic scale, which is related to the emitted or reflected radiation (as it is) by the object existing in the photograph , resulting in an image in which each pixel 30 identifies, with a different color, each part of the object (s) in the photograph as a spectral signature. Finally, and in order to know what color each object corresponds to, it is necessary to calibrate the image by means of a field visit whenever possible, since this will allow having a reference for future research or data collection in that area. The advantages of this new procedure in relation to the prior art are several. In the first place it allows us to eliminate the dependence of those in charge of the direct obtaining of remote sensed images, which significantly reduces the cost of the investigations carried out in certain areas where it is difficult to have images for a more or less long period of time. In addition, it is a useful procedure for making decisions of a different nature, such as whether it is necessary to install drainage networks, weed treatments, pest treatments, fertilize a crop or provide nutrients, etc. Finally, comment on its usefulness in territorial planning at the provincial level and its versatility when it comes to detecting camouflaged objects under the vegetable canopy or by any other artificial means. Four. Five

The present procedure consists in obtaining remote sensed images from photographs taken by a simple and current camera.

For this it is necessary that the photograph taken is in color, since otherwise it is impossible to obtain the emissivity or reflectivity values (depending on the case) of each object.

Subsequently, we have to obtain a hyperspectral image (so we will have made 5 the first of the two digital treatments), with all the bands of the spectrum, for which color photography has to be transformed to panchromatic respecting a series of important points:

1.- The "average" and "fashion" values of the gray-scale histogram must be 10 similar and at least equal to 25.

2.- The resulting image must have 8 bits per pixel.

3.- In addition, the "magnification" parameter must be at least 0.17. fifteen

With these guidelines in mind, we obtain an image with all the bands of the spectrum in panchromatic format, distorted and with opacity equal to zero, in which it is practically impossible, depending on the objects photographed, to recognize any of them. twenty

In the second and last place, another digital treatment is performed consisting of identifying the exact coloration of each pixel on the panchromatic scale, which is related to the emitted or reflected radiation (as it is concerned) by the object existing in the photograph, giving as The result is an image in which each pixel identifies, with a different color, each part of the object (s) in the photograph as a spectral signature. For this purpose it is necessary to know:

1.- The "average" and "fashion" values of the RGB histogram must be at least equal for red, present a ratio of 1.2 for green and, show a ratio between 30 1.1 and 1.6 in favor of "fashion" for blue.

2.- The resulting image must have 32 bits per pixel.

3.- The "magnification" parameter must also be at least 0.17. 35

4.- It must be borne in mind that the resulting image must have a 3D Gaussian filter, since in this way the relationship between the object and its exact pixel color will be obtained as a spectral signature.

 40

Finally, remember that it is convenient to make a field visit in order to be able to calibrate the image (s) / is obtained for future research or work in the same area.

The present procedure has important applications in the industry for agriculture, livestock and forestry, since it allows to know the moment in which the cattle have sufficient quality natural foods, in order to ensure greater input savings in those areas with few resources. In the same way it allows to know if the crops are in perfect condition to reach the optimum of production or not, which minimizes deviations from the average, 50 resulting in a lower expense by the producer.

With respect to the chemical industry, or those of other fields that produce effluents with discharges into public channels, it is an application to know if the general level of discharge has been exceeded or, if on the contrary, is at the limit of being exceeded, giving time to take measures to minimize the impact on the environment that are appropriate to the dynamics of the contaminant in the water. 5

The application for the army is none other than to find out the movement of allied or enemy troops, as well as to know if there are camouflaged equipment and in what percentage.

For the public administration in general, as well as for those NGOs that need it, 10 is a valid, useful and applicable procedure not only in land management, but also in those areas affected by natural disasters.

Embodiment of the invention

 fifteen

As an example, comment that images corresponding to the herbaceous crop practices plot of ETSIA (X = 239,685.82 m; Y = 4,137,830.10 m; Zone S; Spindle 30) have been processed at the University of Seville (during part of the 2010/11 academic year) through a small UAV provided with a small normal camera and current specific for these devices. twenty

All images were treated according to the proposed procedure in order to obtain a scale that related the color of each pixel with each of the existing crops in that field of practice, as well as with the diseases with which they were inoculated. 25

The results obtained, showed not only the percentage of the surface of the farm that has the same crop or object, but also if the vegetation is stressed or not as a result of the inoculated disease. The comparison of these results with those obtained from SPOT satellite images were very positive. 30

Claims (5)

1. Procedure for obtaining a remote sensed image from photography, characterized by using any type of camera, and comprising the following phases or stages: - Obtaining the photograph of that area that is intended or needed to have a remote sensed image. - Carrying out a treatment of the image in order to make it panchromatic, 10 but correcting the hue, saturation and luminosity of the image, resulting in a seemingly distorted image where it is impossible to differentiate and identify existing objects in the photograph. - Subsequent realization on the distorted black and white image of a final digital treatment consisting in identifying the exact coloration of each pixel on the panchromatic scale, which is related to the emitted or reflected radiation (as it is concerned) by the object existing in the photograph, resulting in an image in which each pixel identifies, with a different color, each part of the object (s) in the photograph as a spectral signature, thanks to the application of a 3D Gaussian filter. - Calibration of the image obtained through field visit. 2. Procedure for obtaining a remote sensed image from a photograph, according to claim 1, characterized in that the photograph taken must necessarily be in color. 3. Procedure for obtaining a remote sensed image from photography, according to previous claims, characterized by having only two different digital treatments. 4. Procedure for obtaining a remote sensed image from photography, according to previous claims, characterized in that the photograph can be obtained from the ground by any person or, by means of normal cameras coupled to 35 small UAVs directed by radio control. 5. Procedure for obtaining a remote sensed image from photography, according to previous claims, characterized by using a 3D Gaussian filter.  40