I. Overview of comprehensive treatment methods
The combined processing method of remote sensing images was first summarized by Cao Yu in the process of remote sensing geological application research in western Junggar Basin, Xinjiang, and then continuously supplemented and improved in the application practice of related oilfields in eastern China (Liaohe, Jiangsu and Henan), forming a variety of combined processing flows designed for different purposes.
The basic idea of the combined processing method is that the electromagnetic wave information received by the sensor is mainly on the surface and near the surface, while the information in the deep underground is weak. The purpose of using special digital processing function is to suppress the surface information to the maximum extent and highlight the underground geological information, so as to obtain a new image that can reflect the underground geological structure and structure.
The related processing flow is briefly described as follows.
(A) Linear expansion-histogram adjustment-local enhancement-color gradation coloring processing function combination
1. linear extension (PLIM)
By linearly stretching the histogram of the image, the brightness of the original image is extended to the clearest range of the naked eye, which is in line with the linear distribution range of photosensitive materials, so that the image with a single tone becomes clear. Linear expansion formula:
Selected papers on remote sensing of oil and gas geology in Yehefei
Where: Xa and XB- the minimum and maximum values of the image histogram before expansion;
Ya, Yb- the minimum and maximum of the extended image histogram;
X pixel value before expansion;
Y- Extended pixel value.
This function can transform the pixel value of an image as needed, and can provide piecewise linear transformation of up to 256 breakpoint pairs, and the transformation between two breakpoints is interpolation. The information of the extended image has been greatly enhanced. However, due to the great correlation of image bands, the image tones in some areas are very single. In this way, it is necessary to further adjust the expanded image to make the color of the image richer, which depends on the following processing functions.
2. Histogram adjustment (adjustment)
This function linearly transforms the image by translating and expanding (or compressing) the image histogram, so that the pixel brightness value of the new image has the average value and standard deviation determined by the user. Proper selection of input statistics (determined from the whole image or a predetermined color area) can maintain the variance of the expanded image and adjust the average value of each band, thus staggering the peak values of the histogram between image bands and making the image colorful.
3. Local enhancement
Divide the image into small enough parts, and calculate their variance (S) and mean (M) respectively. Then, based on their respective average values, the range of gray levels is adjusted to make the new gray values obey the normal distribution. The conversion formula is as follows:
Selected papers on remote sensing of oil and gas geology in Yehefei
Where: x is the pixel value of the original image;
Y—— pixel value of enhanced image;
S, m- variance and mean in the window with X as the center of the original image;
S', m '- variance and mean of user's expectation.
After local enhancement, the histogram of the image is full of 0 ~ 255 gray levels, thus enriching the color of the image, smoothing its variance, and effectively enhancing the weak information in the image, especially the hidden geological information. On the contrary, the information that was originally strong on the surface was suppressed.
4. Color gradation coloring (false)
This function divides the histogram distribution of the image into several levels according to the sensitivity of the naked eye to the color resolution, and the computer automatically gives the image color according to its hue, brightness and contrast. Through the transformation of this function, the related information of the image can be enhanced. When the correlation between image pixels is high, their colors also show almost the same color, thus achieving the purpose of highlighting the hidden geological information in the image.
In addition to the above functional combinations, there are many functional combinations for different purposes. It is:
(2) Linear geological structure combination treatment
Contrast enhancement-matrix transformation-trackball transformation;
B. contrast enhancement-template convolution.
(3) Comprehensive treatment of lithofacies zones
Contrast enhancement-matrix transformation-trackball transformation;
B contrast enhancement-histogram adjustment-local enhancement-gradation coloring.
(D) enhanced image information, TINT processing.
Linear expansion-histogram adjustment-local statistical enhancement.
These methods are similar to the function combination of "linear expansion-histogram adjustment-local enhancement-color gradation coloring" mentioned above, so they are not detailed here.
Secondly, the remote sensing data is selected by combining functional image processing.
Experience shows that there are obvious differences in the results of processing three commonly used image data: MSS, TM and SPOT, as far as the current I2S- 10 1 system is concerned. The former is ideal, which suppresses most ground information such as vegetation and water system, and generates continuous and clear new images of underground structures (plates ⅵ-2 and ⅹ-2). The latter two are not ideal, some or most of the ground information is not suppressed, and the generated image is still dominated by ground features (plate ⅳ- 1). It is speculated that this phenomenon is related to density and ground resolution (SPOT20m×20m > TM；; 30m×30m > MSS； 79m×79m), but not a combination of tapes. In other words, the more information per unit area, the higher the resolution, the more difficult it is to strip and extract the underground information, so more ground information is left. SPOT and TM belong to these three, and MSS is one of the three with less data per unit area and lower resolution, which is just suitable for the multi-functional combined processing method in design to generate an ideal image that can reflect the underground structure information. Of course, it doesn't mean that the less data and the lower the resolution, the more appropriate this function combination processing method is. For example, NOAA satellite data with resolution of 1000m× 1000m cannot be processed by multi-function combination method to obtain the above images. Therefore, the combination function processing method discussed in this paper only selects MSS image data, which is also related to the threshold required by I2S-1KLOC-0/image processing system and combination processing method.
Third, the characteristics of MSS combined images
The usual MSS standard false color image refers to the image synthesized by 7, 5 and 4 bands and red, green and blue computers. This image contains not only information in visible light band, but also some information in near red band. It can not only provide the data of macroscopic geological structure, but also provide geological information reflecting locality, which is an indispensable basic data for comprehensive interpretation and analysis. It has been pointed out that the characteristics of this kind of images are mainly image data reflecting surface geological information. This kind of image is synthesized by giving different colors according to different bands. A certain color on the image only represents the ground objects with similar electromagnetic spectrum in a certain interval, but does not represent the true color of the ground objects themselves, so it is called a false color image (plates ⅸ and ⅹ). In the process of interpretation using MSS false color images, people have different working experiences and different understandings of geological problems, and different people will come to different conclusions. At the same time, although the general remote sensing images also contain all kinds of weak geological information in the deep underground, they are ignored by interpreters because they are difficult to identify with the naked eye. Therefore, finding a new image that can effectively extract underground geological information has become the key to solve many geological problems.
Based on the standard false color image, the combined processing image uses the "deep filtering" method similar to geophysical data processing to enhance the weak part of underground geological information, and uses the combined processing function of "linear expansion-histogram adjustment-local enhancement-color gradation coloring" to greatly enhance the original surface information, thus highlighting the deep underground geological information.
The comparative analysis shows that there is a great difference between the synthetic image and the standard false color image. In addition to overcoming the above shortcomings, the surface information of rivers, farmland and vegetation, which was originally clearly reflected, was mostly suppressed and replaced by the deep information originally hidden in the surface information. The color bands, color rings and color blocks were composed of different "color levels" (see plates III-2 and X-2). These color rings and color blocks have different colors, and the contrast is obvious, and the interpretation accuracy is greatly improved. Comparing seismic and gravity and magnetic data, it is found that they have a close corresponding relationship in underground geological structure and structural reflection.
Fourthly, the influencing factors of MSS combined processing are analyzed.
More than ten years' practice has proved that combined image processing is very effective in solving geological problems. However, there are many factors that affect the effect of image processing and application. Such as terrain conditions, geological conditions and image processing level in different regions.
(a) the influence of landform conditions
It is pointed out that the combined image processing effect in outcrop area is poor. The reason is that outcrop rocks are exposed to the surface, which causes strong reflection and radiation of electromagnetic waves on the surface, thus completely suppressing the weak information in the deep underground. Unless the underground structural characteristics have obvious inheritance, it is difficult to show them through consolidated rock strata. In the coverage area, the reflectivity of various ground objects is relatively low, the surface information is relatively weak, and the underground information is relatively strong. Moreover, underground information can be reflected in the subtle changes of water system, vegetation and landform on the surface in a certain way. Therefore, the underground information in the combined processing image can be enhanced more effectively.
(B) the impact of geological conditions
There are many and complicated geological conditions that affect the processing and application of synthetic images. The sedimentary history and tectonic development history of an area, as well as the geological structure and structure determined by it, and so on.
Generally speaking, the more complex the structure and the deeper the buried depth, the more difficult it is to transmit underground information. Of course, the geological structure and structure of the underground can affect the deposition, diagenesis and structural evolution of the overlying strata to a certain extent, or it can be gradually reflected to the surface by inheritance and information superposition. Remote sensing geological interpretation of the coverage area is gradually deepened on the premise of "inheritance" to a great extent, at least at present.
Liaohe basin is located above luji fault block, and its fold basement is upper Proterozoic. There are thick Mesoproterozoic and Upper Proterozoic on the west side of the Tanlu fault zone, and most of the rest areas were uplifted in the early Paleozoic. Only part of the Upper Paleozoic (C, P) is scattered in Jinxi and Chaoyang areas, and the Tertiary is almost directly covered on the fold basement. Therefore, the Paleozoic and Mesozoic sediments in this area have been interrupted for a long time, and there is a big gap in the formation age of the strata (above and below the unconformity surface) that are in contact with each other. The geological structure and structure are relatively simple, and the combined image display effect is ideal. The ring and block images in the image are highly consistent with underground geological structures (uplift, bulge, depression, depression, fault block, etc.). However, due to the complex geological structure, there are both Precambrian metamorphic basement and Hercynian folded basement in Tabei area of Tarim Basin. Moreover, it has experienced many structural cycles such as Zhongtiaoshan-Jinning, Keping-Caledonian, Hercynian, Indosinian-Yanshan and Himalayan. Therefore, the interpretation effect of remote sensing geological interpretation, especially the deep geological structure related to oil-bearing strata, is not as obvious as that in Liaohe area. However, it is found that although the area is affected by multi-stage tectonic movements, many geological structures still have inherited characteristics, which are closely related to the surface micro-geomorphology. It is worth noting that all the known oil and gas fields in this area are related to micro-positive landform, that is, light color anomalies on remote sensing images (combination map and TM standard map). The structures such as Donghetang and girac explained by this method are basically consistent with the oil-bearing area preliminarily proved (Figure 1). Other oil and gas prediction areas (such as Quelingsi structure, etc.). ) subject to further drilling confirmation.
Map 1 geological interpretation map of MSS (156 ~ 32) in northern Tarim.
(3) the influence of human factors in the treatment process
In the process of combined processing on I2S- 10 1 system, some steps, such as color gradation coloring (false), are converted by tracking balls, so there are certain differences in the degree of mastering color standards between people, which can not but affect the processing results of images, which is also the influencing factor of geological interpretation using MSS combined map.
The influence of trackball on the processing effect is first manifested in the different colors of the same object in adjacent images. Secondly, it is also manifested in the connection of color levels to form patterns.
As long as the interpreters go through a period of running-in and exploration, the above problems are not difficult to solve.
Five, MSS combined image application analysis
MSS synthetic image is mainly used for geological interpretation and analysis of Quaternary coverage area. From the current review, its application effect is ideal. Because no matter Nanyang Oilfield, Zhongyuan (Dongpu) Oilfield, Subei Oilfield or Liaohe Oilfield, the favorable oil-bearing area prediction made according to the interpretation results of MSS synthetic images at that time has been verified by later exploration and obtained industrial oil flow in many areas. For example, Sanchunji and Guaying structures in the south of the remote sensing prediction area of Zhongyuan Oilfield encountered industrial flowing oil at 1990, and reported proven reserves of100000 tons, which became the main area for increasing reserves in the oilfield in that year. In addition, some hidden structures translated according to MSS combination map have also been confirmed by geophysical exploration. Let's briefly review the application results of remote sensing in oil fields.
(I) Liaohe Oilfield
During the period of 1989, many annular images related to oil and gas information were found in Liaohe basin and its periphery through geological interpretation of combined processing images. Later, drilling confirmed that industrial oil gas flow was found in the original remote sensing prediction area. For example: ① the ring structure of 17 and 18 to the west of Rongxing Reservoir (see figure 1 on page 62): the ring or rectangular light-colored image is abnormal, and the whole trend is near north and south, which is a relatively high place in the tidal flat. This area is a thermal anomaly area disturbed by water network, and the image is shaded, which is similar to known oil fields. The analysis shows that the annular image is a favorable oil-gas bearing area (Figure 3). 199 1 year obtained industrial oil and gas flow in this prediction area, and further exploration also found the overseas river structure, which controlled the overseas river oilfield. In addition, industrial oil and gas flow is also found in the annular image of Dawa (15) in the northeast of the structure and the annular image of Rongxing oil layer (22) in the south.
(2) Tarim basin
The geological interpretation of MSS composite processing film (156 ~ 32) in northern Tarim area has obtained many image features that cannot be recognized on MSS standard film (see chart X-2). For example, RR 123, RR 124, RR 125, RR 126 to the south of Hadadun (Figure 1) and so on. On the standard film, it appears as relatively uniform yellow sand, while on the composite image, it appears as green, blue patches or green tones on the yellow tone background. Another example is RR82 and RR90 in the northeast of Daheyan. On MSS standard film, their yellow-green tone is slightly darker, and the shape of sand dunes is somewhat different from its periphery, but the characteristics are not obvious. On the images jointly processed by MSS, they are irregular images marked with blue spots and yellow spots, corresponding to abnormal high gravity and magnetic force. Earthquakes are interpreted as underground igneous rocks.
Taking MSS joint processing image as the base map, the northern Tarim area is interpreted emphatically. The interpretation results show that many known oil and gas fields correspond to annular image characteristics on the image. Rudong Tanghe structure is a Carboniferous sandstone reservoir, with bright yellow and white light color anomalies on the image, dendritic appearance, cut by water system, and its northern edge is restricted by linear structure, and it strikes northwest (R47). The R59 image in the southeast of Luntai is abnormal, with striking white tone and regular oval shape. Well Lunnan 59 was drilled with industrial oil and gas flow, which was a Carboniferous uplift structure. Now girac Oilfield has been built. In addition, R43 and R50 are consistent with the Erbatai-Luntai buried hill structural belt, R44 is roughly equivalent to Yakela buried hill belt, and R49 is a famous structure 168. All these ring-shaped and block-shaped images, without exception, show light colors (light yellow, yellow and white) in the combined images, which are obviously different from the surrounding background colors.
According to the above-mentioned known relationship between oil and gas fields and annular images, we have made an analogy analysis of R55 (i.e. the annular image on Queling side), and think that it is an important target area for recent oil and gas exploration in northern Tarim. The reasons are as follows: ① the micro-positive landform of the surface corresponds to the characteristics of light-colored ring images on remote sensing images; ② The known oil fields in northern Tarim area have obvious correlation with the characteristics of remote sensing light-colored ring images, which are consistent with the characteristics of geochemical oil-gas adsorption (C 1-C4), ultraviolet fluorescence sealing effect, mercury radon edge effect and negative anomalies of thermal infrared images; ③ It is in good agreement with the underground structural trap delineated by earthquake. Compared with girac structure, they are all located in the NE-trending fault structural belt, all of which are Carboniferous uplifts, slightly positive landform, and surrounded by water systems. They all appear as huge light-colored rings in the image, which are located on the northern slope of the northern depression. Queling side structure is closer to the oil-generating sag, and has many similarities with other known oilfields in northern Tarim, so it is a promising oil-bearing prospect. But we still have to drill.
A large number of facts have proved that choosing an effective combination processing image in the coverage area can not only improve the geological interpretation effect of remote sensing information, but also directly predict the oil field by analyzing the characteristics of remote sensing information in known oil fields under similar geological background through analogy. If the above prediction methods are fixed in quantitative form, it will have far-reaching significance for future oil and gas exploration.