CN103489168A - Enhancing method and system for infrared image being converted to pseudo color image in self-adaptive mode - Google Patents

Abstract

The invention provides an enhancing method and system for an infrared image being converted to a pseudo color image in a self-adaptive mode. The method comprises the steps that de-noising processing is conducted on the input infrared image according to median filtering; edge extraction is conducted on the infrared image according to Sobel edge detection; normalization processing is conducted on generated data with enhanced image details, and the data are converted to gray data; histogram statistic and mapping color code article bound calculation are conducted on the gray data; according to a histogram statistic result and color code bounds, color space conversion is conducted on the basis of a mapping rule. Compared with the prior art, the enhancing method and system for the infrared image being converted to the pseudo color image in the self-adaptive mode have the advantages that the purposes of enhancing the image details and reducing image noise are achieved, the dynamic range of the infrared image can be stretched in real time, an observed object is protruded at the same time when the background is compressed, image definition and the image contrast ratio are improved, and the high-quality pseudo color image is output. The image quality is ensured and the imaging effect is greatly improved when the dynamic range of the image is stretched and the image details are enhanced.

Description

A kind of infrared image self-adaptation turns pseudo-color Enhancement Method and system

Technical field

The present invention relates to the infrared imagery technique field, relate in particular to a kind of infrared image self-adaptation and turn pseudo-color Enhancement Method and system.

Background technology

At present, the infrared focal plane detector technology is increasingly mature, is widely used in the fields such as public security, fire-fighting, military affairs.The data of infrared focus plane collection form gradation data after Nonuniformity Correction, and gradation data need to, through pseudo-color conversion, just can become displayable image and be shown.

Due to infrared picture data, to have a spatial coherence strong, resolution is low, the characteristics that dynamic range is little, the imperfection of the random disturbance of external environment and thermal imaging system can bring the noise of Various Complex simultaneously, different conversion methods causes the observable image quality formed that larger difference is arranged, and different-effect also can directly affect user's experience.

Specifically, in realizing process of the present invention, the inventor finds that there is following shortcoming in existing scheme:

In prior art, conversion method commonly used, in expansion stretching image space, has also enlarged the image noise, variation that simultaneously can't fine adaptation scene, and imaging effect is not ideal enough.Need badly and will a kind ofly when stretching dynamic range of images, reinforcement image detail, can guarantee that the new self-adaptation of picture quality turns pcolor image intensifying scheme.

Summary of the invention

The object of the invention is to overcome the shortcoming and defect of prior art, provide a kind of infrared image self-adaptation to turn pseudo-color Enhancement Method and system.

A kind of infrared image self-adaptation turns pseudo-color Enhancement Method, and described method comprises:

According to medium filtering, the infrared image of inputting is carried out to denoising;

According to the Sobel rim detection, infrared image is carried out to edge extraction;

The data that strengthen with image detail process normalized by generating, be converted to gradation data;

Described gradation data is carried out to statistics with histogram and the calculating of mapping colour code bar bound;

According to described statistics with histogram result and colour code bound, according to mapping ruler, carry out the color space conversion.

Described according to medium filtering to the input infrared image carry out denoising, comprising:

Adopt the 5x5 medium filtering to carry out denoising to the infrared image of input;

Filtered result and original view data are carried out to addition, obtain new image.

Describedly according to the Sobel rim detection, infrared image is carried out to edge extraction, also comprises:

By edge detection results with the weighting scheme image that is added to.

Described described gradation data is carried out to statistics with histogram and mapping colour code bar bound is calculated, comprising:

Detect the histogram distribution data of scanned infrared image, find out the roughly distribution parameter of its gray scale: the average gray value avg of the minimum gradation value low of image, the maximum gradation value high of image, image;

The minimum gradation value low of image is mapped to " color_low "; The maximum gradation value high of image is mapped to " color_high "; The average gray value avg of image is mapped to " color_avg ";

According to current gray level mean value calculation avg_id, with min_level, compare, if average gray avg is less than threshold value min_level, increase the colour code lower limit; Otherwise the colour code lower limit is constant;

In like manner calculate the colour code upper limit.

Describedly according to mapping ruler, carry out the color space conversion, comprising:

The colour code bar that calculating makes new advances, traversal histogram horizontal ordinate, calculate the color that all gray-scale values are corresponding one by one;

Greyscale image transitions is become to coloured image.

In described Sobel rim detection, Sobel operator template adopts $\left[\begin{array}{ccc}-\mathrm{xf}& 0& \mathrm{xf}\\ -\mathrm{yf}& 0& \mathrm{yf}\\ -\mathrm{xf}& 0& \mathrm{xf}\end{array}\right]$ With $\left[\begin{array}{ccc}-\mathrm{xf}& -\mathrm{yf}& -\mathrm{xf}\\ 0& 0& 0\\ \mathrm{xf}& \mathrm{yf}& \mathrm{xf}\end{array}\right],$ Adjust according to demand rim detection; Described xf, yf are level, vertical direction weighting coefficient.

A kind of infrared image self-adaptation turns pseudo-color enhancing system, and described system comprises median filter unit, edge detection unit, normalization unit, colour code bar computing unit and map unit, wherein,

Described median filter unit, for carrying out denoising according to medium filtering to the infrared image of inputting;

Described edge detection unit, for carrying out edge extraction according to the Sobel rim detection to infrared image;

Described normalization unit, the data that strengthen with the image detail process normalized for generating, be converted to gradation data;

Described colour code bar computing unit, for carrying out statistics with histogram and the calculating of mapping colour code bar bound to described gradation data;

Described map unit, for according to described statistics with histogram result and colour code bound, carry out the color space conversion according to mapping ruler.

Described system also comprises the first superpositing unit, for the filtered result of described median filter unit and original view data are carried out to addition, obtains new image.

Described system also comprises the second superpositing unit, for the edge detection results by described edge detection unit with the weighting scheme image that is added to.

Described colour code bar computing unit comprises statistics with histogram subelement and colour code bar bound computation subunit, wherein,

Described statistics with histogram subelement, for carrying out statistics with histogram to described gradation data;

Described colour code bar bound computation subunit, for calculating mapping colour code bar bound.

The present invention, by having introduced medium filtering and Sobel edge algorithms, carries out pre-service to infrared image, by adjusting suitable weighting factor, can reach the strengthening image detail, reduces the purpose of picture noise.Adaptive color space conversion method, can the be real-time infrared image dynamic range be stretched, give prominence to object observing in the time of the compacting background, increased image definition, contrast, the pcolor picture of outputting high quality.Guarantee picture quality when stretching dynamic range of images, reinforcement image detail, made imaging effect that very large lifting be arranged.

The accompanying drawing explanation

The infrared image self-adaptation that Fig. 1 provides for the embodiment of the present invention 1 turns pseudo-color Enhancement Method principle flow chart;

The color space conversion schematic diagram that Fig. 2 provides for the embodiment of the present invention 1;

The infrared image self-adaptation that Fig. 3 provides for the embodiment of the present invention 2 turns the pseudo-color system architecture schematic diagram that strengthens;

Colour code bar computing unit 400 structural representations that Fig. 4 provides for the embodiment of the present invention 2.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in detail.But embodiments of the present invention are not limited to this.

The principle of each embodiment of the present invention is: by image is carried out to the intermediate value noise reduction filtering, strengthen image detail according to the Sobel algorithm, the Normalized Grey Level data, the statistic histogram data also calculate suitable colour code bar scope, according to mapping ruler, gradation data are changed into to the pcolor picture.Introduce medium filtering and Sobel edge algorithms, infrared image has been carried out to pre-service, by adjusting suitable weighting factor, can reach the strengthening image detail, reduced the purpose of picture noise.Adaptive color space conversion method, can the be real-time infrared image dynamic range be stretched, give prominence to object observing in the time of the compacting background, increased image definition, contrast, the pcolor picture of outputting high quality.

As shown in Figure 1, the infrared image self-adaptation provided for the embodiment of the present invention 1 turns pseudo-color Enhancement Method principle flow chart, specific as follows:

Step 10, carry out denoising according to medium filtering to the infrared image of inputting.

Adopt the 5x5 medium filtering to carry out denoising to the infrared image of input.Image denoising is the pith of infrared preprocessing, can produce material impact to subsequent treatment.Medium filtering can carry out smoothly image, can retain image detail simultaneously.Filtered result and original view data are carried out to addition, obtain new image.

Y _out＝Y _gray+Y _med*factor1；

Wherein, Y _outfor last filtering output, Y _grayfor original image, Y _medfor median-filtered result, factor1 is weighting factor.

Step 20, carry out edge extraction according to the Sobel rim detection to infrared image.

Adopt the Sobel edge detection method to carry out edge extraction to infrared image.Because infrared image itself is indifferent to the target detail embodiment, adopt edge detection algorithm can strengthen the detail section of infrared image.Edge detection results is added to weighting scheme in image, and making image retain raw information has increased again the edge details feature.Reach the purpose that details strengthens.

Y＝Y _out+Y _sobel+factor2；

Wherein, Y is output, Y _outfor formula 1 output, Y _sobelfor the Sobel testing result, factor2 is weighting factor.

Sobel operator template adopts $\left[\begin{array}{ccc}-\mathrm{xf}& 0& \mathrm{xf}\\ -\mathrm{yf}& 0& \mathrm{yf}\\ -\mathrm{xf}& 0& \mathrm{xf}\end{array}\right]$ With $\left[\begin{array}{ccc}-\mathrm{xf}& -\mathrm{yf}& -\mathrm{xf}\\ 0& 0& 0\\ \mathrm{xf}& \mathrm{yf}& \mathrm{xf}\end{array}\right],$ Adjust according to demand rim detection; Described xf, yf are level, vertical direction weighting coefficient.

Step 30, the data that strengthen with the image detail process normalized by generating, be converted to gradation data.

The data that strengthen with image detail process normalized by generating, be converted to 8 gradation datas.Can be undertaken by following formula.

Wherein, Y ' _{(x, y)}for coordinate x, y place pixel value, Y _maxfor image maximal value, Y _minfor the image minimum value.

Step 40, carry out statistics with histogram and the calculating of mapping colour code bar bound to gradation data.

8 gradation datas are carried out to statistics with histogram and the calculating of mapping colour code bar bound.Infrared gradation data need to be mapped in displayable color space by a certain suitable colour code bar.Its color space color is determined by color lower limit color_low and color upper limit color_high.

By detecting the histogram distribution data of scanned infrared image, find out the roughly distribution parameter of its gray scale: the average gray value avg of the minimum gradation value low of image, the maximum gradation value high of image, image, these parameters will be carried out the image space mapping for generating conversion colour code bar new and image correlation.

Order:

$\mathrm{color}\_\mathrm{avg}=\frac{\mathrm{color}\_\mathrm{low}+\mathrm{color}\_\mathrm{high}}{2};$

Basic mapping ruler is defined as follows:

The minimum gradation value low of image is mapped to " color_low ";

The maximum gradation value high of image is mapped to " color_high ";

The average gray value avg of image is mapped to " color_avg ";

Histogram smallest point 0 corresponds to the bottom thresholding;

Histogram maximum point 255 corresponds to the top thresholding.

For fear of infrared image in amplitude hour, directly completely spatial mappings arrives the colour code bar, can cause like this image too to stretch, thereby significantly amplified the problem of noise, when the colour code bar formed in order to image transitions, we introduce a threshold value min_level, stipulate that this value is for being mapped to the minimum value at colour code center.

Colour code bar " lower limit " calculation process of take is example:

Initialization arranges colour code bound color_low and color_high, and the threshold value min_level that can be mapped to the colour code center is set.

Go out avg_id according to the current gray level mean value calculation, compare with min_level, if average gray avg is less than threshold value min_level, need so to increase the colour code lower limit, reduce the scope that is mapped to colour code, formula is as follows:

$\mathrm{new}\_\mathrm{color}\_\mathrm{low}=\mathrm{color}\_\mathrm{low}+\frac{(\min \_\mathrm{level}-\mathrm{avg}\_\mathrm{id})*(\mathrm{color}\_\mathrm{avg}-\mathrm{color}\_\mathrm{low})}{\min \_\mathrm{level}};$

If avg_id be greater than min_level so the colour code lower limit without change.

new_color_low＝color_low。

In like manner can calculate the new upper limit new_color_high of colour code.

Step 50, according to statistics with histogram result and colour code bound, carry out the color space conversion according to mapping ruler.

According to statistics with histogram result and new colour code bound, according to mapping ruler, carry out the color space conversion.As shown in Figure 2.In Fig. 2, low_gap and high_gap are the spaces of shining upon without detail content in the regulation picture.

At first calculate the colour code bar made new advances, traversal histogram horizontal ordinate, calculate color corresponding to all gray scales one by one.Take the lower limit part mapping equally as example, suppose that current gray level is gray, need to calculate colour corresponding to current gray level value.Operate as follows.

As gray<low,

$\mathrm{ColorBar}\left[\mathrm{gray}\right]=\mathrm{OldColorBar}[\mathrm{new}\_\mathrm{color}\_\mathrm{low}+\frac{\mathrm{low}\_\mathrm{gap}*\mathrm{gray}}{\mathrm{low}}];$

Work as gray >=low,

$\left[\begin{array}{c}\mathrm{ColorBar}\left[\mathrm{gray}\right]=\\ \mathrm{OldColorBar}[\mathrm{new}\_\mathrm{color}\_\mathrm{low}+\mathrm{low}\_\mathrm{gap}+\frac{(\mathrm{gray}-\mathrm{low})*(\mathrm{color}\_\mathrm{avg}-\mathrm{new}\_\mathrm{color}\_\mathrm{low})}{\mathrm{avg}-\mathrm{low}}].\end{array}\right]$

In like manner can calculate the colour code color of gray under other tonal ranges.

After having obtained the mapping colour code bar relevant to picture, greyscale image transitions is become to coloured image.

Y _rgb＝ColorBar[Y _gray]；

Y wherein _gray8 gray level images, Y _rgbfor the pcolor picture.

In order further dynamically to adjust picture mapping effect, details and the content of outstanding user's real concern, suppress irrelevant interference content, and regulation can be offset adjustment to the low in figure, high and color_avg parameter.

As shown in Figure 3, the infrared image self-adaptation provided for the embodiment of the present invention 2 turns the pseudo-color system architecture schematic diagram that strengthens, this system comprises median filter unit 100, edge detection unit 200, normalization unit 300, colour code bar computing unit 400 and map unit 500, specific as follows:

Median filter unit 100, for carrying out denoising according to medium filtering to the infrared image of inputting;

Edge detection unit 200, for carrying out edge extraction according to the Sobel rim detection to infrared image;

Normalization unit 300, the data that strengthen with the image detail process normalized for generating, be converted to gradation data;

Colour code bar computing unit 400, for carrying out statistics with histogram and the calculating of mapping colour code bar bound to gradation data;

Map unit 500, for according to statistics with histogram result and colour code bound, carry out the color space conversion according to mapping ruler.

Further, this system also comprises the first superpositing unit 600, for the filtered result of median filter unit 100 and original view data are carried out to addition, obtains new image.

Further, this system also comprises the second superpositing unit 700, for the edge detection results by edge detection unit 200 with the weighting scheme image that is added to.

Further, as shown in Figure 4, above-mentioned colour code bar computing unit 400 comprises statistics with histogram subelement 401 and colour code bar bound computation subunit 402, specific as follows:

Statistics with histogram subelement 401, for carrying out statistics with histogram to gradation data;

Colour code bar bound computation subunit 402, for calculating mapping colour code bar bound.

It should be noted that: the infrared image self-adaptation that above-described embodiment provides turns pseudo-color enhancing system when the infrared image self-adaptation turns pseudo-color the enhancing, only the division with above-mentioned each functional module is illustrated, in practical application, can above-mentioned functions be distributed and completed by different functional modules as required, the inner structure of the system of being about to is divided into different functional modules, to complete all or part of function described above.In addition, the infrared image self-adaptation that above-described embodiment provides turns pseudo-color enhancing system and infrared image self-adaptation and turns pseudo-color Enhancement Method embodiment and belong to same design, and its specific implementation process refers to embodiment of the method, repeats no more here.

The invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.

To sum up, the present invention, by having introduced medium filtering and Sobel edge algorithms, carries out pre-service to infrared image, by adjusting suitable weighting factor, can reach the strengthening image detail, reduces the purpose of picture noise.Adaptive color space conversion method, can the be real-time infrared image dynamic range be stretched, give prominence to object observing in the time of the compacting background, increased image definition, contrast, the pcolor picture of outputting high quality.Guarantee picture quality when stretching dynamic range of images, reinforcement image detail, made imaging effect that very large lifting be arranged.

Above-described embodiment is preferably embodiment of the present invention; but embodiments of the present invention are not restricted to the described embodiments; other any do not deviate from change, the modification done under Spirit Essence of the present invention and principle, substitutes, combination, simplify; all should be equivalent substitute mode, within being included in protection scope of the present invention.

Claims (10)

1. an infrared image self-adaptation turns pseudo-color Enhancement Method, it is characterized in that, described method comprises:

According to medium filtering, the infrared image of inputting is carried out to denoising;

According to the Sobel rim detection, infrared image is carried out to edge extraction;

The data that strengthen with image detail process normalized by generating, be converted to gradation data;

Described gradation data is carried out to statistics with histogram and the calculating of mapping colour code bar bound;

According to described statistics with histogram result and colour code bound, according to mapping ruler, carry out the color space conversion.

2. the method for claim 1, is characterized in that, described according to medium filtering to the input infrared image carry out denoising, comprising:

Adopt the 5x5 medium filtering to carry out denoising to the infrared image of input;

Filtered result and original view data are carried out to addition, obtain new image.

3. the method for claim 1, is characterized in that, describedly according to the Sobel rim detection, infrared image carried out to edge extraction, also comprises:

By edge detection results with the weighting scheme image that is added to.

4. the method for claim 1, is characterized in that, described described gradation data carried out to statistics with histogram and mapping colour code bar bound is calculated, and comprising:

Detect the histogram distribution data of scanned infrared image, find out the roughly distribution parameter of its gray scale: the average gray value avg of the minimum gradation value low of image, the maximum gradation value high of image, image;

The minimum gradation value low of image is mapped to " color_low "; The maximum gradation value high of image is mapped to " color_high "; The average gray value avg of image is mapped to " color_avg ";

According to current gray level mean value calculation avg_id, with min_level, compare, if average gray avg is less than threshold value min_level, increase the colour code lower limit; Otherwise the colour code lower limit is constant;

In like manner calculate the colour code upper limit.

5. the method for claim 1, is characterized in that, describedly according to mapping ruler, carries out the color space conversion, comprising:

The colour code bar that calculating makes new advances, traversal histogram horizontal ordinate, calculate the color that all gray-scale values are corresponding one by one;

Greyscale image transitions is become to coloured image.

6. the method for claim 1, is characterized in that, in described Sobel rim detection, Sobel operator template adopts $\left[\begin{array}{ccc}-\mathrm{xf}& 0& \mathrm{xf}\\ -\mathrm{yf}& 0& \mathrm{yf}\\ -\mathrm{xf}& 0& \mathrm{xf}\end{array}\right]$ With $\left[\begin{array}{ccc}-\mathrm{xf}& -\mathrm{yf}& -\mathrm{xf}\\ 0& 0& 0\\ \mathrm{xf}& \mathrm{yf}& \mathrm{xf}\end{array}\right],$ Adjust according to demand rim detection; Described xf, yf are level, vertical direction weighting coefficient.

7. an infrared image self-adaptation turns pseudo-color enhancing system, it is characterized in that, described system comprises median filter unit, edge detection unit, normalization unit, colour code bar computing unit and map unit, wherein,

Described median filter unit, for carrying out denoising according to medium filtering to the infrared image of inputting;

Described edge detection unit, for carrying out edge extraction according to the Sobel rim detection to infrared image;

Described normalization unit, the data that strengthen with the image detail process normalized for generating, be converted to gradation data;

Described colour code bar computing unit, for carrying out statistics with histogram and the calculating of mapping colour code bar bound to described gradation data;

Described map unit, for according to described statistics with histogram result and colour code bound, carry out the color space conversion according to mapping ruler.

8. system as claimed in claim 7, is characterized in that, described system also comprises the first superpositing unit, for the filtered result of described median filter unit and original view data are carried out to addition, obtains new image.

9. system as claimed in claim 7, is characterized in that, described system also comprises the second superpositing unit, for the edge detection results by described edge detection unit with the weighting scheme image that is added to.

10. system as claimed in claim 7, is characterized in that, described colour code bar computing unit comprises statistics with histogram subelement and colour code bar bound computation subunit, wherein,

Described statistics with histogram subelement, for carrying out statistics with histogram to described gradation data;

Described colour code bar bound computation subunit, for calculating mapping colour code bar bound.

Images