Bloodstains: Detection and Age Estimation
Written by Meez Islam, PhD, Bo Li, PhD, Liam O’Hare, PhD & Peter Beveridge   

Bloodstains are among the most important types of forensic evidence at scenes of violent crime. There are well-established molecular biology methods for DNA identification, and blood-pattern analysis can be important for crime scene reconstruction. However, there are significant problems with the techniques currently used to detect blood, and there is no reliable method to estimate the age of bloodstains.

Currently, identification of blood at a crime scene or in a forensic laboratory is primarily based on chemical presumptive tests. The Leuco Malachite Green (LMG), Kastle-Meyer (KM), and Luminol tests cause color changes or fluorescence in the presence of blood. While these tests can be very sensitive, they are not very specific and can give rise to both false positives and negatives. In addition, these chemical tests may cause contamination of the stain, leading to dilution and alteration of the shape of the stain—as well as potentially affecting subsequent DNA analysis. Other challenges facing the forensic examiner include problems detecting latent stains or stains on dark backgrounds on recovered evidence.

The age estimation of bloodstains recovered from a crime scene could provide valuable information relating to the timeline of a violent crime and could lead to the inclusion or exclusion of persons of interest during an investigation. Currently there are no established techniques to provide robust and reliable age estimations of bloodstains at the crime scene.

There is therefore a need for a technique that is non-contact, non-destructive, and able to positively identify bloodstains from other substances of similar appearance, and also able to estimate the age of the bloodstain. Instruments based on visible wavelength hyperspectral imaging may provide the means of making such measurements. These are currently being investigated and developed at Teesside University in the United Kingdom. They rely on the visible spectrum of hemoglobin that is present in red blood cells (Figure 1). The spectrum between 400 and 600 nm is dominated by a relatively sharp peak centered at approximately 415 nm that is commonly called the Soret band while two weaker peaks around 540 and 575 nm, called the b and a peaks respectively, are also present. The Soret band is used to positively identify bloodstains while the a and b peaks are used for the age estimation of bloodstains.

Figure 1—Visible reflectance spectrum of a bloodstain when fresh and at 15 days old.

Instrumental setup

The current prototype hyperspectral imaging instrument at Teesside University has been constructed by coupling a liquid crystal tuneable filter (LCTF) to an imaging camera, allowing wavelengths between 400 and 700 nm to be scanned at 5 nm intervals. Illumination is provided by a solid-state light source that emits about 10W of radiant power between 400 and 700 nm. Measurements are made in reflectance mode and instrument control and data analysis is achieved through custom software that runs on a laptop PC. While the current instrument is portable, it still requires access to an electrical power point.

Identification of bloodstains

An object appears red to our eyes because, when illuminated by white light, the object absorbs in the blue part of the spectrum. Thus, the red color of blood is predominately due to the absorption of light around 415 nm by hemoglobin, known as the Soret band absorption. However, when compared to the absorption spectrum of other red substances, the Soret band absorption is found to be much sharper; this forms the basis of our methodology.

A statistical correlation can be used to compare the spectrum of a suspect stain against a reference bloodstain spectrum. If the correlation is above a selected threshold value, the stain is identified as blood. In a processed hyperspectral image this can be done pixel by pixel; we choose to mark those pixels where the presence of blood is confirmed as white, while the absence of blood is colored black. Thus, our processed images are predominantly black with white areas showing the locations of bloodstains. Figure 2 shows examples of red-colored stains in the top color photograph, while on the processed hyperspectral image on the bottom, only the actual bloodstain appears as white. To date, our methodology has been tested on over 50 red-colored substances and other stains that could be confused with blood, and no false positives have been generated.

Figure 2—Color photograph of three red stains (above) and the processed hyperspectral image (below).

The technique currently works best on light-colored substrates where the intensity of reflected light is high. On darker substrates, the background absorbs a large fraction of the light and the processed image contains more noise, making reliable discrimination of bloodstain more difficult. However, successful measurements have been demonstrated on backgrounds where detection is difficult by eye—for example, red backgrounds and on dark blue jeans and, with a more sensitive camera and a more powerful light source, we hope that successful detection should also be possible on black substrates.

On light-colored backgrounds, diluted bloodstains can be detected even at the level at which they become visually latent. On white paper, blood diluted to greater than 100 fold becomes latent but the processed hyperspectral image is currently able to identify a bloodstain at up to 500-fold dilution. The system can also detect a latent bloodstain at 32-fold dilution on a red background. Again, we are confident that the use of a more sensitive camera and more powerful light source will also result in lower limits of detection for diluted bloodstains.

Age estimation of bloodstains

Historically, the change in the color of a fresh bloodstain from bright red to dark brown as it ages is well known. The chemical process underlying this change has been found to be due to the oxidation of oxyhemoglobin (HbO2) that is formed when hemoglobin comes into contact with oxygen in air. Spectrally, this process manifests itself in the region of the a and b peaks, which show significant change in intensity and shape as a bloodstain ages (Figure 1). Our methodology uses this change as the basis for the age estimation of bloodstains.

Two methods are used. The first is based on a statistical model that identifies the spectral wavelengths at which the greatest changes take place, and then uses these for a predictive model based on linear discriminant analysis (LDA). The second method uses a wavelength ratio to create a false color scale that can then be used to visually indicate the age of the bloodstain. As the LDA model uses more wavelengths, its accuracy is greater. Daily measurements have been made on bloodstains kept under controlled environmental conditions over a one-month period. The LDA model is able to predict the age of a test bloodstain with an average error of approximately 1 day over a 1-month period. The false color method shows visible daily changes in color over the first 6 days after which the color changes over 2- to 3-day time periods. Figure 3 shows an example of the color changes between 0 and 21 days.


Figure 3—The images above show: a) a false-color image of four bloodstains of different ages, obtained from a processed hyperspectral image; and b) a color photograph of the same four bloodstains.

Analysis of the rate of change of the spectra shows that the greatest change actually occurs over the first day of aging and thus greater accuracy can be obtained for measurements over a 1-day time period. To test this, hourly measurements were made on bloodstains over a 22-hour time period and both the LDA method and the false color method were used to predict the age of test bloodstains. The LDA model gave an average accuracy of approximately 0.7 hours over 22 hours, while for the false color method, visible changes in color could be observed hourly over the first 10-15 hours. Either of these methods has the potential to be used as an alternative way of estimating the post mortem interval for violent crimes where bloodstains are present or to individually estimate the times of a series of events involving blood spatter.

Further work

The methodology for the age estimation of bloodstains needs further measurements and development as the current aging measurements have been made under controlled environmental conditions to allow the underlying chemical change to be accurately studied. However, the aging process is likely to depend on environmental variables such as temperature, humidity, and light intensity, as well as physical variables such as substrate. Given that these variables will be uncontrolled at crime scenes, it will be necessary to make further measurements to investigate the effects of these variables on the aging process separately. It should in principle be possible to parameterize the dependencies and therefore create a bloodstain aging model that will take into account the actual or estimated environmental and physical conditions at the crime scene.

The methodology developed for the detection and identification of bloodstains is already robust and reliable and appears to work well on most substrates aside from very dark substrates. The use of a more sensitive camera and a more powerful light source should allow detection on even black backgrounds. The prototype instrument also needs to be engineered to allow portable use and ruggedized to the levels required for a commercial instrument.

The goal is to ultimately develop a single instrument that will be able to detect and identify a bloodstain and also estimate its age. We are currently seeking active collaboration with instrument manufacturers who can help with the further development of a robust and sensitive prototype and in taking the product to market. We are also interested in working with forensic providers and police forces to help validate the method.

About the Lead Author

This e-mail address is being protected from spam bots, you need JavaScript enabled to view it is a Reader in Physical Chemistry at Teesside University (UK) whose research interests lie in the development and application of techniques based on optical spectroscopy. He obtained his undergraduate degree in Chemistry from Oxford University (UK) and his PhD in Physical Chemistry from Birmingham University (UK).


Li, B., P. Beveridge, W.T. O'Hare, and M. Islam. "The application of hyperspectral imaging to the detection and identification of bloodstains." (Submitted to Science and Justice, December 2013)

Li, B., P. Beveridge, W.T. O'Hare, and M. Islam. "The age estimation of bloodstains up to 30 days old using visible wavelength hyperspectral image analysis and linear discriminant analysis." Science and Justice. 53(3), pp. 270-277 (2013)

Li, B., P. Beveridge, W.T. O'Hare, and M. Islam. "The estimation of the age of a bloodstain using reflectance spectroscopy with a microspectrophotometer, spectral pre-processing and linear discriminant analysis." Forensic Science International. 212, 198-204 (2011)

< Prev   Next >

Court Case Update

FINGERPRINT EVIDENCE went through a nearly three-year ordeal in the New Hampshire court system, but eventually emerged unscathed. On April 4, 2008, the New Hampshire Supreme Court unanimously reversed the decision of a lower court to exclude expert testimony regarding fingerprint evidence in the case of The State of New Hampshire v. Richard Langill. The case has been remanded back to the Rockingham County Superior Court.