The Importance of Near-Infrared Sensors and Their Typical Applications in Hyperspectral Cameras and Spectral Imaging

Near-infrared (NIR) sensors capture and analyze light within the 900 to 1700 nm wavelength range, which is beyond visible light, and can assess object details that are not detectable by the human eye. Hyperspectral cameras and sensors operating in the NIR region are increasingly being adopted in research labs and industrial machine vision. Applications such as plastic and textile sorting, food and pharmaceutical quality control, and thin-film thickness and moisture measurement demonstrate the value and versatility of NIR hyperspectral imaging (HSI).

In this article, we explain NIR sensors’ role in hyperspectral cameras and spectral imaging and elucidate the capabilities and advantages of NIR hyperspectral imaging technology when successfully applied in industrial settings.

Figure 1: NIR hyperspectral imaging captures spectral information within the 900 to 1700 nm wavelength range in the electromagnetic spectrum.

Understanding Near-Infrared Hyperspectral Imaging

Hyperspectral cameras combine imaging and spectroscopy. Hyperspectral cameras acquire hyperspectral data cubes, gathering spectral and spatial information pixel by pixel. Each pixel yields a spectral signature, providing unique “fingerprints” for different materials. The technology aids in classification and analysis, discerning distinct materials and their properties based on their spectral characteristics.

Certain materials can be detected specifically in the near-infrared range between 900 to 1700 nm due to organic compounds’ unique molecular vibrations and absorbance characteristics within this wavelength range.

Figure 2: NIR hyperspectral cameras identify different materials in real-time based on their unique spectral “fingerprints”.

Applications That Benefit From NIR Hyperspectral Imaging

NIR hyperspectral imaging has been proven helpful across various industries, significantly improving inspection, sorting, and analysis processes. Hyperspectral imaging provides real-time information about the manufacturing process, helping ensure compliance with regulations, reduce costs, and improve quality.

NIR Sensors for Accurate and High-Speed Plastics Sorting

NIR HSI can be used to sort plastics efficiently for recycling, achieving high purity rates. While short-wave (SWIR) and middle-wave (MWIR) infrared HSI cameras can also sort plastic, the NIR sensors are typically the most cost-efficient solution for plastic sorting and can sort the most common plastics (Figure 3). In the case of black plastics, the Specim FX50 MWIR HSI camera is the only viable solution.

Figure 3: The Specim FX17 NIR HSI camera can distinguish the most common polymer materials, including PE, ABS, PVC, PS, PA, PP, and PET.

Prodecologia, a Ukrainian machine builder, has developed a NIR separator employing a Specim FX17 camera, which identifies, sorts, and separates different types of polymers, enhancing the efficiency of the polymer recycling process. Despite the complexity of polymer processing and the diversity of recyclable polymers, Prodecologia has achieved polymer purity scores greater than 98% with the NIR HSI technology. The Prodecologia’s NIR separator separates plastics such as PET, PVC, PE, and PP. It has a 600–700 kg/h capacity for a mix of PP and PE and 500–600 kg per hour for PET flakes.

Figure 4: The NIR plastic separator from Prodecologia.

NIR Sensors for Rapid Textile Sorting and Recycling

NIR hyperspectral imaging also allows for identifying the composition of textile products, as different fibers have unique spectral characteristics. Separating different textile materials is complicated because they contain various raw materials and combinations of natural and synthetic fibers. NIR hyperspectral imaging has revolutionized textile sorting for PICVISA, a Spanish company specializing in optical sorting, robotics, AI, and deep learning.

PICVISA developed an automated textile sorting machine that integrates a Specim FX17 camera along with several complementary technologies and AI. According to Daniel Carrero, product manager of PICVISA, “NIR HSI is a game-changer. It enables us to identify materials and compositions of garments for preselection before recycling, providing a complete analysis with excellent spectral resolution.”

NIR spectroscopy is used in the textile and fiber industry also to control quality and analyze parameters such as fiber composition, moisture content, dye concentration, and other properties.

Figure 5: Powered by NIR HSI camera, the PICVISA machine ensures precise classification of endless combinations of fibers and colors.

NIR Sensors for Food Grading, Quality Control, and Foreign Object Detection

NIR HSI is employed in the food industry for sorting, grading, quality control, and authentication purposes. NIR HSI facilitates the accurate and dependable assessment of chemical composition and identification of foreign objects by leveraging the distinct spectral patterns inherent to each material. Various food properties — such as moisture, fat, protein, sugar content, and other nutritional parameters — have unique light reflection properties within the NIR spectral range.

As shown in Figure 6, with an RGB camera it’s challenging to distinguish between fat and meat and contaminants such as plastic or wood. However, NIR HSI accurately identifies these contaminants and detect the fat’s location in the meat, as their spectral signatures differ.

Figure 6: RGB vs. NIR HSI image of foreign objects (plastic and wood) and fat in meat acquired with Specim FX17 camera.

NIR HSI has many benefits over traditional color and filter cameras and point spectrometers. Hyperspectral imaging enables comprehensive, rapid, and non-destructive inspection of large areas or multiple samples. This is particularly important in the food industry, where high throughput, real-time monitoring, and hygiene are essential.

Figure 8: Alpma uses the Specim FX17 mounted on a robotic arm with illumination that moves around large blocks of cheese to detect defects and contaminants such as thin plastic residues and mold.

NIR sensors for moisture measurement

Water absorption in the NIR spectral range is a significant phenomenon with implications in various scientific and technological fields. In the NIR range, water molecules absorb light due to overtones and combinations of the fundamental vibrational modes of water (stretching and bending motions of the H-O-H molecule). The absorption bands are primarily observed around 760 nm, 970 nm, 1200 nm, 1450 nm, and 1940 nm.

The intensity of absorption varies with the wavelength. For instance, at 1450 nm, water exhibits a strong absorption peak, making the NIR wavelength particularly useful for detecting water content in various substances. This property of water absorption in the NIR range is exploited in numerous applications, including remote sensing, agriculture, food, paper, wood, cosmetics industries, and biomedical imaging, among others.

One application where water absorption in the NIR spectral range can be leveraged is in measuring moisture and fat content in minced meat (Figure 9).

Figure 9: The Specim FX17 camera and its NIR spectral range proved highly effective in accurately measuring the moisture and fat content of the minced meat.

The analysis of fresh produce is another application where water absorption in the NIR spectral range can be exploited. Wageningen University & Research (WUR) in the Netherlands has developed a stand-alone All-in-One Spectral Imaging (ASI) system based on the VNIR (visible and near-infrared) and NIR hyperspectral cameras to monitor fruit freshness and maturity.

Figure 10: Wageningen University & Research’s smart All-in-one Spectral Imaging (ASI) laboratory system for automated fresh produce analysis.

NIR sensors for Inspecting Heat-Sealed Packages

Another relevant application for the NIR spectral range is inspecting heat-sealed packaging. If a contaminant is present at the sealing point, it could cause a leak, potentially spoiling the product inside.

NIR imaging goes beyond what we can see, allowing the detection of contamination like melted grease hidden in sealed packages. The human eye would typically miss this invisible contamination. Conventional RGB imaging would be ineffective if the melted grease were positioned behind an opaque printing area on the packaging. However, the greater penetration depth of the NIR HSI can detect it, highlighting the technology’s power for quality and safety checks. A successful case was reported by Minebea that used the Specim FX17 in combination with X-ray imaging to achieve this.

Figure 11: NIR hyperspectral imaging can detect melted grease in heat-sealed areas of packages that are undetectable by the human eye. (Image courtesy of Minebea)

NIR Sensors for Pharmaceutical Inspection and Analysis

In pharmaceutical inspection and analysis, the NIR range is particularly useful for ensuring the quality and composition of medications for safety and efficacy. NIR HSI has been successfully applied for monitoring various types of active ingredients, even through blister packaging. HSI is an especially valuable inspection method when pharmaceutical products, e.g., tablets or powder, are visually identical.

Figure 12: The Specim FX17 allows pharmaceutical companies to check for contaminants and ensure that each type of tablet has the correct composition of ingredients in the right quantities.

NIR Sensors for Measuring Thin-Film Thickness

NIR HSI is also helpful in inspecting the thickness of thin films and coatings. NIR HSI can enhance thin-film efficiency and coating quality control systems, offering 100% in-line inspection for improved quality and waste reduction, which enables faster and more comprehensive inspection compared to current solutions based on point spectrometers that provide only partial monitoring.

Figure 13: NIR hyperspectral cameras, such as the Specim FX17, offer a solution by capturing high-resolution spectroscopic data across the entire width of the film.

In addition to the examples provided in this article, NIR hyperspectral imaging offers a spectrum of possibilities for accurate and real-time inspection for various scientific and industrial use cases.

Related products:

Specim FX17

Specim GX17

Specim AFX17