Automotive production plants continually face two common problems in the application of coatings: imperfect coverage and contaminants. Thermal vision systems provide a solution.

Thermal imagery, or thermography, can be defined as two-dimensional scanning of a surface to acquire its thermal emission using focal plane arrays sensitive to infrared rays. The thermal emissions are acquired and processed as static images, which can be done in either a passive or active mode. Passive imaging exploits the built-in thermal signature in a process. Active thermal imaging requires an external heating or cooling source to create a temperature gradient across the product or process being monitored.

Fuel Tank Coating

Figure 1. Results from the on-site tests: (a) thin coating (pinholes) detected at the exit of curing oven, and (b) corresponding processed thermo-gram image, pinholes extracted.

A passive thermal imaging system can detect small pinholes (1 mm in diameter) in the 500-micron-thick coating applied over steel automotive fuel tanks. (These tiny voids expose the tank to accelerated corrosion.) Thermal imaging solves problems associated with reflection-based detection methods, such as CCDs (Charged Coupled Devices). These problems include tank size and complex geometry, plus the limited ability of CCDs to evaluate coating cracks under industrial lighting conditions.

A thermal vision approach exploits the difference in contrast between the thermal emission from a pinhole and that coming from the coating.

For the fuel tank project, laboratory testing utilized the FLIR ThermaCam SC2000 micro-bolometric detector for acquisition, while xenon heating lamps and an industrial fan were used to simulate oven heating. These tests showed that the pinholes’ signature is detectable and is in agreement with the simulation.

The actual online system installation included an industrially packaged micro-bolometric array, FLIR’s ThermaCam A40M, at the end of the curing oven cycle. Sample results from the system’s online operation and processing are shown in Figure 1. To display the thermal images to a user and enable system automation, an in-house processing algorithm was developed to process the thermal maps.

Body Paint Contamination

An active thermal vision system can be use to detect contaminants such as dust and fibers on newly painted car bodies. Again, thermal imaging is more reliable than reflective-based systems.

Figure 2. Paint defect detection head, showing two optical heating elements and the cooled infrared 256 x 256 focal plane InSb detector array (Merlin-Indigo, product of FLIR)

Xenon lamps are used as the external heat source in this application because of the large variations in the convection oven heating patterns relative to the contaminants’ thermal contrast signals. Figure 2 illustrates the compact detection head. It is envisioned that this assembly will be deployed on a robotic manipulator to scan the large, complicated geometry of a car body.

An in-house algorithm was developed to (1) neutralize the emissivity variations though a pre-processing step, (2) characterize the deviations in thermal emission based on size and temperature, and (3) display the size and location of defects to the operator in an automated manner. Data storage was configured as a quality assurance database.

More Information

For more information on FLIR’s thermal imaging systems, visit

Imaging Technology Magazine

This article first appeared in the June, 2007 issue of Imaging Technology Magazine.

Read more articles from the archives here.