AUTOMATIC PAINT DEFAULT REPAIR

The invention relates to a process for automated repair of a damaged coated surface. The invention further relates to a system for carrying out the process.

A process for automated repair of a coated substrate is known from United States patent US 6714831 B. This document relates to a method and assembly for inspecting painted surfaces of a vehicle body, locating and tracking defects in the painted surface, and repairing such defects. The method includes developing paint defect data using electronic imaging of the vehicle body. The electronic imaging is referenced with vehicle CAD data to develop three- dimensional paint defect coordinates for each paint defect. A repair strategy is developed based upon the paint defect data and the paint defect coordinates. Finally, an automated repair is performed on the paint defects based upon the repair strategy.

The assembly includes an imaging system which generates paint defect data by electronically imaging the vehicle body. The imaging system may be an optical system, such as a vision scanner with telecentric optics. The imaging system generates paint defect data as it scans the vehicle body. Paint defect data includes the size, type, and location of the paint defect. The paint defect data is passed on to a vision cell controller. The vision cell controller references the information from the imaging system with vehicle CAD data to develop three- dimensional paint defect coordinates for each paint defect. The assembly further includes a robot cell controller which develops a repair strategy based upon the paint defect data and the paint defect coordinates. The repair strategy may include path and processing parameters, tools, and robot choice. The assembly also includes an automated robotic repair system which performs an automated repair on the paint defects based upon the repair strategy. The known process is suitable for repairing defects which arise from errors during application of paint in a motor vehicle assembly plant, where the original, factory applied coating material is readily available. The known process is not adapted for repair of damage resulting from collision of vehicles. In particular, it is not

described how to carry out an invisible or practically invisible repair of a damaged coated surface when the original, factory applied colour- and/or effect- imparting coating composition is not available. Although modern colour matching techniques allow for very good matching of automotive repair paints with the original, factory applied colour of the coating layer, it is still difficult to prepare coating compositions which match the factory applied paint colour exactly. This is due, among others, to small variations within batches of factory applied paints, aging effects of the factory applied coating during the service life of the vehicle, and the limited number of base tints from which matching repair paints can be prepared by mixing. In cases where a repaired spot is surrounded by the original factory applied coating, small variations in colour are particularly easily perceived by the human eye.

It would be desirable to use an automated repair process also for collision repair and other minor damage during the service life of a motor vehicle. This would be particularly attractive in view of rising labour costs.

It is the object of the present invention to provide a process for automated repair of a damaged coated surface which does not have the above-mentioned drawback. More in particular, the process should lead to invisible or practically invisible repairs also when areas of body panels of motor vehicles are repaired which are surrounded by the original, factory applied coating.

It has now been found that an invisible or practically invisible automatic repair of damaged spots can be achieved when the colour- and/or effect-imparting repair coating composition is applied with a hiding gradient. Accordingly, the invention provides a process for automated repair of a damaged coated surface under the control of a data storage and processing unit, characterized in that it comprises the steps of a) determination of the desired visual properties data of the damaged surface after repair, and transmittal of these data to the data storage and processing unit,

b) selection of a colour- and/or effect-imparting coating composition matching the data determined in step a), c) determination of the hiding gradient in which the colour- and/or effect- imparting coating composition has to be applied in order to ensure that the slope of the colour difference between the repaired area and the surrounding area does not exceed a predetermined threshold value, d) automatic application of the selected colour- and/or effect-imparting coating composition to the damaged portion of the coated surface with the hiding gradient determined in step c).

By hiding gradient is meant that the colour- and/or effect-imparting coating composition is applied in such a way that the hiding power of the colour- and/or effect-imparting coating layer decreases towards at least one border of the applied coating layer. By applying this technique, which is also known as fade- out, any colour differences between the original, factory applied coating and the repair coating become less visible or even invisible, because there is a smooth, gradual transition between the colour of the original coating and the repair coating. The automated process according to the invention also minimizes or eliminates the exposure of workers to solvents and other toxic substances present in coating materials.

The process of the invention is suitable for the repair of almost all types of coated substrates. It can be used with particular advantage for the repair of transportation vehicles, such as automobiles. Also large transportation vehicles, such as trucks, buses, trains, and aircraft, as well as smaller vehicles, such as bicycles or motorcycles, can be repaired with the process. The process can be used to repair the coating on an entire body panel of an automobile. However, the advantages of the process can be more fully exploited for the repair of minor damage which would not require recoating of an entire body panel. In such cases, only the area of the coated substrate which is actually damaged or

additionally a small area surrounding the damage has to be recoated. This type of repair is also known as spot repair.

The process of the invention is carried out under the control of a data storage and processing unit. A suitable data storage and processing unit may be a general purpose computer having a central processing unit, a memory, means for loading an application program into a defined address space of the memory, and a computer monitor. Also connection means to a communication network for data transfer to and from remote locations, such as a modem for connection to the Internet, are useful.

The data storage and processing unit suitably has access to one or more databases. The database may be stored internally, for example on a hard disk, or externally. The data processing unit may have access to the database via a data communication line, for example via the Internet or via an intranet. In that case, the at least one database may be physically located in a remote location.

Generally, the process comprises the automatic determination of the location, geometry, size, shape, roughness, and texture data of the damaged portion of the coated substrate. This data can suitably be determined using computer vision technologies. Examples of methods to determine these properties are described in the thesis of G. Bradshaw, Non-Contact Surface Geometry Measurement Techniques, Trinity College, Dublin, Ireland, 1998/1999. Also, different technologies capable of measuring one or more of the above- mentioned data categories can be used successively or simultaneously. The data determined in this step is transmitted to the data storage and processing unit, which is configured to receive such data.

An essential step of the process of the invention is the determination of the desired visual properties data of the damaged surface after repair. The visual

properties after repair should match the visual properties of the original coating sufficiently closely, in order to minimize the risk of any difference being visible. Examples of visual property data are colour data, gloss data, effect data, and texture data. Depending on the type of surface to be matched for visual properties, one or more of said data have to be determined. Besides colour, a paint film shows numerous further visual properties. Colour can be expressed by the paint film reflection as a function of wavelength of visible light. Alternatively, colour can be expressed in accordance with the so-called CIE Lab system, as defined by the Commission International d'Eclairage, or similar systems, such as the CIE Luv, CIE XYZ systems or the Munsell system. Particularly when effect pigments, such as for example aluminium flake pigments or pearlescent pigments, are used, the look of a paint film is not one of uniform colour, but shows texture. This can include phenomena such as coarseness, glints, micro-brilliance, cloudiness, mottle, speckle, sparkle or glitter. In the following, texture is defined as the visible surface structure in the plane of the paint film depending on the size and organization of small constituent parts of a material. In this context, texture does not include roughness of the paint film but only the visual irregularities in the plane of the paint film. Structures smaller than the resolution of the human eye contribute to "colour", whereas larger structures generally also contribute to "texture".

Texture data can for instance include the particle size distribution of the effect pigments in the toner and the optical contrast, defined as the difference in lightness, between the effect pigment and the other pigments present in the coating. Also particles which are not directly observable by themselves can contribute to the overall visual appearance of a paint film. Des-orienters are an example of such particles. Effect pigments are generally flakes tending to take a horizontal orientation in a cured film. To prevent this, and to obtain more variation in flake orientation, spherical particles are used, referred to as des- orienters. Using des-orienters in a metallic paint results in more glitter. The

determination of texture data is described in more detail in International patent application WO 01/25737.

The visual property data can be determined by known instrumental analysis. Using an angle-dependent spectrophotometer colour and colour effect data can be uniquely identified. The instrumental analysis quantifies colour and colour effect parameters, examples of which are lightness, hue, and chroma data, and their angle dependence.

If the surface to be matched is the surface of a motor vehicle, an alternative way of determining visual property data is the retrieval of the so-called colour number, optionally refined by a colour variant code. The colour number is the code that represents the manufacturer's paint colour and colour effect, and can be found on the vehicle of interest. Paint colour and colour effect may vary within one colour number, for example due to minor variations between paint batches or application variables, such as relative humidity or temperature. Those variations can be taken into account using a variant database.

Still another way of recording visual property data involves the retrieval of the Vehicle Identification Number (VIN), which can be found on every motor vehicle. The VIN is a structured combination of characters assigned to a vehicle by the manufacturer for identification purposes. The VIN consists of three parts, to wit the World Manufacturer Identifier (WMI), the Vehicle Description Section (VDS), and the Vehicle Indicator Section (VIS). The VIS, in conjunction with the VDS, ensures a unique identification of all vehicles produced by each manufacturer. Once colour code variants have been linked to the combination of VIS and VDS, the VIN can be used to determine the visual property data of a vehicle. The use of the VIN to determine the visual property data of vehicles is described in more detail in European patent application EP 1355242 A.

It is also possible to use a combination of the above-mentioned methods for determination of the visual property data. The most appropriate and thus preferred method of determination may vary. For example, if the visual properties of an aged automobile are to be determined, an instrumental analysis

of the visual properties may be preferred, because the visual properties may have changed too much over time to be adequately described by the colour code or the VIN.

The process of the invention also includes the step of selecting a colour- and/or effect-imparting coating composition matching the visual properties data determined previously. After transmission of the data to the data storage and processing unit, the composition data of a coating composition having an acceptable match of visual properties are determined. Coating composition data can be determined in a number of ways, i.e. by means of search procedures, calculations, or combinations of the two.

For example, use may be made of a database comprising composition formulae having visual property data linked thereto. Using the determined visual property data of the surface to be matched in visual properties, the most closely matching coating composition formula can be found.

Alternatively, it is possible to use a database having visual property formulae with spectral data linked thereto. Known calculation methods can be used to calculate the visual property data of the visual property formulae and compare them. Also, a databank can be used in which the absorption and reflection data, the so-called K and S data, of pigments are stored. Using K and S data in combination with pigment concentrations makes it possible to calculate the formula of which the visual property data most closely match the visual property data of the surface to be matched. It is possible to combine the aforesaid search and calculation methods. In one embodiment, the coating composition is a base coat composition such as is typically used in base coat/clear coat systems on motor vehicles. Such a base coat composition is a liquid coating composition comprising pigments and/or effect-imparting particles, such as mica particles or metal flakes. Application thereof is preferably carried out by spraying. However, other application methods are also possible, provided that the effect-imparting particles are

oriented in such a way that a colour effect can be matched independent of the viewing angle. Suitable base coat compositions are available commercially, for example from Akzo Nobel Car Refinishes under the trade designations Autowave and Autobase Plus.

The colour- and/or effect-imparting coating composition may be selected from a number of previously prepared available coating compositions. Alternatively, the visual properties-matching coating composition can prepared by mixing the required components in the required ratio. In one embodiment, the computer can transfer the coating composition data directly to an automated mixing machine. The automated mixing machine may automatically prepare the visual properties-matching coating composition based on the composition data. It is possible to calculate the required amount of visual properties-matching coating composition on the basis of the surface to be coated. In this case, the visual properties-matching coating composition is suitably prepared in the required amount. Otherwise, it is possible to prepare a standard amount of the visual properties-matching coating composition.

The visual properties-matching coating composition typically is a liquid coating composition which can be applied by spraying. The visual properties-matching coating composition may be water borne or solvent borne. It may suitably be prepared by mixing one or more colour- and/or effect-imparting modules, one or more binder modules, and a diluent module. Also a crosslinker may be added to the visual properties-matching coating composition. If the colour- and/or effect-imparting coating layer formed from the visual properties-matching coating composition is not covered by a clear coat layer, it is preferred that the colour- and/or effect-imparting coating layer is crosslinkable. Crosslinkable coating layers can be obtained from two- component coating compositions based on hydroxy-functional binders and isocyanate-functional crosslinkers. Such compositions are well-known in the art

and they are available as solvent borne compositions and as water borne compositions. Crosslinking may occur at ambient temperature or at elevated temperature.

It is also possible to use coating compositions which are curable by actinic radiation. In a still further embodiment, so-called dual cure compositions may be used, which are curable thermally and by actinic radiation. Such dual cure compositions may for example comprise binders having hydroxyl groups and (meth)acryloyl groups, and an isocyanate-functional crosslinker optionally having (meth)acryloyl groups.

A further step of the process of the invention is the determination of the hiding gradient in which the colour- and/or effect-imparting coating composition has to be applied in order to ensure that the slope of the colour difference between the repaired area and the surrounding area does not exceed a predetermined threshold value.

The required hiding gradient can be calculated on the basis of several input variables, for example the difference between the desired visual properties after repair and the visual properties data of the selected coating composition. A higher difference generally requires a larger gradient area in order to bridge the visual properties difference. The difference in original surface colour and repair coating colour can be leveled by a controlled fade-out generation. The colour difference δE is limited to a certain maximum to assure invisibility and the slope of the colour difference over a distance x must not exceed a certain critical threshold value. This actually means that when measuring colour starting from x=0 to the center of the refinish spot, the colour measurements must lie within a specific bandwidth, with a specific slope. Depending on how large the colour difference between repair coating composition and the original coated surface is, a certain fade-out size δx is chosen to design a colour fade-out with a slope under the critical maximum slope. This critical slope is colour- and effect- dependent and may be part of a database. The critical slope is chosen in such a

way that resulting repair is practically invisible to the human eye. The fade-out is also used to control other differences in visual properties, i.e. effects and texture in the case of metallic and/or pearlescent pigments.

Another feature having influence on the visibility of the refinished spot is the shape of the surface after refinishing, especially the absence of edge marks. Figure 1 is a schematic representation of the cross-section of a repaired coated substrate. (1 ) shows a dent filled with a hardened filler composition, (2) is a layer of primer/filler coating, (3) is a layer of colour- and effect-imparting base coat, and (4) represents a clear coat layer. In Figure 1 , (α _P ) is the edge angle between the coated substrate surface and the repair primer coating layer, (CI _B ) is the edge angle between the coated substrate surface and the repair colour- and effect-imparting base coat layer, (αc) is the edge angle between the coated substrate surface and the repair clear coat layer. The angles α between coated substrate and repair coating edge for the different repair coating layers is relevant for spot repair invisibility. Depending on the coating layer, these angles should have values close enough to 180 degrees in order to assure invisibility of the spot repair edges. The critical values of these angles also depend on the colour and the coating composition recipe. The critical angles can be integrated in a recipe database as process input.

The process of the invention further includes automatic application of the selected colour- and/or effect-imparting coating composition to the damaged portion of the coated surface with the previously determined hiding gradient. The colour- and/or effect-imparting coating layer is applied to the damaged area so as to achieve a hiding power gradient within the coated area, with full hiding power being achieved in the centre of the area of the applied coating layer and the hiding power decreasing towards at least one border of the area of the applied coating layer. This can for example be achieved by gradually decreasing the applied layer thickness of the coating towards the at least one

border of the area of the applied coating layer. This technique is commonly known in the refinish sector as fade-out. There are several ways to achieve a required layer thickness gradient. The speed of the movement of the application tool, such as a spray gun or air brush, relative to the substrate to be coated can be varied. A faster movement of the application tool leads to less application of coating material per surface unit, i.e. a lower layer thickness. Also the distance between application tool and substrate, as well as the angle between the mean movement direction of atomized paint emitted from a spray gun and the substrate to be coated will influence the layer thickness of the applied coating. It is also possible to decrease or increase the paint flow rate in the spray gun. Application of a more diluted paint can also be used to generate a layer thickness gradient. The non-volatile content of an applied paint can be gradually decreased by increasing the proportion of volatile organic diluent or water. Alternatively, the hiding gradient can also be achieved by variation of the coating composition. In that case, variations in the proportion of unpigmented clear binder and pigmented toner would cause a gradient in hiding power. It is also possible to use combinations of the above-mentioned techniques for generation of a hiding power gradient. Automatic application of the coating composition can be suitably carried out by a coating robot under the control of the data storage and processing unit. The application unit and the damaged substrate are moveable with respect to each other. In one embodiment, the application unit may for example be attached to a robotic arm moveable about various control axes. Alternatively, the application unit can be fixed and the substrate is moveable about various control axes. It is also possible that both the substrate and the application unit are moveable.

The application unit may for example be in the form of a spray gun, air brush equipment, or a dispense-jetting unit. The application unit suitably is connected to at least one reservoir of coating material.

Depending on the coating composition recipe used and on the shape of the coated substrate to be repaired, application and fade-out process planning is optimized, performed, and monitored using computer vision. Coating layer edge angles, determined by automatic measurement after application, optionally in combination with measured surface texture data, can be used to determine the need for an automatic sanding process step. The surface parameters, including edge angles, can also be used as input for the automatic sanding process.

When the determination of the visual property data indicates that a clear coat layer is required to match the visual properties, a clear coat layer is applied on top of the colour- and/or effect-imparting coating layer. The clear coat layer may be applied on top of the dried colour- and/or effect-imparting coating layer. Alternatively, the clear coat may be applied wet-on-wet on top of the colour- and/or effect-imparting coating layer. Wet-on-wet application means that the underlying colour- and/or effect-imparting coating layer is not fully dried or cured before application of the clear coat. It is also possible for the colour- and/or effect-imparting coating layer to be partially dried, for example by subjecting it to a so-called flash-off phase of 5 to 15 minutes, during which a substantial part of the liquid diluent can evaporate.

Automatic application of the clear coat may be done by spraying. However, other means of application, such as mentioned above for the visual properties- matching coating composition, may be employed as well. The choice of the appropriate application method may depend on the visual properties, such as gloss, to be matched. The application method giving the best match of visual properties is generally preferred.

The clear coat layer is preferably crosslinkable, as described above in respect of the visual properties-matching coating composition. Suitable clear coat compositions are available commercially, for example from Akzo Nobel Car Refinishes under the trade designation Autoclear.

In one embodiment, the process of the invention additionally comprises the automatic sanding of the damaged portion of the coated substrate. Sanding can improve the smoothness of the surface before application of a subsequent coating layer. Additionally, sanding improves the smoothness of the transition between the damaged area and the surrounding undamaged area. Sanding is usually carried out before application of a coating layer. When multiple layers are applied, the dried previously applied layer may be sanded before application of a subsequent layer.

In one embodiment, the process of the invention additionally comprises the automatic application of a primer/filler layer to the damaged portion of the coated substrate. The primer/filler layer is applied to the damaged substrate as described above for the colour- and/or effect-imparting coating composition. In a multi-layer coating system the primer/filler layer is the coating layer closest to the coated substrate. When a primer/filler layer is applied, it is always applied before the colour- and/or effect-imparting coating composition. Suitable primer/filler coating compositions are well known to the skilled person and are commercially available.

Suitable coating compositions harden after application. Hardening can be caused by physical processes, i.e. heating and/or evaporation of volatile diluents. Alternatively or additionally, hardening can be caused by chemical curing reactions. An optional curing and/or hardening step may be included in the process. Curing and/or hardening can be induced by supply of thermal energy or by actinic radiation, such as UV radiation, depending on the type of coating composition used. Also two-component coating compositions which are mixed immediately prior to application and which cure at ambient temperature can be used. Curing and/or hardening can be carried out after the application of individual coating layers. It is also possible to cure and/or harden a plurality of layers or even all layers together at the end of the process.

In one embodiment, the process of the invention is preceded by automatic filling of a dent with a hardenable filling material. Automatic dent filling suitably comprises the steps of a) determination of the actual geometry of the surface of the dent, b) determination of the desired surface geometry after filling of the dent, c) calculation of the difference in geometry between the actual geometry of the surface of the dent and the desired surface geometry and calculation of the volume and geometry to be filled and determination of the position and orientation thereof in space, d) determination whether the volume determined in step c) is below or equal to a predetermined value and i) if the difference is below or equal to the predetermined value, stopping the process, and ii) if the difference is above the predetermined value, continuation with steps e) to g), e) calculation, on the basis of the calculated volume and geometry to be filled, of layers \-\ to I _n corresponding to the volumes and geometries required to fill the dent, wherein n is an integer corresponding to the number of layers required to fill the dent, and h is the layer closest to the damaged substrate, f) application to the damaged substrate of layers h to l _m of a filler composition corresponding to the volume and geometry calculated for the respective layer, layers h to l _m corresponding to the layers which are applied subsequently without intermediate determination of the difference according to step d), and wherein m is an integer which is equal to or smaller than n, g) optionally, determination of the surface geometry after application of layers h to l _m ,

h) repetition of steps c) to g) until the difference in geometry determined in step c) is below or equal to the predetermined value.

The actual geometry of the surface of the dent to be filled can be determined by any suitable technology available. Examples of suitable technologies are point based techniques, line based techniques, and area based techniques. Generally, determination of the geometry of the surface of the dent can be carried out using contact methods or non-contact methods. A typical industrially applied contact method is implemented in a coordinate measuring machine. Further examples of suitable contact methods are the use of a touch probe or a linear position sensor, such as a linear variable differential transformer (LVDT). Examples of suitable non-contact methods are laser triangulation, time of flight measurements which use time as a surrogate measure for distance, fringe projection, X-ray, photogrammetry, and interferometry. These and other surface geometry determination methods are generally known, for example from the thesis of G. Bradshaw, Non-Contact Surface Geometry Measurement Techniques, Trinity College, Dublin, Ireland, 1998/1999. If required, the geometry of the surface surrounding the damaged area can be determined.

The geometry determination of step a) is suitably carried out by a geometry determination unit. The geometry determination unit and the three-dimensionally shaped damaged substrate are moveable with respect to each other. The geometry determination unit suitably is a computer controlled robotic unit. In one embodiment, the geometry determination unit can be attached to a robotic arm moveable about various control axes. Alternatively, the geometry determination unit can be fixed and the substrate is moveable about various control axes. It is also possible that both the substrate and the geometry determination unit are moveable.

The determined geometry data are generally transmitted from the geometry determination unit to an input unit of a data storage and processing unit capable of reading and processing the geometry data.

As mentioned above, the desired surface geometry after filling of the dent is determined. Generally, the desired surface geometry after filling of the dent is identical to or very similar to the original surface geometry before the damage occurred. In one embodiment, the desired surface geometry after filling of the dent can be determined by a reverse engineering step, such as extrapolation on the basis of the surface surrounding the dent. In particular in the case of relatively simple geometries, this method can be very suitable to determine the original surface geometry at the location of the dent. The surface geometry of some areas of automobiles, such as door panels, hoods, or roofs, can often be described by relatively simple geometrical functions. If the damaged area is sufficiently small, it can sometimes even be approximated as a plane. For such cases extrapolation from the surface surrounding the dent can give a very reliable approximation of the desired surface geometry after filling of the dent. In the case of symmetrical substrates, it is often possible to find a mirrored surface of the damaged spot on the opposite side of the symmetry plane. Automobiles generally are almost symmetrical about their centre plane. By using geometry information from the corresponding area of the other side of an automobile it is often possible to reconstruct the geometry of the damaged surface.

In another embodiment, it is possible to use geometrical data provided by the manufacturer of the three-dimensionally shaped substrate. Many articles are developed and manufactured using computer aided design (CAD). If available, CAD data can be used to determine the original surface geometry of the damaged substrate. Determination of the desired surface geometry after filling of the dent is suitably carried out by a data storage and processing unit under the control of a program.

The difference in geometry between the actual geometry of the surface of the dent and the desired surface geometry can be calculated by subtracting the geometry data of the damaged surface from the desired surface geometry data. The result of this step is a volume and geometry that has to be filled with filler composition, and the position and orientation thereof in space. This step is generally carried our by the data storage and processing unit mentioned above.

Depending on the required precision of the dent filling process, a value of the difference in geometry is predetermined below which no further filling is required.

Hence, when the difference between the actual and the desired surface geometry is below or equal to the predetermined value, the dent filling process is stopped or not even started. If the difference is above the predetermined value, the process is continued.

As mentioned above, the volume and geometry to be filled is determined. This volume and geometry is subdivided into volumes and geometries of filler composition which can be applied in a single step. Typically, the volume and geometry is subdivided into layers I which can be applied in a single step. The total number of layers into which the volume and geometry is subdivided is an integer n, and the layer which is applied first and which therefore is the layer closest to the substrate is labeled h. The subsequent layers are numbered consecutively up to layer I _n . The result of the calculation thus is a number of layers \-\ to I _n which can be applied to fill the dent. The number of layers needed to fill said volume and geometry depends on the filler composition used, the application process, and the characteristics of the optional curing step. The thickness of individual layers need not be the same. In one embodiment, the first layer(s), i.e. the layer(s) closest to the substrate, can be calculated to have a higher layer thickness than the outermost layers. When such a protocol is

used, faster filling of a dent can be achieved without compromising the accuracy of the dent filling process.

In the following step the layers of a filling material as calculated in the previous step are actually applied to the damaged substrate. Layers h to l _m of a filler composition corresponding to the volume and geometry calculated for the respective layer are applied to the damaged substrate, layers h to l _m corresponding to layers which are applied subsequently without intermediate determination of the difference, and wherein m is an integer which is equal to or smaller than n. In one embodiment, only a single layer is applied before the subsequent steps are carried out. In that case, m is 1. However, it is also possible to apply a plurality of layers or all required layers n before carrying out the subsequent steps. In that case, m is smaller than n and equal to n, respectively. As mentioned above, the individual layers need not be of the same layer thickness. Application of the filler composition is carried out by an application unit, such as a computer controlled robotic unit. The application unit and the damaged substrate are moveable with respect to each other. In one embodiment, the application unit can for example be attached to a robotic arm moveable about various control axes. Alternatively, the application unit can be fixed and the substrate is moveable about various control axes. It is also possible that both the substrate and the application unit are moveable.

The application unit can for example be in the form of a nozzle or a spout suitably connected to a reservoir of filler composition. Application by dispensing or dispense-jetting is preferred. In the case of dispensing, the filler composition is contained in a barrel and dispensed through a nozzle by applying force or pressure. By this technique individual drops or continuous lines can be dispensed through a nozzle. The nozzle is very close to the surface to which the filler composition is applied in order for the filler composition to make contact with that surface. In the case of dispense-jetting, a drop of filler composition is jetted from a nozzle and travels some distance through the air before making

contact with the surface. Although dispense-jetting is more complex than dispensing, it generally offers higher flexibility and the position of the nozzle is less critical.

Suitable filler compositions applied in the process are those materials known to the skilled person and commonly used for filling dents caused by damage, such as liquid or semi-liquid filler compositions or putties. Also so-called hot-melts can be used. Suitable filler compositions harden after application. Hardening can be caused by physical processes, i.e. cooling and/or evaporation of volatile diluents. Alternatively or additionally, hardening can be caused by chemical curing reactions. An optional curing and/or hardening step can be included in the process. Curing and/or hardening can be induced by supply of thermal energy or by actinic radiation, such as UV radiation, depending on the type of filler composition used. Also two-component filler compositions which are mixed immediately prior to application and which cure at ambient temperature can be used. Curing and/or hardening can be carried out after application of individual layers of filler composition. It is also possible to cure and/or harden a plurality of layers or even all layers together at the end of the filling process.

In one embodiment, filling additionally includes a sanding step. Sanding of the filled dent can improve the smoothness of the surface of the filled dent.

Additionally, sanding improves the smoothness of the transition between the filled dent and the surrounding undamaged area. Sanding is usually carried out after all layers of filler composition have been applied and hardened and/or cured. However, it is also possible to sand individual layers. When sanding is carried out, it is possible to use a standard sanding step in order to smooth the surface. Alternatively, sanding can be used to selectively remove hardened filler composition so as to achieve a desired surface geometry. It is also possible to have a sanding step interrupted by a geometry determination step in order to determine whether the selective removal of hardened filler composition is sufficient. Sanding is likewise carried out automatically by a sanding unit under

control of the data storage and processing unit. The surface geometry data determined in previous steps or determined specifically for sanding control can suitably be used as input data for control of a selective sanding step. In one embodiment of dent filling, the surface geometry of the damaged area is determined after an application step. Determination of the geometry after application and subsequent repetition of steps c) to g) can improve the accuracy of the process. In particular in cases where the actually applied filler layer or filler layers differ in thickness and/or geometry from the calculated filler layer(s), intermediate geometry determination is beneficial. In that case, deviations from the calculated results can be compensated for in subsequent steps.

The above steps are repeated until the difference in geometry determined in step c) is below or equal to a predetermined value. This value can be predetermined so as to achieve the required degree of matching of the original surface contours of the damaged substrate. If a high degree of matching is required, a lower predetermined difference in geometry will be selected, possibly leading to a higher number of layers to be applied. On the other hand, if a low degree of matching of the original surface is sufficient in an individual case, a higher predetermined difference will be selected as stop criterion.

Figure 2 is a flowchart which represents an embodiment of the automated repair process of the invention. The process starts with the determination of the desired visual properties data of the damaged surface after repair. The damaged surface typically is a damaged portion of the body panel of a motor vehicle. For this purpose a computer controlled robotic unit having an a multi- angle spectrophotometer attached to a robotic arm moveable about four control axes can suitably be used. The spectrophotometric data of an area close to the damaged surface, with the same visual properties as have to be matched by the repair coating, is determined. The determined spectrophotometric data is transmitted to a data storage and processing unit. Subsequently, a colour-

imparting liquid coating composition is selected from an electronic database containing coating composition data of available coating compositions and spectrophotometry data assigned to the compositions. The coating composition having the best match with the determined spectrophotometric data is selected from the database.

The required hiding gradient in which the selected coating composition is to be applied is calculated on the basis of the difference between the determined spectrophotometric data and the spectrophotometric data assigned to the selected coating composition. Subsequently, the selected colour-imparting coating composition is automatically applied to the damaged portion of the coated surface with the previously determined hiding gradient, so as to achieve a hiding power gradient within the coated area, with full hiding power being achieved in the centre of the area of the applied coating layer and the hiding power decreasing towards at least one border of the area of the applied coating layer. This is achieved by gradually decreasing the applied layer thickness of the coating towards the at least one border of the area of the applied coating layer. To achieve this fade- out effect, the speed of the movement of the spray gun relative to the substrate is varied. A faster movement of the spray gun leads to less application of coating material per surface unit, i.e. a lower layer thickness

Automatic application of the coating composition is carried out by a coating robot comprising a spray gun, under the control of the data storage and processing unit. The spray gun typically is attached to a robotic arm moveable about four control axes. The spray gun is connected to a reservoir of the selected coating material.

The invention also relates to a system suitable for carrying out the process for automated repair of a damaged coated surface. The system comprises

a) a determination unit capable of determining the visual properties data of a coated substrate, wherein the geometry determination unit and the damaged substrate are moveable with respect to each other, b) a data storage and processing unit under the control of a program, configured to receive and process visual properties data and configured to select a colour- and/or effect-imparting coating composition, to determine the hiding gradient in which the colour- and/or effect-imparting coating composition has to be applied in order to ensure that the slope of the colour difference between the repaired area and the surrounding area does not exceed a predetermined threshold value, and capable of controlling an application unit, and c) an application unit for applying a coating composition which is under the control of the data storage and processing unit, wherein the application unit and the damaged substrate are moveable with respect to each other.

The system may be positioned on glide tracks or a gantry for movement along the coated substrate to be repaired, such as an automobile. The system generally is implemented as a computer controlled robotic unit, having arms provided with various attachments moveable about various control axes. It is also possible to use a computer controlled robotic unit having only a single arm moveable about various control axes and having exchangeable attachments, such that specific tools as required can be affixed to the arm. Movement of the arms can be caused hydraulically or electrically, or by other suitable means known in the art. In a typical embodiment, the visual properties determination unit is attached to such a robotic arm moveable about various control axes. The visual properties determination unit may be implemented as a spectrophotometer. A further example of a suitable visual properties determination unit is a digital camera. The data storage and processing unit is under the control of a program. The data storage and processing unit is implemented to receive and read visual

properties data transmitted from a visual properties determination unit and optionally from other sources, such as a computer readable visual properties data file. The data storage and processing unit is programmed to select a colour- and/or effect-imparting coating composition, to determine the hiding gradient in which the selected colour- and/or effect-imparting coating composition has to be applied in order to ensure that the slope of the colour difference between the repaired area and the surrounding area does not exceed a predetermined threshold value. The data storage and processing unit also serves as a controller for directing the movement of the moveable units of the system.

The moveable application unit suitably is attached to a robotic arm moveable about various control axes. The application unit may be implemented as a spray nozzle or spout connected to a reservoir of coating material. Suitable means for controlled release of the coating material from the spray nozzle or spout, such as valves and pressurizing equipment, are present as well. Also air brush equipment may be suitable as application unit.

Optionally, the system also comprises a curing unit for curing the applied coating material. The curing unit and the substrate are likewise moveable with respect to each other. In one embodiment, the curing unit is attached to a robotic arm moveable about various control axes. Depending on the curing mechanism of the coating material employed, a suitable curing unit can be selected. If, for example, UV-curable coating material is employed, the curing unit will be implemented as a source of UV radiation. For thermally curable coating materials the curing unit may be implemented as a hot air blower or as a source of (near)infrared radiation.

Optionally, the system may further comprise a sanding unit. The sanding unit is likewise suitably attached to a robotic arm moveable about various control axes.

The sanding unit may be implemented as an electrically driven rotating sanding head.