A feature matching (algorithm based) on classification is proposed, which uses the correlation of the classified images to realize a precise registration after a coarse (registration).

The experiments shows that image registration method based on control point is effective in image registration, and it has higher registration precision.

When these images are fused in pixel-level,they should be registered first and registration precision is required in pixel- level.

The new concept lean and correct assembly based on the application of lean production in assembling process will change the traditional assembly model.

There are many factors affecting the registration accuracy.

The experimental results indicate that the significant features can be detected and matched reliably with this method of being invariant to intensity and contrast variations,and a high registration accuracy of less than systematic mismatch error 0.

A positioning method of high speed SAR plane based on number map after high precision matching is proposed.

Based on the multi channel scanning imagery radiometer of FY-4, some factors which may lowerthe accuracy of imaging register were described, and the approaches to eliminate them were givenaccordingly, which will promote the radiometer research and development.