Following our earlier report of the use of online quality engineering to reduce inspection loss in the printed wiring board assembly（PWBA）process through overall optimization by integrated inspection, we now report on a low-cost inspection system in which the MT-system is used to replace visual inspection. Many expensive high-precision image processing systems are used to automate visual inspection, but by use of the MT-system, even low-resolution images can provide discrimination accuracy equivalent to that achieved by visual inspection. For two PWBA connector fault modes, when the multirecognition-Taguchi multi-RT）method was applied with the normal state as the unit space, the images of the relevant parts were partitioned, and the distance D was calculated synthetically from the individual βand ηvalues, the result was a pass/fail difference exceeding a factor of ten, demonstrating that this low-cost method is adequate for practical use. When the RT-method is employed for image inspection, the discrimination sensitivity and processing speed can also be adjusted by partitioning the image, provided the position and shape of the imaged object and its fault modes are known. Given this simplification of the inspection equipment, further loss reduction can be achieved by optimal positioning in the production process. Cost reductions of 4.9％for inspection and 1.0％for equipment were obtained, totaling 2.29￥per part.